**Contents**



## *Article* **Soil-Improving Cropping Systems for Sustainable and Profitable Farming in Europe**

**Rudi Hessel 1,\*, Guido Wyseure 2, Ioanna S. Panagea 2, Abdallah Alaoui 3, Mark S. Reed 4,5, Hedwig van Delden 6, Melanie Muro 7, Jane Mills 8, Oene Oenema 1, Francisco Areal 4, Erik van den Elsen 1, Simone Verzandvoort 1, Falentijn Assinck 1, Annemie Elsen 9, Jerzy Lipiec 10, Aristeidis Koutroulis 11, Lilian O'Sullivan 12, Martin A. Bolinder 13, Luuk Fleskens 14, Ellen Kandeler 15, Luca Montanarella 16, Marius Heinen 1, Zoltan Toth 17, Moritz Hallama 15, Julián Cuevas 18, Jantiene E. M. Baartman 14, Ilaria Piccoli 19, Tommy Dalgaard 20, Jannes Stolte 21, Jasmine E. Black <sup>8</sup> and Charlotte-Anne Chivers <sup>8</sup>**


**Abstract:** Soils form the basis for agricultural production and other ecosystem services, and soil management should aim at improving their quality and resilience. Within the SoilCare project,

**Citation:** Hessel, R.; Wyseure, G.; Panagea, I.S.; Alaoui, A.; Reed, M.S.; van Delden, H.; Muro, M.; Mills, J.; Oenema, O.; Areal, F.; et al. Soil-Improving Cropping Systems for Sustainable and Profitable Farming in Europe. *Land* **2022**, *11*, 780. https://doi.org/10.3390/ land11060780

Academic Editor: Richard Cruse

Received: 14 April 2022 Accepted: 23 May 2022 Published: 25 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the concept of soil-improving cropping systems (SICS) was developed as a holistic approach to facilitate the adoption of soil management that is sustainable and profitable. SICS selected with stakeholders were monitored and evaluated for environmental, sociocultural, and economic effects to determine profitability and sustainability. Monitoring results were upscaled to European level using modelling and Europe-wide data, and a mapping tool was developed to assist in selection of appropriate SICS across Europe. Furthermore, biophysical, sociocultural, economic, and policy reasons for (non)adoption were studied. Results at the plot/farm scale showed a small positive impact of SICS on environment and soil, no effect on sustainability, and small negative impacts on economic and sociocultural dimensions. Modelling showed that different SICS had different impacts across Europe—indicating the importance of understanding local dynamics in Europe-wide assessments. Work on adoption of SICS confirmed the role economic considerations play in the uptake of SICS, but also highlighted social factors such as trust. The project's results underlined the need for policies that support and enable a transition to more sustainable agricultural practices in a coherent way.

**Keywords:** soil quality; sustainable soil management; adoption; crop management; environmental dimension; sociocultural dimension; economic dimension

#### **1. Introduction**

Crop production in Europe faces the challenge to remain profitable while at the same time achieving environmental sustainability. Average wheat yields in several European. countries are less than what is locally attainable [1–4], possibly because of suboptimal management and/or impairment caused by poor soil quality (defined as 'the capacity of a soil to function within ecosystem and land-use boundaries to sustain biological productivity, maintain environmental quality, and promote plant and animal health', following [5]). In addition, agricultural land faces a number of other threats that may lead to physical, chemical, and biological degradation of the soil [6–9]. These include erosion, compaction, salinization [10], soil pollution, loss of organic matter [11], and loss of soil biodiversity [12]. For example, the use of heavy machinery can lead to soil compaction and impaired root growth [13]; increased soil cultivation and climate change can lead to soil organic matter decline [14]; and narrow rotations may cause biodiversity decline and increased incidence of soil-borne diseases [15]. These forms of soil degradation are often neglected by land managers because of low awareness, low visibility during initial stages of degradation, and a lack of appropriate tools, benchmark values, and policies. As a result, production levels in some cropping systems are maintained by high input (e.g., nutrients and pesticides) and technology (e.g., machinery and breeding), which may mask losses in long-term productivity due to reduced soil quality [16,17]. Such increased use of agricultural inputs may reduce long-term farm profitability because of their costs while also negatively affecting the environment because of unsustainable use of energy and resources in producing inputs [18] and as a consequence of their application (e.g., [19–21]). Soil improvement is necessary to break the negative spiral of degradation, increased inputs, increased costs, and damage to soil and the environment [22]. Maintaining or improving soil quality is crucial for crop production [23] and can especially contribute to remediating forms of soil degradation that are initially hardly visible, such as gradual loss of soil biodiversity and soil organic matter.

Soils are at the intersection of a broad range of land use and environmental challenges. They are critical for economic and environmental well-being, because they form the basis for agricultural production, support high-quality food output [24], and provide a range of other ecosystem services. For example, good-quality soils are more resilient to weather extremes [25] and provide better buffering and cycling of nutrients [26], water purification and regulation, and resilience to pests [27] and climate variability/change [28]. Other ecosystem services provided by soils [29] include provision of biodiversity [30,31] and carbon sequestration, cycling, and regulation [32,33]. Thus, to ensure that sufficient healthy

food for expanding human populations can be grown within planetary boundaries [34], soil management should aim at improving the quality and resilience of land and soil [35].

Attention on soil quality is increasing (e.g., [5,7,36–43]). In Europe, various projects (see, e.g., CORDIS|European Commission (europa.eu), domain 'Food and Natural Resources') have worked on soil threats, prevention of soil degradation, sustainable land management, agricultural management practices, soil functions, and soil quality. There is also increasing recognition of the fact that crop production should be enhanced without compromising the environment [44,45]. More than ever, the important role that soil plays in sustaining life on the planet is being recognized, with high-level objectives at the E.U. scale (e.g., [46]) and the UN Sustainable Development Goals (SDGs) being reliant in large part on sustainable land and soil management [47].

More sustainable farming systems (defined as 'Farming systems that use land resources, including soils, water, and plants, for the production of crops, while simultaneously ensuring the long-term productive potential of these resources and the maintenance of their environmental functions', following the definition of sustainable land management given by WOCAT (www.wocat.net/en/slm (accessed on 13 April 2022)) ' and practices, such as organic farming, conservation agriculture, and precision farming have taken a foothold in Europe [48,49]. For example, Bioland, an association for organic farmers in Germany and Austria, already had more than 5800 members in 2014 [50] and 8500 in 2021 (see https://www.bioland.de/fileadmin/user\_upload/Verband/Entwicklung\_Betriebe\_ und\_Flaeche\_01.svg (accessed on 13 April 2022)). However, these farming systems were not adopted to their full potential and were in some cases even abandoned [51]. Reasons behind this may be the possible negative effect of conservation agriculture on crop yield [39]; the complexity of conservation agriculture, which is management and knowledge intensive [52]; problems with weed and residue management [51]; or the increased occurrence of pests and diseases. There are also cultural and political barriers to the adoption of more sustainable agricultural practices [53]. Barriers to adoption often involve issues around land tenure, access to credit and inputs [7], and other socioeconomic factors, and the lack of knowledge, credible scientific evidence, and good-quality technical advice has also been highlighted [54].

This paper proposed and operationalized a multidisciplinary, multi-actor approach to identifying soil-improving cropping systems (SICS) that are both sustainable and profitable, and hence are more likely to achieve mainstream adoption in agriculture. The focus is on two main aspects, namely evaluation of SICS based on field experiments and modelling and adoption of SICS. To do this, we:


The paper starts by describing the concepts and methodology used for evaluating SICS and studying their adoption (Section 2) and then proceeds by presenting and discussing key findings from SoilCare (Sections 3 and 4). For a literature review that summarizes the main findings of published meta-analyses on SICS, the reader is referred to [55].

#### **2. Concepts and Methodology**

#### *2.1. Conceptualization of SICS*

The term 'cropping system' refers to the crop type, crop rotation, and agronomic management techniques used on a particular field over a period of years [56]. Choices made for these factors can influence the profitability and sustainability of crop production [57–59]. We considered these systems soil-improving if they resulted in a durable increased ability

of the soil to maintain its functions, including food and biomass production, buffering and filtering capacity, and provision of other ecosystem services.

The basic concept adopted in the SoilCare project was that profitability and sustainability of crop production in Europe should be integrated and enhanced. Both are influenced by choices made in farm management, which are in turn influenced by external drivers and factors (Figure 1). External drivers and factors include E.U. policies and international agreements, supply chain and market effects (suppliers, industry, processing, retail, and consumers), macroeconomic conditions, society (public opinion), and pedoclimatic conditions. These external drivers and factors are dynamic and change because of socioeconomic developments, geopolitics, and climate change. As the focus of SoilCare was on arable cropping systems, grazing systems, multisystem farms, and other on-farm activities were not considered.

**Figure 1.** Methodological framework for assessing sustainability and adoptability of soil-improving cropping systems, showing the influence of farm management levels (FML 1–3) on soil quality, environment, crop yield, profitability, and sustainability. LIT refers to literature and other published data, LTE to long-term experiments, and SS to work in the study sites.

At the highest farm management level (FML1, see Figure 1) a choice is made among different types of farming; cropping systems are decided on at FML2, while choices regarding agronomic techniques that are used for management of soil, water, nutrients, and pests are made at FML3. Which farm type is chosen depends on external factors but also on the farm's ownership, resources and social context, such as the education, age, and preferences of the farmer (e.g., [60]). Choices made at this level also influence FML2 and FML3. For example, a choice for organic farming made at FML1 implies crop rotation at FML2 and biological pest management at FML3.

Choices made at all three FMLs have impacts on soil quality, on the environment, and on yield (thus farm economy) (Figure 1). These also influence each other. For example, the occurrence of a soil threat such as erosion influences soil quality as well as crop yield [61]. Crop yield can also influence soil quality, for example, through nutrient mining, rooting effects, and below-ground biomass. When impacts on soil quality and environment are positive, and the balance between production costs and revenues is also positive, the dual targets of farm profitability and environmental sustainability are reached.

The use of SICS improves soil quality and environmental benefits and has positive impacts on the farm economy (Figure 2). Some benefits result directly from the application of proper agronomic techniques; for example, avoiding overapplication of nutrients reduces greenhouse gasses (GHGs) and pollution (soil degradation). Other benefits of

SICS are indirect, as they result from improved soil quality brought about by application of the SICS. For example, improved soil quality improves infiltration and hydrological properties, increases rooting depths and resilience to climate change impacts, and stimulates soil biodiversity [11]. Finally, SICS also have above-ground impacts on vegetation and landscape (e.g., through the use of hedges, buffer strips, trees, terraces, ditches). Such impacts may also contribute to the conservation of biodiversity and wildlife, which may in turn positively influence soil quality.

**Figure 2.** Impacts of agronomic techniques for managing soil, water, nutrients, and pests. One-sided arrows indicate impact, while two-sided arrows indicate that factors influence each other. Note that agronomic techniques are part of cropping systems and correspond to FML3 in Figure 1.

Profitability is a key factor influencing the adoption of SICS [62–66] that is partly influenced by the choice of cropping system and its management and partly by factors that farmers (in Europe) cannot typically control, such as global markets and policies [53]. A key aspect of profitability is production costs, as farmers have more control over this aspect than over the prices they get for their products. Different cropping systems require different types and levels of inputs (e.g., [67]) with different costs. In addition, the choice of cropping system influences the price of the product, which is often higher for organic than for conventional farming.

Conventional farming may become increasingly costly because of rising costs for external inputs and/or for mitigation/restoration measures against soil degradation. In addition, prices of external inputs fluctuate. For example, refinery curtailments due to the COVID-19 pandemic have limited supplies of raw materials, raising input costs by increasing the price of fertilizers for farmers [68]. Price fluctuations of agricultural products are expected to persist and continue to challenge the ability of consumers, producers and authorities to cope with the consequences [69]. In this context of rapid change and longterm challenges, farm profitability is at risk. In line with the Europe 2020 Strategy [70] on achieving smart, sustainable, and inclusive growth, boosting profitability is not only about reducing production costs, or increasing productivity, but also about more sustainable agriculture and the transformation of the food market to green, high-quality products. Smarter and greener agriculture also has the potential to contribute to a more circular bioeconomy and increase the value of agricultural products and the willingness of consumers to buy European agricultural products both inside and outside of the European Union [71,72].

SICS have the potential to reduce costs in the long run by reducing the need for external, costly inputs such as fertilizers and pesticides, reducing energy use for operating machinery, and/or reducing labour input [73–75]. While some SICS may lead to reduced productivity, they may make more efficient used of inputs and thus be more profitable. Costs associated with current unsustainable land use and management are estimated to be in excess of EUR 50 billion per year in the European Union [46]. In the long term, adoption of SICS should help reverse the current trajectory, and when soil quality has improved, efficiency is expected to increase further as a consequence of the reduced need for external inputs and possibly higher production. Additional long-term benefits lie in the reduction of expenditures due to reduced land degradation, GHG emissions, and risk to damages from natural disasters such as storms, droughts, or floods [25].

Various factors influence where SICS are most needed and best suitable and thereby determine the balance between the benefits and drawbacks of SICS and the ways in which these drawbacks can be minimized. These factors include the pedoclimatic zone (zones that are relatively homogeneous concerning climate and soil; see, e.g., [76]), the type of problem that constrains soil quality and crop production, biophysical conditions, and socioeconomic and political conditions. These different conditions require the use of different SICS and determine the applicability, profitability, and environmental impacts of the SICS across Europe. Hence, an assessment of SICS should incorporate environmental, economic, social, and policy aspects while also taking into account future trends in land use and climate change.

#### *2.2. Methods Used for Evaluation of SICS*

The first step in evaluating selected SICS was an in-depth analysis of the benefits and drawbacks of SICS as reported in literature and other published sources [55,77]. This was followed by investigating data from existing long-term experiments (LTEs). Next, we conducted field experiments and stakeholder research in 16 study sites located in different parts of Europe (Table 1, Figure 3), covering different pedoclimatic, socioeconomic, and policy conditions. Literature and other published data were mainly used to assess external drivers and factors (Figure 1). This was supplemented by stakeholder consultation at the E.U. level and modelling. Data from LTEs were mainly used to investigate SICS that show effects only in the long term. The focus of field experiments and stakeholder research in the study sites was primarily on FML3, since soil, water, nutrient and pest management can be adapted in the course of the year and these choices generally have more immediate effects than choices made at FML1 and FML2.

**Table 1.** Overview of SoilCare study sites. Types of crops listed here represent the study site region, not the sites where monitoring was conducted.



**Table 1.** *Cont.*

<sup>1</sup> climatic zones based on the Environmental Stratification of Europe (version 8) [76]; <sup>2</sup> SOC = soil organic carbon decline.

**Figure 3.** The 16 SoilCare study sites. Details on each study site can be found in Table 1.

Within the study sites, different SICS were selected, tested in field, and evaluated in collaboration with stakeholders. Evaluation of SICS was conducted by applying the same assessment methodology at each study site. This general methodology was based on a shared database [78], a common monitoring plan, a unified statistical analysis (according to the experimental design of each experiment) and sustainability assessment. In the field experiments, SICS were compared with a control (usually a standard conventional practice) [79], and SICS were monitored for 2–4 years. Data from the field trials were assessed using a decision tree in terms of soil quality (physical, biological, and chemical); environmental, economic, and sociocultural dimensions; and sustainability, resulting in a score between −1 and 1 for each dimension [80]. For the three dimensions, the following methods were used for scoring:


Detailed results of the evaluation of environmental, economic, and sociocultural dimensions were presented in [79]. For SICS for which data on all three dimensions were available, we calculated the impact on sustainability as the average of the impact on the three dimensions [80].

Finally, the study site results were upscaled to the European level using a storyline, simulation, and policy support process [81–83]. This process combined participation and modelling to better understand the impacts of SICS across Europe and to provide policy support to facilitate the uptake of SICS under different contexts and conditions. As part of the approach, an integrated assessment model (IAM) consisting of spatial, socioeconomic, and environmental simulation models (i.e., the AGMEMOD [84], METRONAMICA [85], PESERA [86], dyna-QUEFTS [87], and MITERRA [88] models) was developed [81]. The IAM was used to simulate possible effects of four scenarios that captured diverse pathways for European agriculture until 2050 (Figure 4). These scenarios differed with regard to challenges to voluntary instruments and mandatory instruments. We used a combination of qualitative and quantitative techniques in a multi-actor approach to develop these scenarios in order to assess how agricultural practices could contribute to sustainable and profitable European agriculture and, finally, to discuss what is needed to enable adoption and implementation of these practices. In addition, for a range of 27 SICS, Europe-wide maps and modelling were combined with expert judgement from study site partners and their stakeholders to provide a SICS potential index based on the applicability, relevance, and impact of each SICS [82]. An interactive web-based tool was developed to help land users and decision makers select suitable SICS throughout Europe (imt.soilcare-project.eu; accessed on 13 April 2022) [83]. This tool allows users to compare different SICS with regard to various aspects, including IAM results and the SICS potential index.

**Table 2.** Results of SICS analysis based on the developed assessment methodology [80]. Values were scored on a range from −1 to 1 for those experiments where data on all three dimensions were available (see [79]). Details on experiments can be found in [79]. Impact on sustainability was the average of environmental impact, economic impact, and sociocultural impact. Negative impacts are indicated by red and positive ones by green. More details are provided in Table S1.


**Figure 4.** Overview of scenario framing linked with scenario titles and motivating factors [82].

*2.3. Concepts and Methodology Used to Study Adoption of SICS*

In the last decade, there have been numerous policy initiatives at the European level that, directly or indirectly, promoted the adoption of beneficial agricultural practices [89,90]. Most recently, the European Green Deal (COM/2019/640 final. https://eur-lex.europa. eu/legal-content/EN/TXT/?qid=1576150542719&uri=COM%3A2019%3A640%3AFIN (accessed on 13 April 2022) and the new Soil Strategy (COM/2021/699 final. https://eur-lex. europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52021DC0699 (accessed on 13 April 2022)) set out the roadmap for making the European Union's economy more sustainable and identified several key actions that will be crucial in advancing land and soil protection in Europe. With this shift to more sustainable practices comes increasing pressure on farmers to change how they operate and adopt new techniques and practices. However, innovations associated with potential benefits to soil quality have not yet been adopted to their full potential and have, in some cases, even been abandoned, raising the question of why support for and adoption of these practices by European farmers is still weak.

Adoption of new or modified agricultural practices by farmers is a complex process that is governed not only by physical effectiveness and economics of agricultural practices but by a range of other factors, including individual, social, cultural, and policy-related factors [91]. These include internal factors, such as the farmer's own views on farming, the influence of peers and advisers, their perceived difficulties in implementing practices, and sociodemographic characteristics, and external factors, such as pedoclimatic conditions, markets, and policies [91,92]. Economics is an important factor and is often considered to be the main driver for adoption. However, overlooking some of the other factors may be one of the main reasons why seemingly advantageous measures have not been adopted widely by farming communities (e.g., [93,94]). Factors influencing the adoption of sustainable farming practices in Europe range from the land managers' access to information, training, and technical advice [95], to the performance of a particular practice in terms of yield increase or reduction in production costs or work time [96,97], to aspects rooted in the social and cultural context or in the personality of the individual land user. Social factors include the underlying motives (e.g., social or personal rewards) and attitude towards risks [98]; personality traits such as openness to new experience or resistance to change; what land users perceive others expect from them; and land users' perceptions of the relative benefits, costs, and risks associated with a particular practice [97,99]. In addition, farming practices, e.g., conservation measures, must be compatible with the values of

landowners [97], cultural constructions of 'good farming' [100,101], and farmers' sense of professional identity and aesthetic preferences [102]. Finally, social factors such as trust and acceptability also influence adoption [59]. The dynamics of trust (across space, time, social groups, and culture) can explain how innovations are adopted through social learning and collaborative learning processes. The speed and spatial scale at which trust can develop likely depends on the extent to which it is possible to find or develop shared values, converge towards compatible epistemologies, and find common interests that can transcend sociocultural, political, and economic differences. It should be noted that engagement processes work differently and can lead to different outcomes when they operate over different spatial and temporal scales [103] so that engagement processes should be adapted to local conditions.

To understand all the factors that influence adoption and take them into account, a multidisciplinary integrated approach is needed, including, e.g., soil science (physics, chemistry, and biology), agronomy, hydrology, ecology, climatology, economics, and social sciences. In addition, a variety of stakeholders should be involved, as multiple stakeholders influence the ways in which crops are produced. This makes adoption site-specific, as every area has its own unique combination of biophysical, sociocultural, economic, and policy factors, as well as its own set of stakeholders. Thus, adoption research necessitates the involvement of scientists and practitioners from multiple disciplines, as well as active involvement of stakeholders. For SoilCare, this contextual nature of sociocultural and political drivers meant, on the one hand, that a robust assessment of adoption factors could be performed only at the study-site scale, so the broader suitability of SICS across Europe was considered primarily based on biophysical and environmental characteristics. On the other hand, the adoption work could still offer insights into more general trends with respect to the typical factors that can influence the adoption of particular SICS.

The SoilCare research on the adoption of SICS focussed on understanding the reasons why SICS are being adopted or not adopted and how farmers can be encouraged through appropriate incentives to adopt suitable SICS. The methods applied addressed four types of factors affecting adoption:


Adoption should be considered not only with regard to a range of factors but also at different scales, from the farm scale to the European scale, because operations and actors in the agricultural value chain stretch out over these scales in the supply, purchase, processing, and distribution of agricultural products. Furthermore, socioeconomic developments, such as changing public awareness of the importance of sustainable production and the consequences this has for the prices consumers and companies are willing to pay for sustainably produced food, have an influence on adoption.

The storyline, simulation, and policy support approach presented in Section 2.2 was used to assess the adoption potential of SICS at the European scale. By developing different scenarios or pathways for European agriculture using a combination of sociocultural, technological, economic, environmental, and political factors and drivers of change, the impact of (policy) actions on enhancing adoption of SICS was assessed under various current and future conditions to arrive to options that would be robust across scenarios or target specific factors/barriers and enablers within scenarios.

#### **3. Key findings**

#### *3.1. Main Effects of SICS*

Table 2 provides an overview of monitoring results from 11 countries, derived from [79], which contained details on the experiments. Overall, these results showed a small positive impact of SICS (when compared with the control) on environment (including soil quality), no effect on sustainability, and a small negative impact on economics and the sociocultural dimension. Some treatments showed both high and low values of impact scores on the dimensions of the sustainability assessment, which illustrated trade-offs in the performance of a SICS. Some treatments yielded only zero or negative impacts (e.g., early wheat sowing, FR), and other treatments gave positive impact scores in all dimensions (e.g., N fertilization with straw/stalk, HU).

#### 3.1.1. Environmental Dimension

In general, the SoilCare field experiments were too short to show clear statistically significant effects on productivity (yield or relative yield), SOC, structure stability (water stable aggregates), infiltration rate (hydraulic conductivity), biological activity (earthworm counting), or soil bulk density. Hydraulic conductivity and bulk density have large spatial and temporal variability in the field, which made it difficult to detect significant differences without dramatically increasing the number of measurements. The study site in Poland illustrated this spatial variability well [108]. Overall, SICS showed a small but positive effect on soil properties and the environmental dimension (Table 2); 6 out of 16 experiments showed a positive impact of SICS, 1 experiment showed a negative impact, and 9 experiments showed no change. Although not significant from a statistical point of view, slight improvements were found for most of the experiments. In addition, stakeholders and scientists in many cases could visually detect and evaluate positive effects of SICS, in properties such as soil structure or infiltration, or negative effects, such as weed infestation.

In addition, the SoilCare monitoring results provided the following insights based on the evaluation of the environmental dimension for all SICS [79].

**Tillage**: For most experiments, reduced tillage and noninversion tillage had a positive effect on soil characteristics and did not lead to lower yields. The noninversion tillage in a Belgian experiment presented better physical characteristics (hydraulic conductivity and aggregate stability). The minimized tillage in a Hungarian LTE [109] also improved the aggregate stability and SOC content when compared with conventional ploughing and increased the plant available water content [110]. A Czech experiment [111] showed that zero tillage was difficult for heavy soils and root crops but significantly improved the topsoil SOC, bulk density and aggregate stability when compared with conventional ploughing. However, the increase in SOC did not affect the plant available water content [110]. Pest and weed control was a challenge in the Belgian experiments under strip tillage and significantly impacted plant growth and crop yield. Weed control was also a major issue in several no-tillage systems; this resulted in increasing use of herbicides.

**Soil compaction**: Subsoiling is a means to alleviating compaction [112] by breaking up the compaction of deeper soil layers. In a Romanian experiment, subsoiling was suggested to a depth of 60 cm every 3 to 4 years to improve the aggregate stability and hydraulic conductivity and reduce the soil bulk density while maintaining a good crop yield. A Swedish experiment on a naturally compacted soil found that mechanical subsoiling, with or without incorporation of organic materials, had a positive impact on root growth and rooting depth. In a U.K. experiment, different physical and biological methods for compaction alleviation were explored. Ploughing was the most effective method for opening up the soil structure and alleviating topsoil compaction, but no effect on crop yield was observed in the two years of study [113]. The results of an Italian experiment that used different crops and tillage methods to reduce soil compaction indicated a higher risk of crop failure and difficulties with weed control (requiring herbicides) under no-tillage systems. Nevertheless, reduced-tillage systems had the potential to increase farm environmental and agronomic sustainability according to the relative sustainability index, which was based on 11 physical chemical and biological properties [114].

**Fertilizers and amendments**: An LTE in Hungary [115] showed significant positive effects on yield and soil structure (water stable aggregates and bulk density) when incorporating crop residues into the soil or when applying farmyard manure. The SOC content and plant available water content were not significantly increased [110] despite the positive effects on yield and soil structure. A Belgian experiment compared adding woodchips, compost, and pig manure with a control (no additions). The C/N ratio of the amendments helped to explain the availability of nutrients for crops. In a Portuguese experiment, urban sludge from wastewater treatment plants increased SOC and soil nutrient contents and earthworm population without affecting the heavy metal concentration in the soil in the short term. In a Danish experiment [116], the use of manure helped to reduce the crop yield gap between organic cultivation treatments and conventional control treatment with mineral fertilizers and to reduce soil bulk density. A study in Italy [117] examined the effects of SICS with different crop residue management and concluded that crop residues reduced the need for fertilizers. The Controlled Uptake Long-Term Ammonium Nutrition (CULTAN) method in Switzerland reduced the risk of nitrate leaching.

Data from LTEs in Belgium, Denmark, the United Kingdom, and Hungary indicated that soil management influenced soil biota, which in turn influenced soil quality [118]. The fungal communities were found to be very variable across sites located in different soil types and climatic regions, and only fertilization showed a consistent effect on arbuscular mycorrhizal fungi and plant pathogenic fungi, whereas the responses to tillage, cover crops, and organic amendments were site, soil, and crop-species specific. A study in Poland [119] examined the effects of adding spent mushroom substrate and chicken manure to soils on soil fungal community composition and mycobiome diversity. Both increased the abundance of fungi and reduced the relative abundance of several potential crop pathogens. These results provided a novel insight into the fungal communities associated with organic additives, which should be beneficial in the task of managing the soil mycobiome as well as crop protection and productivity. Both additives were also found to result in increased SOC [120].

**Cover crops**: Over the last decade, the increased use of cover crops between growing seasons has motivated the inclusion of this practice in the field experiments of many study sites. The benefits of cover crops are generally well accepted, and recent research has indicated that they can also enhance the availability of soil P and have positive effects on the soil microbial community [121–123] and earthworm abundance [116]. Positive effects were also illustrated by experiments in the study sites in Norway, Portugal, Denmark, France, Italy, and Germany [79]. However, because of global warming, which was visible in the results of the meteorological analyses for these study sites, the lack of freezing during recent winters meant that cover crops survived the winter. In that case, either herbicides or mechanical measures were required to kill them in spring. This is an important issue for further investigation. In the German experiment, the possible negative effect of glyphosate on soil quality was investigated by using different soil microbiological methods. An increase was found in ß-glucosidase activity (C-cycling enzyme) as a stress response of soil

microorganisms after a period of seven days of application (unpublished data). Since no significant changes in microbial community composition occurred after the application of glyphosate in the field experiment, these effects were considered minor. Nevertheless, transport of glyphosate by preferential flow into deeper zones of soils might hinder the fast decay of this compound by bacterial glyphosate degraders [124]. Banning herbicides would require high-precision shallow tillage/mechanical weeding before seeding of the crops so as not to destroy the benefits of cover crops on soils again. Furthermore, mechanical weeding might mean more fuel use and GHG emissions.

In Greece and Spain, the tested cropping systems were vineyards, stone fruit, and olive orchards. In Crete (Greece), erosion reduction was the major challenge. Crete had historically high rainfall in October 2017 and some other heavy rainfall events afterwards. It was concluded that cover crops in vineyards and minimum tillage in olive orchards could reduce the erosion rates during extreme rainfall events and increase the earthworm density. The conversion of the traditional orange orchards to avocado cultivation resulted in a statistically significant reduction in erosion and increased SOC content and hydraulic conductivity [125]. Almería (southeast Spain), as the driest and hottest place in Europe, focused on water savings by deficit irrigation and erosion reduction with different soil cover or cultivation methods. The application of different combinations of irrigation led to water savings of up to 15%, but topsoil management did not cause significant differences in yield, fruit quality, or soil quality apart from an unexplained increase in the electrical conductivity when cover crops were used. [79].

#### 3.1.2. Economic Impact (Profitability)

Table 2 indicates that the economic impact was positive for 4 out of 16 experiments, while it was negative for 6 and did not show change for the remaining 6. The average impact was −0.13, but the median impact was 0.01. Closer inspection of detailed data on costs and benefits (available for 15 SICS in Table S2) reveals that:


This indicates that, at the field/farm level, short-term profitability was generally positive for the SICS (13 out of 15), but in half of the cases, it was lower for the SICS than for the control.

#### 3.1.3. Sociocultural Impact

Table 2 indicates that for 3 out of 16 SICS, the sociocultural impact was positive; for 7, it was negative; and for 6, there was no change. The average impact was −0.04, and the median impact was −0.02. Analysis of data from 16 SICS showed (Table S3):


This indicates that application of SICS had a positive impact on farmer reputation, as land users applying SICS were usually considered to be innovative. Workload did not show a clear trend, as for some SICS it was higher, while for others, it was lower. Many SICS are perceived to be associated with potential risks, most importantly the risk of crop failure and/or other economic risks (such as, e.g., high investment costs). The respondents often related the risk of crop failure to specific weather conditions such as prolonged dry spells or heavy rainfalls.

#### 3.1.4. Main Results Upscaling SICS

Upscaling results included the potential for applying SICS across Europe as well as an assessment of the impact of SICS application under future uncertainty using the four developed scenarios (Figure 4). Figure 5 shows the SICS Potential Index for cover crops (for 2018) as an example result of the first type of upscaling activity. The figure shows that differences in climate, soil, and land use conditions resulted in differences in the applicability, relevance, and impact (on SOM, erosion, and yield) of cover crop use and hence the potential to apply them across Europe. Regarding the second type of upscaling activity, the results of the IAM indicated that over time (until 2050), in the different scenarios, different changes were expected in consumption, production and net exports, yield, gross margin, SOC, and erosion. This was due to, amongst other factors, growth in population, changes in diets, trade flows, climate change, technological changes, and changes in agricultural practices (i.e., through application of SICS). While some drivers were expected to result in impacts in the same direction in all scenarios (e.g., population growth was likely to lead to more consumption), other drivers could impact in very different ways. This was caused by regional differences such as, e.g., climate change impacting on yield levels and gross margins based on country-specific crop prices and location-specific biophysical conditions.

**Figure 5.** Examples of modelling results. Left: SICS potential index for cover crops (2018) [82]; right: scenarios leading to the highest yield increase in 2050 [81]. RttB = race to the bottom, UP = under pressure, LS = local and sustainable, CS = caring and sharing (see Figure 4).

As expected because of its formulation, the Caring and Sharing (CS) scenario, which assumed wide application of SICS (Figure 4), was likely to provide the best environmental impacts (i.e., increased, or stable SOC content and reduced erosion rates), and the Race to the Bottom scenario, assuming limited application, was likely to provide the worst.

An important finding, however, is that although the CS scenario in most regions led to highest yield impacts (Figure 5), the gross margin of SICS uptake under this scenario was negative in many NUTS-2 regions [81]. The most important factor contributing to this was the high implementation costs assumed when combinations of SICS were implemented. Despite sustainability being high on the agenda in the CS scenario, (financial) policy support would therefore likely be needed to enhance uptake of SICS. Alternatively, value added through additional products and services and valuation of environmental co-benefits could be a pathway to widespread SICS adoption.

The cost–benefit analysis showed a mixed spatial pattern of scenarios that had the highest gross margin across Europe. The reason for this was that the combination of drivers played out differently in different parts of Europe, indicating the complexity of the issue and the importance of understanding local dynamics. Using these scenarios for policy support also illustrated the importance of tailored/context-specific policy design/development, as selected options were often expected to have different performance under different scenarios.

#### *3.2. Adoption of SICS*

As illustrated in Table 3, there is a wide range of issues affecting adoption of sustainable soil management. Following this, country-specific issues stem from the fundamental E.U. level factors listed below:



#### **Table 3.** Adoption factors in SoilCare study sites.

#### **4. Discussion and Conclusions**

#### *4.1. Evaluation of SICS*

SoilCare provided scientific evidence on the potential of SICS at 16 study sites and Europe-wide. Although monitoring in study sites did not provide conclusive results in all cases, it did show positive effects on most soil properties as well as a small positive impact on the environmental dimension. This was in line with the main results reported by meta-analyses such as those reviewed in [55]. No significant changes were observed for sustainability or for the economic dimension at the farm level. Nevertheless, most SICS were found to be profitable, since benefits were often higher than costs. However, in a small majority of cases, the profitability of the SICS was lower than for the control. The sociocultural dimension was slightly negative on average, mainly because SICS were perceived to be risky by farmers. The respondents often related the risk of crop failure to specific weather conditions, such as prolonged dry spells or heavy rainfalls. Indeed, it is known that some SICS are more sensitive to yearly variations than conventional practices, such as, for example, organic farming (e.g., [126–128]). On the other hand, weather conditions would in most cases also challenge the performance of the controls, but the risks associated with these practices were not assessed in our study. As described in Section 3.2, risks can also be higher during the transition period from conventional to more sustainable practices, although our economic data overall showed similar revenues for SICS and control. A final reason why SICS are perceived to be risky may have to do with uncertainty and risk aversion on the part of farmers, as switching from normal practices to

SICS means a switch from familiar ways to something new. A repeated questionnaire after a few years of implementation of SICS might help to investigate whether risk perception of SICS changes over time.

It should be noted that our results were obtained at the plot/farm level and based on only 3 (max. 4 for some study sites) years of monitoring. This has several implications:


In addition, the assessment methodology for SICS that was applied may need further development and refinement. Both the assessment methodology and its application relied on expert opinion, not only with regard to the weights assigned to different parameters and to the environmental, economic, and sociocultural dimensions but with regard to the underlying concepts. For example, the economic dimension did not give very positive results for the SICS, which was at least partly due to the fact that more importance was attached to the relative difference between SICS and a control than to the difference between benefits and costs. As a result, SICS with a positive benefit/cost ratio scored negatively on the economic dimension because the control had a more positive benefit/cost ratio. This may actually reflect reality, as this meant that farmers would earn less by applying SICS, but the point here is to illustrate that assumptions made in the assessment methodology did have an impact on the outcome. Such assumptions are open to discussion and can be subject to revision as more data become available.

Furthermore, the outcome of the assessment was, of course, influenced by the input. Although this may seem trivial, it is not, as the input by necessity has to be a combination of different types of data (quantitative as well as qualitative) originating from different sources (including scientific experiments but also stakeholder perceptions), sometimes with gaps or limitations.

For all of these reasons, the results of the evaluation should not be seen as a final result, but rather as an indication that forms a starting point for discussion with stakeholders (from farmers to scientists and policy makers).

#### *4.2. Adoption of SICS*

SoilCare also delivered knowledge on how to promote the adoption of SICS to individual farmers, European institutions, member state authorities, and agricultural advisory services. The analyses carried out in SoilCare delivered increased insight into biophysical, economic, social, and political barriers to adoption, several of which corresponded to barriers already identified in [52] for conservation agriculture. SoilCare also provided solutions that could help to overcome such barriers. The results confirmed the crucial role of social factors such as trust in adoption and underlined the need for policies that support and enable a transition to more sustainable agricultural practices in a coherent way.

Historically, soil has been an overlooked component in studies on ecosystem service and policy decision making [139]. At a policy level, the removal of the proposed Soil Framework Directive (COM (2006) 232 final) in 2014 highlighted a need and an opportunity to think about soils differently [140]. The SoilCare project represents a short timeline when set against its objectives; however, it is also noteworthy that the role of soils transitioned to being at the heart of high-level ambitious European policies such as the European Green Deal and the CAP Farm-to-Fork and Biodiversity Strategies during the project lifetime. This was complemented by a focus on soil research and innovation in the European Joint Programme on soil and a mission in the area of soil health and food. E.U. policies to target soil and environmental objectives have been criticized for their lack of nuance to account for localized conditions in the past. In this regard, the SoilCare project has framed a methodology for SICS that reflects the key dimensions that must be considered in governance for local but also wider-scale dynamics. Although more work is required, the lessons learned, particularly in relation to those SICS that exhibited promise, should be further explored and leveraged under the new opportunities that now exist within the policy, research, and innovation space. Table 4 provides an overview of policy recommendations resulting from SoilCare.

**Table 4.** Policy recommendations resulting from SoilCare, after [141].

*Recommendation I: Define long-term ambitions and targets*


*Recommendation II: Increase coherence and exploit synergies between policies more effectively* There are many different pieces of legislation that can work better together if coherence and integration between them is improved. In addition, stakeholders noted that some SICS might not align with existing policy objectives. At the E.U. and country levels, policy conflicts and synergies need to be carefully analysed and aligned to avoid discouraging a transition to sustainable farming.

#### **Table 4.** *Cont.*

*Recommendation III: Design targeted economic instruments that facilitate a transition to sustainable practices and reward environmental benefits delivered*

The CAP should strive to be less prescriptive and avoid one-size-fits-all approaches, instead providing farmers with a general direction clearly defined by targets and empowering them to take steps towards these targets. There is a need to consider the different conditions in which farmers operate (e.g., differences in tenure), and measures need to be flexible enough to allow for regional differences. Priority should be given to farming techniques that are also means of food production and are both profitable and sustainable.

*Recommendation IV: Strengthen existing and establish new opportunities for learning and knowledge exchange for farmers*

Strengthen capacity of Farm Advisory Services: These are valuable sources of information for farmers, but their independence and neutrality should be ensured. Advisers need to learn about new practices, their practical application and costs, and benefits to support farmers. Ref. [142] gave suggestions for achieving more effective advisory services.

Inform farmers about new developments and insights: Dissemination of knowledge, awareness raising, and education are important components of policy interventions, and they should be used in parallel with economic and legislative instruments [143].

*Recommendation V: Strengthen monitoring and enforcement*

At the E.U. level, there is a need to establish a clear, robust, and reliable monitoring and enforcement system for the CAP. At the country level, stronger monitoring and enforcement systems require the training of farm inspectors, who, like farmers, need to understand regulatory requirements and their practical implementation.

#### *4.3. Sustainability and Profitability*

Results obtained at the farm level indicated a small decrease in profitability and a small positive effect on the environmental dimension (Table 2). As discussed above, however, there is a need to consider larger temporal and spatial scales. This was done in the modelling approach, which was used to upscale results from the different study sites and integrate these results with factors operating at the European scale, such as policy development, macroeconomy, societal developments, and climate change. Several scenarios of possible developments with a time horizon of 2050 were simulated. Simulations showed that scenarios in which sustainability was given priority resulted in better soil quality and better environmental conditions. However, while SICS would be profitable to society in the long term, they may not always be profitable to farmers in the short term. As shortterm benefit over conventional practice is a key point for farmers [63], and as modelling suggested that SICS outperformed control treatments in the longer term, some form of compensation and support to farmers would be required to stimulate adoption of SICS, for example, in the form of bridge payments.

#### *4.4. Conclusions*

The need for sustainable soil management is evident from the literature. Soils are critical for economic and environmental well-being because they provide a range of ecosystem services and form the basis for agricultural production. They are at the intersection of a broad range of agricultural and land use challenges. Soil management should aim at improving the quality and resilience of land and soil. Within the SoilCare project, the concept of soil-improving cropping systems (SICS) was developed and applied. SICS can play an important role in the transition towards more sustainable agricultural production that can also be profitable. In practice, the effectiveness of SICS is difficult to demonstrate within the lifespan of a single project, as results vary from year to year because of different conditions, such as different weather and price fluctuations of inputs and crops. Furthermore, many SICS are expected to reach their full potential only after a long time. SoilCare paved the way for further research on SICS by developing an assessment methodology for SICS, a database for SICS data, and a modelling approach for upscaling and scenario evaluation. In addition, SoilCare contributed to the understanding of adoption factors and

provided a first assessment of a range of SICS. Whilst our work on adoption confirmed the role economic considerations play in the uptake of SICS, it also highlighted the influence of social factors, such as trust, and of knowledge. This underlines the need for policies that support and enable a transition to more sustainable agricultural practices in a coherent way.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10.3 390/land11060780/s1, Table S1: Results of environmental dimension, Table S2: Results of economic dimension, Table S3: Results of sociocultural dimension.

**Author Contributions:** Conceptualization: R.H., O.O., M.S.R., A.A., G.W., H.v.D., M.M., J.M., E.v.d.E., S.V. and F.A. (Falentijn Assinck); methodology: R.H., M.S.R., A.A., G.W., H.v.D., M.M., J.M., I.S.P., E.v.d.E., S.V., F.A. (Francisco Areal) and J.S.; software: I.S.P. and H.v.D.; validation: H.v.D., A.A., A.E., J.L., A.K., M.A.B., E.K., Z.T., M.H. (Moritz Hallama) and J.S.; investigation: A.A., A.E., J.L., A.K., M.A.B., E.K., Z.T., M.H. (Moritz Hallama), J.C., I.P., T.D. and J.S.; data curation: S.V., E.v.d.E., G.W., I.S.P. and H.v.D.; writing—original draft preparation: R.H., O.O., M.S.R., A.A., G.W., H.v.D., M.M., J.M. and J.E.B.; writing—review and editing: F.A. (Francisco Areal), M.H. (Marius Heinen), I.S.P., A.E., M.A.B., E.K., M.H. (Moritz Hallama), J.C., L.O., L.F., L.M., J.E.M.B. and J.S.; visualization: R.H., H.v.D., J.M. and C.-A.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 677407 (SoilCare project). R.H., O.O., E.v.d.E., S.V., F.A. (Falentijn Assinck) and M.H. (Marius Heinen) also received funding from the Dutch Ministry of LNV via Kennis Basis programma 34, project KB-34-008-005. The contribution of the University of Hohenheim was partly financially supported by the German Research Foundation (DFG) under the Collaborative Research Centre 1253 CAMPOS (DFG grant SFB 1253/1 2017).

**Institutional Review Board Statement:** Not applicable in this study.

**Informed Consent Statement:** Ethical standards and guidelines have been applied to the collection, processing and storage of data about persons, in accordance with the agreed project ethical statement. All subjects gave their informed consent for inclusion before they participated in the study.

**Data Availability Statement:** The data presented in this study are available in the supplementary material.

**Acknowledgments:** We thank the stakeholders in the SoilCare study sites. Without their collaboration, this research would not have been possible.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **The Effects of Soil Improving Cropping Systems (SICS) on Soil Erosion and Soil Organic Carbon Stocks across Europe: A Simulation Study**

**Jantiene E. M. Baartman 1,\*, Joao Pedro Nunes 1, Hedwig van Delden 2, Roel Vanhout <sup>2</sup> and Luuk Fleskens <sup>1</sup>**


**Abstract:** Healthy soils are fundamental for sustainable agriculture. Soil Improving Cropping Systems (SICS) aim to make land use and food production more sustainable. To evaluate the effect of SICS at EU scale, a modelling approach was taken. This study simulated the effects of SICS on two principal indicators of soil health (Soil Organic Carbon stocks) and land degradation (soil erosion) across Europe using the spatially explicit PESERA model. Four scenarios with varying levels and combinations of cover crops, mulching, soil compaction alleviation and minimum tillage were implemented and simulated until 2050. Results showed that while in the scenario without SICS, erosion slightly increased on average across Europe, it significantly decreased in the scenario with the highest level of SICS applied, especially in the cropping areas in the central European Loess Belt. Regarding SOC stocks, the simulations show a substantial decrease for the scenario without SICS and a slight overall decrease for the medium level scenario and the scenario with a mix of high, medium and no SICS. The scenario with a high level of SICS implementation showed an overall increase in SOC stocks across Europe. Potential future improvements include incorporating dynamic land use, climate change and an optimal spatial allocation of SICS.

**Keywords:** large-scale modelling; Europe; soil health; SOC stocks; soil erosion; scenarios; sustainable soil management

#### **1. Introduction**

A well-functioning, healthy soil is fundamental for sustainable agriculture. Soil quality and soil health are increasingly considered important topics on the political and public agenda (e.g., [1,2]), and are also getting attention in the scientific community (e.g., [3,4]). This is reflected in, among others, the Sustainable Development Goals (SDGs; https:// sdgs.un.org/goals (accessed on 12 April 2022)), where soil together with land use and management play an important role in SDG 1 (no poverty), 2 (zero hunger), 12 (responsible consumption and production), 13 (climate action) and especially SDG 15 (Life on Land) [2,5]. Moreover, in the current Farm to Fork Strategy (F2F, [6]), as part of the European Green Deal [7], sustainable food production is an important goal; the F2F aims at neutral or positive environmental impact, mitigating climate change, reversing the loss of biodiversity and ensuring food security. Land use and land management play a key role in achieving these policy aims and reversing the current trend of land degradation [8]. For example, the F2F strategy targets to 'bring back at least 10% of agricultural areas under high-diversity landscape features (with buffer strips, rotational or non-rotational fallow land, hedges, non-productive trees, terrace walls and ponds)' and 'have 25% of the EU's agricultural land as organic farming by 2030' [9]. These strategies are also developed as the costs of unsustainable land management are estimated to exceed €50 billion per year [10].

**Citation:** Baartman, J.E.M.; Nunes, J.P.; van Delden, H.; Vanhout, R.; Fleskens, L. The Effects of Soil Improving Cropping Systems (SICS) on Soil Erosion and Soil Organic Carbon Stocks across Europe: A Simulation Study. *Land* **2022**, *11*, 943. https://doi.org/10.3390/ land11060943

Academic Editors: Guido Wyseure, Julián Cuevas González and Jean Poesen

Received: 28 April 2022 Accepted: 14 June 2022 Published: 19 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The measures mentioned in the F2F strategy are only a few of the very many existing land management options to improve soil health and reverse or prevent land degradation, ranging from farm and field to village and watershed or community scales (e.g., [11,12] and https://qcat.wocat.net/en/wocat/ (accessed on 12 April 2022)). Among those many options, some measures are common in annual and perennial agriculture across Europe. For example, maintaining a (winter) cover crop is widely applied [13–15]. No-tillage or minimum tillage has been estimated to be applied on 25% of the agricultural land in the EU [16]. Mulching is applied to reduce splash erosion and increase soil moisture [17,18]. Crop residue management [19] and/or maintaining a minimum soil cover is also widely applied [12,17]. Grass strips are applied at field borders [20] to reduce runoff and catch sediments [18,21] and as a means to reduce leaching of nutrients [21,22] and/or pesticides [23]. Rodrigues et al. [19] for example show that reduced tillage and soil protective measures can play an important role in soil carbon sequestration across the EU. Maetens et al. [18] investigated the effect of various soil and water conservation measures on runoff and soil loss across Europe.

These practices affect the farming and cropping systems, aiming to make land use and food production more sustainable. As defined in Hessel et al. [24], cropping system refers to crop type, crop rotation and the agronomic management techniques used. Soil improving cropping systems (SICS) can be defined as cropping systems that result in a durable increased ability of the soil to fulfil its functions, including food and biomass production, buffering and filtering capacity and provision of other ecosystem services [24]. However, the uptake and choice of SICS will vary due to external factors, such as EU policies, market effects, society and pedo-climatic conditions. In addition, these factors are dynamic in time as they are affected by e.g., climate change, geo-politics, consumer purchase power and preferences, technological advances and other developments [25,26]. Hence, when assessing the effects of SICS on improving soil health and combatting land degradation at continental scale it is important to consider divergent trends in these factors that affect the uptake of SICS (e.g., [26,27]).

Soil health and land degradation are both broad terms [2] that include many aspects. Soil health has been defined as the continued capacity of soil to function as a vital living ecosystem that sustains plants, animals and humans [3]; encompasses biological productivity, soil life and biodiversity; enhances its role in water quality and regulation and mitigates climate change. Similarly, land degradation entails many different processes, such as salinisation, nutrient depletion, dehydration, erosion by water or wind, compaction, soil pollution, loss of soil organic matter and soil biodiversity etc., [28–31]. In this study, we focused on one principal indicator for each aspect: Soil Organic Carbon (SOC) as a principal indicator for soil health [32,33] and soil erosion (by water) as an indicator and widely occurring process of land degradation [34,35]. Moreover, Kutter et al. [12] in their review on policy measures for agricultural soil conservation in the EU, found that most measures focused on erosion by water, followed by decline in organic matter.

Upscaling the assessment of the impact of measures from e.g., field or farm level to country or wider (e.g., EU) scale is challenging as measuring at this scale is infeasible [34]. Modelling is a common approach and can also include simulation of scenarios of e.g., climate change effects and policy adoption [36,37]. At EU wide scale, soil erosion was estimated by Panagos et al. [34], based on the RUSLE approach. EU wide SOC estimates include e.g., [38–40]. The RUSLE-based erosion estimates by Panagos et al. [34] also include the effect of mitigation options such as conservation tillage, plant residues and winter crop cover [16] and contour farming, stone walls and grass margins [41]. Modelling estimates of climate change and land use change effects on SOC are abundant, e.g., [42–45], and various studies quantified the effects of agricultural practices on carbon sequestration [46–48]. Lugato et al. [49] included straw incorporation, reduced tillage, their combination, ley cropping systems and cover crops into their spatially explicit modelling scenarios.

The SoilCare project (https://www.soilcare-project.eu/ (accessed on 12 April 2022) [24]) aimed to identify and evaluate promising soil improving cropping systems and agronomic

techniques that increase the profitability and sustainability of agriculture across Europe. In addition to field trials [50,51], the project used a modelling approach to upscale the effects of SICS to EU scale, in a spatially explicit way. To ensure that sufficient healthy food for expanding human populations can be grown within planetary boundaries [52], soil management should aim at improving the health and resilience of land and soil [8]. In this study we evaluated how soil improving cropping systems (SICS) impact land degradation (specifically erosion) and soil health (specifically SOC stocks) across Europe, through the application of the PESERA model. For this purpose, we improved and further developed the PESERA model both in terms of input data improvements and in parameterisation and calibration of SICS and a range of crops, in four climate zones. Moreover, to be able to assess the impacts of SICS, existing land management options have been adapted in the model. Four scenarios, developed within the SoilCare project, were simulated until 2050, with varying application of (combinations of) SICS in each scenario.

#### **2. Methods**

#### *2.1. PESERA Model Description*

The Pan-European Soil Erosion Risk Assessment (PESERA) model simulates biophysical processes including above-ground biomass production, soil erosion risk, soil water deficit and soil humus content, using a monthly time-step. The model was originally developed by Kirkby et al., [53] and has been applied in various agro-ecological zones e.g., [54–57]. A brief technical description is given here, based on Kirkby et al. [53], where all details can be found. PESERA is a process-based and spatially distributed model which combines the effect of topography, climate and soil properties. A schematic model structure is provided in Figure 1. The model has three conceptual stages: (i) A storage threshold model to convert daily rainfall to daily total overland flow runoff; (ii) a power law to estimate sediment transport from runoff and slope gradient. The model interprets sediment transported to the base of a hillslope as average erosion loss. No flow or sediment routing over multiple cells is included; and (iii) integration of daily rates over the frequency distribution of daily rainfalls to estimate monthly erosion rates.

**Figure 1.** Schematic overview of processes in the PESERA model.

In the first step, a simple storage or bucket model is used to convert daily rainfall into daily runoff, which is estimated as the rainfall minus the threshold storage. The threshold storage depends dynamically on soil properties, vegetation cover and soil moisture status, varying over the year. The most important soil factors that determine the threshold storage beneath the vegetation-covered fraction of the surface are texture, depth (if shallow) and organic matter. Where the surface is not protected by vegetation, the susceptibility of the soil to crusting and the duration of crusting conditions generally determine a lower threshold. The final threshold is a weighted average from vegetated and bare fractions of the surface. Corrections are made for the soil water deficit, which may reduce the threshold where the soil is close to saturation. Transpiration is used to drive a generic plant growth model for biomass, constrained as necessary by land use decisions, primarily on a monthly time step. Leaf fall also drives a simple model to estimate soil organic matter.

Precipitation is divided into daily storm events, expressed as a frequency distribution. The distribution of daily rainfall totals is fitted to a Gamma distribution for each month. The rainfall distribution, reflected by the coefficient of variation of rainfall per rain day is given for each month of the simulation period and may be adapted for (future) climate change scenarios. Daily precipitation drives infiltration, excess overland flow and soil erosion, and monthly precipitation, driving saturation levels in the soil. Infiltration excess overland flow is estimated from storm rainfall and soil moisture. Sediment transport is then estimated using a power law approach driven by erodibility, gradient and runoff discharge. Soil erodibility is derived from soil classification data, primarily texture (see Section 2.2.7). Local relief is defined as the standard deviation of elevation within a defined radius around each point (Section 2.2.1 and Figure 2). Accumulated runoff is derived from a biophysical model that combines the frequency of daily storm sizes with an assessment of runoff thresholds based on seasonal water deficit and vegetation growth. Estimates of sediment transport are based on infiltration excess overland flow discharge. In the PESERA model, sediment transport is interpreted as the mean sediment yield delivered to stream channels and includes no downstream routing within the channel network.

The role of vegetation and soil organic matter can modify the infiltration rates through changes in soil structure and/or the development over time of surface or near-surface crusting. Three models are coupled to provide the dynamics of these responses: (i) A vertical hydrological balance, which partitions precipitation between evapotranspiration, overland flow, subsurface flow and changes in soil moisture; (ii) a vegetation growth model, which budgets living biomass and organic matter subject to the constraints of land use and cultivation choices; and (iii) a soil model, which estimates the required hydrological variables from moisture, vegetation and seasonal rainfall history.

The PESERA model works with two phases: an equilibrium phase and a simulation phase. The equilibrium phase model is run first: it calculates long-term average values, using long-term input data on e.g., climate. The equilibrium phase model is calibrated using long-term average data (see Section 2.3). Then, these long-term output maps are used to initiate the simulation phase model. This model uses monthly climate data to run future scenarios (see Section 2.4).

PESERA outputs consist of monthly maps of: vegetation biomass (ton/ha), erosion (risk) (ton/ha/y) and soil organic matter content (ton/ha) for each simulation year. Within the SoilCare project the following improvements were made in the PESERA model: additional crop types (sugar beets, rice, fodder versus consumption maize) have been parameterised and calibrated for Europe; all crops were parameterised and calibrated for four main climate zones across Europe and biomass and SOM were calibrated for each land use/crop type; irrigation has been added as an option in the model; erodibility information for the Northern countries (Norway, Sweden and Iceland) has been updated to solve issues with existing Europe-wide data (see Section 2.2) and soil management options (i.e., SICS) have been defined, parameterised and calibrated (see Section 2.3).

#### *2.2. Input Data*

The required model input data and their sources are summarised in Table 1. All input maps have a spatial resolution of 500 m and projection ETRS 1989 LAEA (Lambert Azimuthal Equal Area). The area modelled is the EU-28, i.e., the current 27 EU countries plus UK. Basic details and the most important maps are given here; a full description and all input maps are given in Supplementary Material S1.

**Category Variable Number of Maps Data Source** Topography Local relief—st. dev. of elevation 1 ESDAC database (RECARE project) Climate *Equilibrium phase model (long-term current climate)* Mean monthly temperature 12 Based on E-OBS version 21.0e, at 0.1◦ spatial resolution and daily scale. [58]. 1981–2010 Mean monthly temperature range <sup>12</sup> Mean monthly rainfall 12 Mean monthly rainfall per rain day <sup>12</sup> Coefficient of variation of mean monthly rainfall per rain day 12 Mean monthly PET 12 Calculated from monthly Tmean and Trange following [59]. *Simulation phase model (climate scenarios)* Mean monthly temperature 12 \* n\_years E-OBS version 21.0e, at 0.1◦ spatial resolution and daily scale. [58]. 2018–2050; RCP4.5 MPI-ES-LR + CCLM4-8-17 Data: JRC EU High Resolution and Precipitation dataset: https: //data.jrc.ec.europa.eu/dataset/jrc-liscoast-10011 (accessed on 18 December 2020) [60] Mean monthly temperature range 12 \* n\_years Monthly rainfall 12 \* n\_years Maximum daily rainfall 12 \* n\_years Soil properties Erodibility class (sensitivity to erosion) <sup>1</sup> Classified RUSLE K-factor map by Panagos et al. [61] https://esdac.jrc.ec.europa.eu/content/soilerodibility-k-factor-high-resolution-dataset-europe (accessed on 1 July 2021) Crusting class (sensitivity to soil surface crusting) <sup>1</sup> Pedotransfer functions based on soil type and texture (ESDB) Scale depth (proxy for infiltration) 1 Based on Texture classes (ESDB) Soil water available to plants (0–300 mm) <sup>1</sup> Pedotransfer functions based on Available Water Content, Texture, Soil packing density and restriction of soil to bedrock; ESDB and SWAT-HWSD [62] for Iceland and Cyprus Soil water available to plants (300–1000 mm) <sup>1</sup> Effective soil water storage capacity <sup>1</sup> Land use & crop data Land use map 1 From Metronamica application, processed data from Eurostat and Corine Land Cover Crop map 1 Planting month (for crops only) 1 Grouped per climate region (see Table 2) Initial ground cover (%) 12 Following PESERA project manual estimations;

**Table 1.** Overview of PESERA input requirements.

adapted where needed


**Table 1.** *Cont.*

#### 2.2.1. Topography

One of the main variables in the model is local relief (Figure 2). It is estimated from the digital elevation model (DEM) as the standard deviation of elevation with a circle of 1.5 km (5 cell radius) diameter around each cell.

**Figure 2.** Local relief (standard deviation of elevation in a 1500 m radius) for Europe.

#### 2.2.2. Climate

Climate input data differs slightly between the equilibrium and simulation phase models. For the equilibrium phase model, E-OBS version 21.0e data, at 0.1º spatial resolution and daily scale was used. Daily data for the ensemble mean of mean temperature, minimum temperature, maximum temperature and rainfall were collected for 1981–2010, representing the reference period used to bias-correct climate scenarios. The monthly parameters shown in Table 1 were calculated from these values, after being interpolated to

a 500 m resolution. The data source is Cornes et al. [58]. Maps of the equilibrium climate input data are presented in Supplementary Material S1.

To minimise bias, climate scenarios at high resolution (0.1º), and already bias corrected with present-day climate (E-OBS) were used in the simulation phase model. The considered emission scenario was RCP4.5 (closer to the average of all emission scenarios). The selected GCM-RCM combination was MPI-ES-LR + CCLM4-8-17. This means that we used the MPI-ES-LR GCM, which has a median sensitivity to climate change [63] combined with the CCLM RCM, which appears to have less bias for temperature and rainfall in several European regions [64]. We used data from the JRC EU High Resolution and Precipitation dataset, which is already bias-corrected using E-OBS [60].

#### 2.2.3. Land Use and Crop Data

The land use and crop map (Figure 3) was made within the SoilCare project, based on Corine Land Cover 2018 (CLC2018) (https://land.copernicus.eu/pan-european/corineland-cover (accessed on 15 September 2021)) crop data from Eurostat (https://ec.europa. eu/eurostat (accessed on 15 September 2021)), and infrastructure (e.g., roads), zoning (e.g., protected natural areas, urban expansion plans) and crop suitability maps from Metronamica. Details on how these data were used to derive the SoilCare land use and crop map are given in [65].

#### 2.2.4. Crop Calendars: Planting Month, WUE and Cover

As crop calendars for the same crop may differ per climate region, we created four major agro-climatic regions in Europe, for which crop calendars were constructed for each crop. We did not use existing maps for cropping calendars, as they are either too coarse [66], not crop-specific [67], or represent related variables which are difficult to translate into planting month [68,69]. We decided instead to aggregate areas per climate region. The existing Köppen-Geiger system determines 19 different climate types in Europe [70]. These were aggregated into the six most representative classes, each occupying at least 5% of the SoilCare study area, and together occupying 92% of the total area; the remainder were assigned to the closest climate class. It should be noted that the division between climate regions is not sharp, and there are often climatic gradients. The six classes were then transformed into four classes with two further aggregations: (1) For cropping purposes, the dry climate regions are similar to the Mediterranean climate regions, so they were reclassified as the latter; and (2) polar climate is important in a large part of mountain regions, but agriculture is not practiced there, so for the model they were reclassified as subarctic climate. Figure 4 shows the climate zones as used in the modelling; they are similar to the environmental stratification of Europe proposed by Metzger et al. [71].

Finally, we aggregated existing crop calendar information for different countries in Europe for the four climate zones using the following datasets according to the dominant climate in the country, in decreasing order of preference:


When extended (>1 month) planting and harvesting dates were given, the latest planting and earliest harvesting date were chosen. The aggregation of calendars gave consistent planting and harvest dates for each region, with the Mediterranean region showing differences from the three other regions, either in earlier planting dates or shorter growing seasons. Cropping calendars were discussed with local partners from the SoilCare project and adapted according to their experience.

**Figure 3.** SoilCare land use and crop map (year: 2018). A GIS compatible version of this map is available in the Supplementary Materials.

Monthly ground cover (%) for each crop was derived mostly from the PESERA project manual, with some exceptions or additions:


**Figure 4.** Climate zones as used in SoilCare to vary crop calendars by agroclimatic zone.

These cover calendars were then adjusted to the crop calendars. In most cases, the cover calendars fit inside the planting and harvest dates. When they did not fit, they were adjusted to keep the same shape as the PESERA growth curves but fitting a shorter or longer interval as needed. When the crop calendars indicated planting or harvesting seasons longer than one month, the cover values of these seasons were extended by repeating the first or last month value (respectively). Table 2 shows the crop calendars per agroclimatic zone and crop, with the cover indicated as value. Monthly canopy cover for permanent crops were based on the PESERA project estimations [74] for Europe and are given in Table S1.

**Table 2.** Crop calendar and ground cover values (%) per crop and agroclimatic zone. Dark green cells indicate the start of the growing season (planting month), orange cells indicate the last month of the growing season.



**Table 2.** *Cont.*

Water use efficiency values were calculated for different crops based on the following sources:

	- Length of the growing stages: (Marjanovi´c-Jeromela et al., 2019)
	- Kc values: (Corlouer et al., 2019) (Figure 2 in their suppl. Material) [73]

WUE calendars, with crop- and growth stage specific WUE values, were also based on planting and harvest dates, and used the same method as that for cover calendars, including stretching or shortening curves to match planting and harvesting dates (Table S2).

#### 2.2.5. Rooting Depth and Surface Storage

Rooting depth was estimated based on three sources: the PESERA project manual [74]; estimates from FAO: http://www.fao.org/land-water/databases-and-software/ crop-information/maize/en/ (accessed on 15 October 2020). These estimates start at 30 cm root depth going to 100 cm at the end of the growing season. As PESERA estimates were lower, a conservative estimate was taken and cross-checked with the third source; the SWAT database, which also estimates slightly deeper (maximum) rooting depths. For initial surface storage (either 0, 5 or 10 mm) and reduction of surface storage (either 0 or 50%), the

PESERA project manual was followed. Values of rooting depth, initial surface storage and reduction of surface storage used in this study are given in Table S3.

#### 2.2.6. Soil Properties

Soil property data are used to calculate storage capacity and therefore the runoff threshold and affect plant growth through soil water availability. Six layers of soil data are required: (1) Erodibility, which is the sensitivity of the soil for erosion; (2) crusting, which is the sensitivity of the soil to surface crusting and affects the infiltration; (3) scale depth, which is a proxy for infiltration; (4) the effective soil water storage capacity; and soil water available to plants for depths 0–300 mm (5) and 300–1000 mm (6) respectively.

#### 2.2.7. Erodibility

The erodibility map has five classes. We used the RUSLE erodibility K-factor, as prepared by Panagos et al., [61], with stoniness effects incorporated, grouped into five classes (Table S4). As indicated earlier (Section 2.3), based on discussions with local partners, the erodibility map for Norway, Sweden and Iceland was adapted. Details of the method used can be found in Supplementary Material S1. Figure 5 shows the final erodibility map as used in this study.

**Figure 5.** Erodibility map as used in SoilCare. Note that bare rock and glacier areas (according to CLC2018) were excluded (grey colours).

#### 2.2.8. Crusting and Scale Depth Maps

The soil sensitivity to crusting index map was created using pedotransfer functions on texture, parent material and physical–chemical soil properties (Figure S1). The scale depth input map (Figure S2) was derived from soil texture classes (Table S5). Texture data were derived from the ESDB database.

#### 2.2.9. Soil Water Availability and Storage Maps

Soil water available to plants (both 0–30 and 30–100 cm) and effective soil water storage capacity maps were derived based on the instructions from the PESERA project [74] and using ESDB data. Available Water Content for topsoil and subsoil (AWC\_top and AWC\_sub) maps of ESDB were used as a starting point. Additional soil property data used in the pedotransfer functions include texture, packing density and restriction of soil depth by bedrock.

The effective soil water storage capacity was then calculated from the soil water available to plants in the top- and subsoil following the PESERA project instructions [74]. Estimations for Iceland and Cyprus, that are not included in the ESDB maps, were derived using the SWAT data in combination with the FAO Harmonized World Soil Database (HWSD), available at https://doi.pangaea.de/10.1594/PANGAEA.901309 (accessed on 15 August 2020). All three maps are shown in Supplementary Material S1 (Figures S3–S5).

#### *2.3. Model Calibration and Evaluation*

During the equilibrium phase, long-term average output of the model was calibrated for erosion estimates and soil organic matter. As it was not feasible to calibrate the model for all countries, calibration was carried out for four countries in various climate zones across Europe (Belgium Spain, Slovakia and Norway), and the Greek island of Crete. Tuning parameters for calibration were: (1) The biomass conversion factor used in the model to calculate gross primary production—affecting ground cover and thereby erosion; and (2) the decomposition factor used in the model to calculate soil organic matter from plant residues. Both parameters are specific for each crop and land use, but generic for all regions. For soil organic matter calibration, the LUCAS topsoil soil organic carbon point data was used: https://esdac.jrc.ec.europa.eu/projects/lucas (accessed on 1 December 2020), which was aggregated to crops and land covers per climate zone (Table S6). In addition, to crossvalidate and make use of the knowledge of the SoilCare local partners, both the spatial patterns and numerical (aggregated) results were shared with selected countries across Europe and their feedback was used for further fine-tuning. Preliminary results were sent to partners in Belgium, Germany, Greece, Spain, Italy, Norway, Poland and Romania. Based on their feedback:


The model output at EU scale was evaluated by comparing the ranges and spatial patterns of the equilibrium phase PESERA erosion and SOC output maps to existing maps reported in the literature (for SOC e.g., [38,75–77]; for erosion e.g., [34,78,79]; see Supplementary Material S2.

#### *2.4. Parameterisation of SICS*

The PESERA model was used to investigate four SICS [80], each representing a different category: soil improving crops, soil amendments, soil cultivation and compaction alleviation. Respectively they were:


These SICS were simulated individually, and in two combination scenarios, combining compaction reduction and minimum tillage with either cover crops or mulching (assuming that cover crops and mulching cannot be combined). The combination measures assumed that no additive effects would occur for each parameter, taking instead the most intensive effect of each individual measure on each parameter. The implementation of each measure in PESERA is described, in general terms, in Table 3. The model implementation of these measures was tested on a synthetic dataset, representative of climatic and crop conditions in the Oceanic climate regions of Europe. The differences between the application of the measure over the control conditions were compared with results taken from a survey of meta-analyses published in indexed journals, on soil erosion and soil organic matter; a detailed list of references is presented in Supplementary Material S3.



#### *2.5. Scenario Description*

Socio-economic scenarios were developed in the SoilCare project in multiple workshops and feedback rounds, including all relevant stakeholders, with the aim to explore plausible agricultural pathways for Europe and assessing their sustainability and profitability impacts. It is beyond the scope of the current study to describe this in detail. The scope of the scenarios and full descriptions can be found in [65]. Here, the scenarios are very briefly described, with emphasis on how the SICS were included in each scenario:


The actual location of which SICS were applied where within the scenarios on the map was randomly distributed within the arable land and perennial crops (i.e., olive groves, vineyards and fruit trees).

These four scenarios were run for the period 2020–2050 and erosion and SOC simulated maps were analysed for the year 2050, and compared to the baseline situation in 2020 with no SICS applied. Note that for erosion calculations, the climate (especially rainfall) of a specific year can affect results (e.g., a large rainfall event in a specific region may lead to high erosion estimates for that year and location, but this does not happen in other years). Therefore, to evaluate erosion output estimates, the average of 2020–2025 was used to represent 2020 and the average of 2045–2050 was taken to represent erosion in 2050.

#### **3. Results**

#### *3.1. Model Calibration Results*

#### 3.1.1. Baseline Long-Term Erosion

Figure 6 shows the calibrated model output for erosion (t/ha/y). These are the equilibrium phase simulation results, based on average long-term climate input data (see Table 1). Overall average erosion across the whole of Europe was simulated at 2.54 t/ha/y, with erosion in arable land estimated at 4.3 t/ha/y on average across Europe. The highest erosion rates were simulated in sugar beet and potato crops and lowest in spring cereals. For the permanent crops, olive groves showed high erosion rates, followed by fruit trees, with mixed and coniferous forest having the lowest erosion rates. This aligns well with estimates by Panagos et al. [34] of 2.46 t/ha/y for erosion prone land covers and 2.22 t/ha/y for all land covers. In line with expectations, the general spatial pattern shows relatively high erosion values in the zone from Northern France and Belgium, across Germany and Poland, known as the Loess Belt with soils susceptible to erosion. Moreover, the mountain areas (Alps, Norway, Apennines, Pyrenees) are visible as areas with high erosion. A third

zone of relatively high erosion is visible in the south of Spain and Italy, where low cover and erodible soils are present. The overall pattern across Europe compares well with estimates using RUSLE2015 [34] (see Figure S16), who also estimate relatively high erosion in the mountain areas (although Norway and Switzerland are not included in their calculations), in southern Spain and Italy and Northern UK. The RUSLE erosion map predicts less erosion in the Loess Belt than the PESERA estimates. Borrelli et al. [79] predict similar areas of relatively high erosion in southern Spain, Italy, across the Loess Belt, but less erosion in the mountain areas and Northern UK (Figure S18). Cerdan et al.'s [78] estimate of more erosion in the Loess Belt is comparable to the PESERA map. However, in the Cerdan et al. [78] map (Figure S17), more areas with relatively high erosion are visible, e.g., in Eastern Europe.

**Figure 6.** Simulated long-term average erosion rates across Europe using the PESERA model.

SoilCare partners' feedback on PESERA simulated erosion maps included for Spain that it seemed relatively low, compared to the national soil erosion map [84]. However, areas in the south show relatively high erosion in both maps. Belgium partners indicated that the relatively high erosion values for row crops like potato and sugar beet seemed valid, but that simulated erosion values for maize and vegetables, which have a wide spacing, were too low compared to their experience, especially when compared to simulated higher erosion in cereals. For Poland and Germany, simulated patterns of erosion were found to be plausible and matching e.g., the German national erosion map [85] with higher erosion in central Germany and very low to no erosion in the northern half of the country.

#### 3.1.2. Baseline Long-Term SOC Stocks

Figure 7 shows the PESERA simulated maps of SOC stocks for Europe based on long-term average climate conditions (equilibrium phase model output). Overall estimates amount to 50 Gt, which is in line with estimates by Aagaard Kristensen et al. [77] (60 Gt), but somewhat higher than estimates of Yigini and Panagos [38] (38 Gt). The Nordic countries (Sweden, Finland) as well as the higher altitude areas clearly show higher SOC stocks (except where soil depth is shallow), while lower SOC stocks are simulated in for example inland Spain, parts of Italy, France and Eastern Europe. This coincides with the patterns of other SOC estimates (see Figures S12–S15). However, the SOC stock map based on the soil profile analytical database for Europe (SPADE; [77]; Figure S15), shows a slightly different pattern with lower SOC stock estimates for Sweden and parts of Finland, where our estimates show high SOC stocks. However, the intermediate stocks are similarly simulated to occur in the wet north-western Iberian Peninsula, the Massif Central in France and relatively low SOC stocks in the Norwegian mountain areas. Highest SOC stocks were simulated for forests, followed by grassland and shrubs. This matches estimates by other studies [38,39,77], although our estimates for grassland (11 Gt) are somewhat higher than those by Yigini and Panagos [38] (6.7 Gt). SOC stocks for fruit trees, olive groves and vineyard were estimated at around 60 t/ha on average across Europe, while the average SOC stocks for arable land across Europe was estimated at 43 t/ha or 5 Gt, which is lower than e.g., Lugato et al. [39] and Yigini and Panagos' [38] estimates of 17.6 and 12.8 Gt respectively.

**Figure 7.** Simulated long-term average SOC stocks across Europe using the PESERA model.

Calibration results for SOC stocks for the three countries and Crete for which the model was calibrated are shown in Table 4. Overall, results were close to the observed data, derived from the LUCAS database. However, for some crops or land uses it was difficult to simulate good values across climate zones. For example, while SOC stocks for potato cultivation were well estimated for Spain, values were too high for Belgium and Slovakia, but too low for Crete. For maize, values were overestimated for Belgium and Slovakia, but underestimated for Spain.

**Table 4.** Calibrated (PESERA baseline long-term results) versus observed (LUCAS database) results for SOC (t/ha) for three countries and the Greek island of Crete in different climate zones. Note: X = crop does not occur. NA = not available.


SoilCare partners' feedback on the PESERA calibrated SOC results indicated that they were in line with national estimates or maps (e.g., Belgium, Poland, Norway, Germany). For example, the German partners provided a German national map with organic matter [86], on which the spatial patterns were similar as those simulated by PESERA.

#### 3.1.3. Calibration of the SICS

Figure 8 shows the simulated changes by the model for the individual SICS, compared with expected values from the literature. It should be noted that expected impacts on soil erosion were only found for cover crops, while the expected impacts on SOC were found for every SICS except compaction reduction; and that specific information for root crops and vegetables was less available than for cereals and permanent crops.

As can be seen, the simulated measures broadly followed what was expected from the literature in terms of erosion reduction and increase in SOC. When analysing per crop type, results for permanent crops tend not to be very good: no changes are simulated to erosion, because the baseline values were zero when using the test dataset; and changes to SOC are very limited. This indicates that the model is better adapted to simulate SOC changes for cereals than permanent crops. There is insufficient data to analyse model performance for root crops and vegetables.

In terms of impacts, PESERA simulates a small effect of compaction reduction when compared to other measures. For soil erosion control, mulching seems to have a limited effect in comparison to cover crops and minimum tillage; this results from the simulated wetter soil conditions when applying mulch, which increase biomass growth (and, indirectly, SOC) by limiting water stress, but also create the right conditions for more frequent runoff generation, counteracting beneficial soil protection effects. For SOC, mulching has a slightly larger benefit than cover crops or minimum tillage.

**Figure 8.** Impact of different SICS (expected; based on literature review, see Table S7, and simulated with PESERA on a test dataset) on different types of crops, for erosion (**top**) and soil organic matter (**bottom**).

As for the combination SICS, both tend to lead to higher increases of SOC when compared with the individual components. For soil erosion, the combinations involving cover crops led to larger reductions than the individual components. However, the combined mulch approach had a very limited impact on soil erosion, despite the erosion decrease expected when applying the individual components. As both individual measures increase soil moisture, the wetter soil conditions and increased runoff counteract the soil protection effects of the measures. In short, results suggest that the combined cover crop approach appears to have a better balance between SOC increase and erosion control, while the combined mulching approach has larger increases of SOC at the expense of the effects on erosion control.

#### *3.2. Simulated Results for SICS Scenarios*

Figure 9 shows the simulated difference in annual erosion between year 2050 and current (2020) for the four scenarios. Note that differences in erosion are affected by differences in climate (e.g., wet months in certain years) as well as by the application of SICS. Some areas show a consistent slight increase in erosion; these are mainly the steep mountain areas (Alps, Pyrenees) and areas that receive a lot of rainfall (e.g., Norwegian coastal zone), where SICS are not applied, as they are covered by e.g., pasture or shrubland. However, for example in the central European Loess Belt, southern Spain and eastern Europe, erosion was simulated to increase in the RttB scenario, while it decreases in the CS scenario due to application of SICS. Overall, across Europe and taking all land uses into account, erosion was simulated to increase slightly for the RttB scenario (+1.3% compared to 2020), while a decrease was simulated for the UP, LS and CS scenarios (75, 79 and 59% respectively, compared to the 2020 situation). When taking only the arable and orchard (fruit trees, olive groves and vineyards) areas into account, where SICS are applied, simulated erosion decreased to 43, 49 and 6.6% (compared to 2020) for the UP, LS and CS scenarios respectively, which is an average decrease of about 1.5 t/ha/y in both the UP and LS scenarios, and 2.6 t/ha/y in the CS scenario. So, especially for the CS scenario, a large decrease in erosion was simulated, which is in line with the large reductions that were parameterised for e.g., cover crops (Figure 8). Simulated erosion maps for RttB 2020 and the four scenarios for 2050 are given in Supplementary Material S4.

Figure 10 shows the simulated changes in SOC content for 2050, relative to the 2020 situation, for each of the four scenarios. All maps show both areas of decrease of SOC as well as areas of SOC increase. However, comparing between the scenarios, the results clearly show a more severe decrease in SOC for the RttB scenario, followed by UP, LS and CS scenarios. The average simulated SOC change across Europe, taking only the arable areas into account, was −23% for RttB, −4.5% for UP, −1.5% for LS and +22% for the CS scenario. This can also be seen in the maps (Figure 10): the CS scenario shows most increases in SOC content. For example, the arable areas in north-central Europe and north-central Spain that in the RttB show a strong decrease in SOC, turned into an increase in SOC in the CS scenario. This reflects the simulated SICS, where in CS all farmers apply a combination of minimum tillage, compaction reduction and either cover crops or mulching. In the UP scenario, all farmers apply only one type of SICS, while in the LS scenario, the application of SICS is mixed. Overall, it seems that, in terms of SOC content, the LS scenario leads to better results than the UP scenario, although local differences are likely greater in LS.

**Figure 9.** Simulated difference in erosion (t/ha/y) between 2020 and 2050 for each scenario.

**Figure 10.** Simulated change in SOC stock (%) between 2020 and 2050 for each scenario.

#### **4. Discussion**

Using the PESERA model, we simulated the effects of SICS on erosion and SOC stock changes across Europe, based on scenarios in which either no SICS were applied (RttB scenario; low sustainability level), or where a medium (UP scenario), high (CS scenario) or a mix of these three levels of sustainability was assumed (LS scenario). Comparing the effects of the simulated scenarios clearly shows that the application of a high level of sustainability, where all farmers apply a combination of SICS: minimum tillage, compaction reduction, and either cover crops (50%) or mulching (50%), results in highest and most widespread

erosion reduction (Figure 9) and a shift from a continuous reduction in SOC stocks to an increase in SOC stocks (Figure 10). In general, our results imply that erosion can be quite well prevented by application of SICS (Figure 9), but less pronounced effects are simulated for SOC change. This is in line with often reported findings of relatively quick effects of measures on runoff and erosion while changing SOC is a slower process [87].

The scenarios we simulated contain a mix of measures, so direct comparison with other estimates is difficult. Panagos et al. [16,41] calculated the effect of the C (cover management) and P (conservation practices) factors. Panagos et al. [16] found that conservation tillage reduced the C-factor (and thus, indirectly erosion, if all other factors remain the same) by 17%, application of crop residues reduced the C-factor by 1.2% and cover crops by 1.3%. Note that these numbers are affected by the (relatively small) area where these practices were found to be applied and that large differences between countries were found [16]. Panagos et al. [41] estimated the P-factor (conservation practices), when including contouring, stone walls and grass margins, to be 0.9702 across Europe, meaning that erosion would be reduced by 3% (if all other factors remain the same). This factor has a wide spatial variation, being lower (i.e., more effective erosion protection) in for example Portugal, Spain and Belgium, due to a high number of stone walls and grass margins [41]. These findings are somewhat different than our results, as they do not include mulching. In our results, cover crops were estimated to reduce erosion significantly (also due to the calibration, see Figure 8). In a review study based on a large database of plot-scale erosion and runoff, including the effects of SWCTs (Soil and Water Conservation Techniques), Maetens et al. [18] found that overall, application of SWCTs reduced the exceedance probability for a soil loss tolerance of 5 t/ha/y and 12 t/ha/y by 14 and 12% respectively. The individual measures ranked in the order (more to less effective) of geotextiles, buffer strips, mulching, contour bunds, cover crops, conservation tillage and strip cropping [18]. They concluded that crop and vegetation management (mulching, cover crops) and mechanical measures (terraces, contour bunds) are more effective than soil management techniques (reduced tillage). While our study did not include mechanical measures in the scenarios, our results are in line with this as the CS scenario, where cover crops and mulching is always applied (in combination with minimum tillage and compaction reduction) is clearly more effective in reducing erosion than the UP scenario, where only half of the farmers applies mulching or cover crops. However, it should be noted that, for soil erosion, even a low intervention scenario (UP) can decrease erosion below 1 t/ha/y (Figure S20), which can be considered as a threshold for sustainability [88]. There is some variability between climate regions, and within them, between regions with different topography and soil types. Nevertheless, these results indicate that the UP scenario might be good enough for most agricultural crops in Europe; and that special attention, and stronger intervention measures, could focus on remaining crop types (pulses, root crops, etc.,) and on areas with higher erosion rates. The results from this work could be used as a first approach to define priority areas for different levels of intervention across Europe.

Similar as for erosion, also a direct comparison with other studies regarding the simulated changes in SOC stocks for our scenarios is difficult. Lugato et al. [49] simulated the effect of six management practices scenarios on possible carbon sequestration, including spatially explicit maps across Europe. They found that, besides conversion of arable land to grassland which showed the highest SOC sequestration rates, ley cropping systems and cover crops results in higher SOC sequestration than straw incorporation and reduced tillage, which is in line with our results. Aertsen et al. [47] investigated the effect of agroforestry, hedges along field boundaries, cover crops and no/low tillage on carbon sequestration for the EU27, concluding that agroforestry has the highest potential, and no spatial maps of Europe were presented. Bellassen et al. [48] did not include no-tillage practices, as they only redistribute SOC instead of sequestering it. They also state that cover crops have a substantial potential for carbon sequestration, but that the large-scale potential of other practices such as hedges and crop residues is probably limited. Lessmann et al. [89] combined global meta-analytical results with spatially explicit data on current management practices and potential areas for implementation of measures at a global scale and found that organic matter inputs led to highest mean SOC changes, followed by crop residue incorporation, reduced tillage and increased crop diversity [89].

While in general terms the simulated values and spatial patterns are in line with other studies, local experiments and observations might deviate. This is a difficulty in any upscaling to large (continental) scales. Factors that play a role include assumptions in the model (e.g., biomass and humus conversion factors are crop-specific, but not adaptable per region), lack of (spatially explicit) input data (for example the difficulty of deriving a reliable erodibility map) and scarcity of (observed; i.e., non-modelled) calibration and validation data across Europe [79], but see [18,77]. Therefore, the absolute values should be taken with caution, but a qualitative and comparative analysis over time and across Europe can be insightful.

In this study, we focussed on erosion, as one of the most important processes of land degradation [35] and SOC changes, as one of the most widely used and important indicators of soil health [32,33]. While these are important indicators, many other indicators and processes play a role in a healthy functioning soil [2,5]. However, simulation of all these functions together is almost impossible, especially at a large (e.g., EU) scale. A few studies are beginning to attempt this. For example, the Soil Navigator decision support system was developed to assess and optimise various soil functions [90] on farm scale, incorporating soil management strategies. This was applied to monitor multiple soil functions at 94 sites across 13 European countries [91]. Vrebos et al. [92] analysed and mapped four soil functions on agricultural lands across the EU. Borelli et al. [79] evaluated soil degradation in Europe, including both erosion and soil carbon fluxes using the WaTEM/SEDEM modelling approach, but did not include the effect of soil and water conservation measures.

Potential additions to the modelling approach that we simulated in this study, would be to include additional indicators, such as biomass growth and effects on yields. While this is possible in the current PESERA model, preliminary results showed some difficulties. However, coupling of PESERA with more sophisticated biomass and yield models such as QUEFTS [93] is feasible and ongoing. This would also allow to evaluate the effects of (changes in) nutrient supply to the crops within the SICS. Another improvement within the PESERA model is to enable the parameterisation/calibration in the SOC calculations in the model (e.g., the decomposition rates) to be both spatially and crop specific (they are at the moment only crop specific), for example by including a spatial map of annual decomposition rates in Europe [46]. In addition to this, land use change as well as climate change can be included in the modelling framework. Finally, while in this study we only evaluated the effects of SICS on environmental indictors, in a really comprehensive analysis and modelling framework, also socio-economic factors and indicators should be taken into account, such as economic profitability and adoption of measures. In the SoilCare project, an important finding was that although the CS scenario leads to highest impacts, the gross margin of SICS uptake under this scenario is negative in many NUTS-2 regions [24,65]. Moreover, note that the spatial allocation of SICS application (e.g., where which SICS was applied within the scenarios) was randomly allocated. Interestingly, overall, results of the UP scenario (medium sustainability level with one SICS applied in all arable lands) were close to those of the LS scenario (a mix of low (no measures), medium (one measure) and high (multiple measures) sustainability levels). However, the spatial variability in LS will be higher, meaning that areas with high erosion and low increase (or decrease) of SOC will be offset by other areas with high erosion reduction and increase in SOC stocks. To avoid this and reach land degradation neutrality (LDN, [94]), careful planning is required and in terms of the scenarios simulated here, regarding the allocation of measures there is room for improvement in the scenarios, for example to base the allocation of certain SICS in areas that need them most and/or are most suitable [95].

#### **5. Conclusions**

In this study we simulated the effects of Soil Improving Cropping Systems (SICS) on SOC stocks and erosion on EU scale using the PESERA model. Four scenarios with varying levels and combinations of SICS were simulated for the time period 2020–2050. We can conclude that, for both SOC stocks, as an indicator for soil health, and erosion, as indicator for land degradation, the scenario with the highest level of SICS, i.e., application of minimum tillage and compaction alleviation in combination with either mulch or cover crops, clearly decreases erosion levels substantially across Europe as well as turning a decreasing trend of SOC stocks (when no SICS are applied) into an increase in SOC stocks, on average across Europe. Scenarios with medium level of SICS application as well as a scenario that implemented a mix of no SICS, medium level and high level SICS throughout Europe showed a decrease in erosion, while SOC stocks remained at the current level.

Future improvements for this modelling study would include to add climate change and dynamic land use. Furthermore, SICS were now randomly allocated in the arable lands; further scenarios including a more targeted spatial allocation of the levels of SICS would be interesting to conduct.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/land11060943/s1. Supplementary material S1: Input data maps for PESERA (Figures S1–S11; Tables S1–S5). S2: EU scale maps of SOC stocks (Figures S12–S15) and erosion (Figures S16–S18) used for model evaluation and SOC calibration data (Table S6). S3: Literature used to compile effect of measures (Table S7); and S4: Simulated erosion maps for the baseline (2020; Figure S19), RttB, UP LS and CS 2050 (Figure S20).

**Author Contributions:** Conceptualisation, J.E.M.B., J.P.N., L.F. and H.v.D.; methodology, J.E.M.B., J.P.N., L.F., H.v.D. and R.V.; software, J.E.M.B., J.P.N., H.v.D. and R.V.; validation, J.E.M.B. and J.P.N.; formal analysis, J.E.M.B. and J.P.N.; investigation, J.E.M.B., J.P.N., L.F. and H.v.D.; resources, J.E.M.B., J.P.N., L.F. and H.v.D.; data curation, J.E.M.B. and J.P.N.; writing—original draft preparation, J.E.M.B. and J.P.N.; writing—review and editing, J.E.M.B., J.P.N., L.F. and H.v.D.; visualisation, J.E.M.B. and J.P.N.; project administration, J.E.M.B., J.P.N., L.F. and H.v.D.; funding acquisition, L.F. and H.v.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 677407 (SoilCare project).

**Institutional Review Board Statement:** Not applicable in this study.

**Informed Consent Statement:** Not applicable in this study.

**Data Availability Statement:** The data presented in this study are available in the Figures in the paper and in the Supplementary Material S1–S4.

**Acknowledgments:** We acknowledge the E-OBS dataset from the EU-FP6 project UERRA (https:// www.uerra.eu (accessed on 18 December 2020)) and the Copernicus Climate Change Service, and the data providers in the ECA&D project (https://www.ecad.eu (accessed on 18 December 2020)). We also acknowledge the feedback on preliminary model results by SoilCare partners, especially J. Stolte and R.J. Barneveld (NIBIO, Norway), M Tits and A. Elsen (Bodemkundige Dienst van Belgie, Belgium), P. Mayer-Gruner, E. Kandeler (University of Hohemheim, Germany), J. Cuevas, V. Pinillos (University of Almeria, Spain), M. Frac, J. Lipiec, B. Usowic (Institute of Agrophysics of the Polish Academy of Sciences, Poland); A. Berti, I. Piccoli (University of Padova, Italy); I. Calciu, O. Vizitiu (National Research and Development Institute for Soil Science, Agrochemistry and Environmental Protection, Romania), D. Alexakis, I. Tsanis (Technical University of Crete, Greece).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **A New Framework to Assess Sustainability of Soil Improving Cropping Systems in Europe**

**Abdallah Alaoui 1,\*, Moritz Hallama 2, Roger Bär 3, Ioanna Panagea 4, Felicitas Bachmann 3, Carola Pekrun 5, Luuk Fleskens 6, Ellen Kandeler <sup>2</sup> and Rudi Hessel <sup>7</sup>**


**Abstract:** Assessing agricultural sustainability is one of the most challenging tasks related to expertise and support methodologies because it entails multidisciplinary aspects and builds on cultural and value-based elements. Thus, agricultural sustainability should be considered a social concept, reliable enough to support decision makers and policy development in a broad context. The aim of this manuscript was to develop a methodology for the assessment of the sustainability of soil improving cropping systems (*SICS*) in Europe. For this purpose, a decision tree based on weights (%) was chosen because it allows more flexibility. The methodology was tested with data from the SoilCare Horizon 2020 study site in Germany for the assessment of the impact of the integration of cover crops into the crop rotation. The effect on the environmental indicators was slightly positive, but most assessed properties did not change over the short course of the experiment. Farmers reported that the increase in workload was outweighed by a reputation gain for using cover crops. The incorporation of cover crops reduced slightly the profitability, due to the costs for seeds and establishment of cover crops. The proposed assessment methodology provides a comprehensive summary to assess the agricultural sustainability of *SICS*.

**Keywords:** sustainability framework; overall sustainability; costs and benefits; cover crops

## **1. Introduction**

Assessing agricultural sustainability is one of the most complex exercises related to appraisal methodologies because it entails not only multidisciplinary aspects (environmental, economic and social dimensions), but also builds on cultural and value-based elements [1]. Thus, agricultural sustainability should be considered as a social concept that can be modified in response to the requirements of society as a whole and the individuals constituting this society [2,3].

According to the current definitions policy-oriented sustainability assessment is a methodology that can help decision- and policymakers decide what actions they should or should not take to make society more sustainable [4]. For this purpose, sustainability

**Citation:** Alaoui, A.; Hallama, M.; Bär, R.; Panagea, I.; Bachmann, F.; Pekrun, C.; Fleskens, L.; Kandeler, E.; Hessel, R. A New Framework to Assess Sustainability of Soil Improving Cropping Systems in Europe. *Land* **2022**, *11*, 729. https://doi.org/10.3390/ land11050729

Academic Editor: Evan Kane

Received: 14 April 2022 Accepted: 10 May 2022 Published: 12 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

assessment practitioners have developed a large number of tools [5]. Finding the appropriate assessment instrument is critical to match theory with practice, and to have successful outcomes in improving sustainability. More specifically, although many methods exist for monitoring and evaluating the environmental dimension of agricultural management practices, no single method has been widely accepted for assessing it, perhaps due to the complexity and variability of agricultural systems [6]. Though the meanings and uses of the term sustainability remain diverse, it is now widely accepted that sustainability is the path to balancing social, economic, and environmental needs [7–9].

There is broad scientific agreement on the fact that sustainable agriculture is defined as the management and the use of the agricultural ecosystem in a way that allows reaching economic (e.g., income growth or economic stability), social (e.g., equity or the cover of basic needs), and ecological objectives (e.g., ecosystem protection or natural resources regeneration) [7,10]. These objectives need to be continuously evaluated with scientific criteria, acknowledging uncertainty and safety margins.

Many frameworks with various combinations of indicator sets aimed at describing farming and cropping systems exist, from simple ones to complex multi-dimensional assessment tools [11–13]. The choice of the indicators depends on the objective of the study. In many studies, indicators are chosen to characterise the sustainability of the system or the intensity of management and practices (i.e., land-use intensity) [14–16]. However, the collection of the data needed to implement such frameworks is tedious and time consuming, and thus simple and reliable indicators, based on data that are reasonably easy to obtain, are required [5].

A review study related to sustainable agriculture revealed that the social dimension is the most difficult to assess in a quantifiable way when compared to the environmental and economic dimensions due to its inherently more subjective nature [17,18]. Research looking into the social sustainability of farming systems deals with issues and indicators related to (subjective) well-being and quality of life of the farming population, working conditions (workload, working time), gender equality, on-farm and off-farm incomes, access to services (education, advisory services), social relations (family, community), social security, finding work meaningful, life satisfaction, physical and mental health, etc. [18–20]. Hence, socio-cultural acceptability is a prerequisite for the adoption of new agricultural practices.

In their review paper, Alaoui et al. [5] selected frameworks based on the following criteria: (1) are validated through a peer-review process, (2) consider a farm-level assessment, (3) cover universal agricultural sectors including livestock and arable farms, (4) include the three dimensions of sustainability, and (5) are suitable both for Europe and countries worldwide. Based on the selected criteria, the following frameworks were identified: RISE (Response-Inducing Sustainability Evaluation [21]), MASC (Multi-attribute Assessment of Sustainability of Cropping Systems [22]), LADA (Land Degradation Assessment in Drylands [3]), SMART (Sustainability Monitoring and Assessment RouTine, [23]), SAFA (Sustainability Assessment of Food and Agriculture systems [24]) and PG (Public Goods [25]).

The EU Horizon 2020 project SoilCare, aimed "to assess the potential of soil-improving cropping systems (*SICS*), to identify and test these *SICS* to determine their impacts on profitability and sustainability in Europe". This required an assessment framework based on an evaluation of environmental, sociocultural, and economic dimensions of crop production. The methodology needed to allow flexibility; it needed to be applicable to all study sites (SS) across Europe to allow comparison and upscaling and at the same time to be flexible enough to consider site-specific circumstances.

Taking into account the above considerations, none of the reviewed frameworks was suitable for SoilCare because they did not include the indicators needed to evaluate *SICS* and/or did not provide results to evaluate the key terms of SoilCare (such as sustainability, profitability, soil quality) in combination.

The main aim of this research was to develop a comprehensive methodology for assessing the overall sustainability of the farm with special attention to the benefits, drawbacks, profitability, and soil quality of the *SICS* as compared to the *control*-conventional system. To set up a tool for the assessment of the overall sustainability, we chose a decision tree based on weights (%). This is because it allows simple aggregation to assess the three dimensions of sustainability and provides flexibility [22].

In this manuscript, we provide the general concept of the assessment tool developed to calculate sustainability of the *SICS* under consideration. We provide information on the indicators, their weighting factors, their threshold values, and their scores. An application with data from the German SoilCare study site is provided to explain how the tool is used for conservation agricultural techniques and serves as a first critical evaluation to document lessons learned. In this study, we assess the sustainability of the farm/field where the *SICS* is implemented.

#### **2. Materials and Methods**

#### *2.1. Assessment Tool*

For the evaluation of the overall sustainability of a farm and to facilitate the assessment of the performance of cropping systems three dimensions of sustainability were considered, i.e., environmental, sociocultural and economic. A decision tree was chosen for the aggregation. It breaks the sustainability assessment decisional issue down into simpler units that comprise quantitative as well as qualitative elementary criteria to rate cropping systems. Such aggregation is needed as the data for the three dimensions include various kinds of quantitative and qualitative data, obtained in different ways, including monitoring and questionnaires [22].

Within the decision tree, weights (%) were assigned to adjust the relative importance of the different indicators used within the three dimensions of sustainability. These weighting factor values were established from expert knowledge based on the literature review and can be modified to fit specific conditions and decision makers [5].

In the SoilCare project, the study sites selected to test the sustainability impact of *SICS* were grouped into 4 key topics to improve sustainability, namely, compaction, soilimproving crops, fertilization/amendments, and soil cultivation. The experiments implemented in the SoilCare project were short-term since the project was a 5-year project. To assess the sustainability of a given farm using the tool developed here, input data is needed. The tool calculates sustainability by assigning a higher score to the key topic considered in comparison to the others. This was the reason why we developed a new tool for the assessment of sustainability.

For the assessment of the sustainability of a farm or a field, we have selected plots with the *SICS* and plots without (*controls*) that best characterise the farm or field under consideration. The assessment was carried out by comparing the *SICS* plot with the *control* plot. Figure 1 provides an overview of the three dimensions considered and the related properties. For future use, the users can adapt the weighing to their specific case. This flexibility would help improve the assessment tool for various purposes.

#### 2.1.1. Environmental Dimension

• Monitoring variables

To assess the sustainability of a farm, in situ measurements of the variables were carried out. For this purpose, a monitoring plan was established to harmonize the monitoring including instructions on the treatment replication, randomization, and sampling in which each experiment within a field/location is composed of 3 blocks (corresponding to 3 replicates). Each block contains two experimental units or plots where sampling is carried out for composite or undisturbed samples [26].

• Selection of the indicators and weights

Based on a literature review [5] and considering the *SICS*-related key topics in the SoilCare project, a list of variables for the evaluation of the environmental dimension of the implemented systems was established. Each key topic is defined by a set of indicators with high weight, e.g., soil compaction is assessed by bulk density and penetration resistance with a value of 0.20, and by infiltration capacity and aggregate stability with weight values of 0.15 and 0.10, respectively (refer to the grey boxes in Table 1).

**Figure 1.** Structure of the aggregation of sustainability dimensions and assessment units.


**Table 1.** Weighing factors attributed to variables as related to the four key topics (soil cultivation, soil fertilization, soil-improving crops, and soil compaction).

The selection of variables and their assigned weights was based on the literature review [5] and is presented in Table 1. For more explanation, refer to File S1.

The assessment tool was designed to assess the change in the environmental dimension resulting from the implementation of the *SICS* compared to the *control* cropping system. Prior to inputting data into the assessment tool, the quantitative change of each variable as measured/estimated in the field is transformed into a qualitative score: positive change, no change, or negative change. For more details, refer to the "Metadata sheet" in File S1 using a statistically based approach.

In this tool, an additional option suggests the appropriate methods to be used for the evaluation of the variables listed in Table 1. The aim was to harmonize the methods across all study sites (refer to File S1, input datasheet). Based on the type of method used, the accuracy of the evaluation is evaluated.

• Statistical analysis

To quantify the difference between the variables of soil with the *SICS* and the ones of the *control*, a statistical approach was used. Mixed-effects models were used to estimate if statistically significant differences exist between the *SICS* and related *control* treatments. Mixed effect models were chosen as they allow a larger variety of designs and implementations [27] enabling a better identification and interpretation of interactions and repeated measurements. The statistical data analysis was performed using-R-Studio, R version 3.6.1 [28]. Differences between treatments or dates were analyzed using the full factorial statement "Treatment x Date", or the factor "Treatment" for variables with repeated measurements and only once measured variables, respectively. The experimental design structure effect (block, whole plot, main-plot, etc.) was introduced in all models as a random effect, using the statement "1|structure" (in the German case study this was 1|Block). The model's optimum fixed structure was selected for the best fit attaining the lowest value of Akaike's Information Criterion (AIC) using a maximum likelihood function (ML). A visual inspection was performed of the residuals' Q-Q plots and the normalised residuals' plots against the fitted values. The final models were fitted using REML estimation. Estimated marginal means by factors were computed by the least square method and contrasted by the Tukey group comparison method (*p* < 0.05).

The comparison between the variables of *SICS* and the ones of the *control* used for scoring the impact of *SICS* was scored by attributing a value of "1" for the statistically significant positive change (PC) *(SICS* is better than the *control*), "0" for no statistically significant change (NoC) and "−1" for statistically significant negative change (NC) (*control* is better than *SICS*).

#### 2.1.2. Sociocultural Dimension

In contrast to the environmental indicators, most factors that determine the sociocultural acceptability of a *SICS* cannot be easily measured or quantified, which is due to their inherently subjective nature. Therefore, a qualitative approach was applied, and a short questionnaire (summarized in Table 2) was used to grasp the land users' assessment of the tested *SICS* in terms of three key topics: changes in workload, perceived risks, and influence on farmers' reputation.

Three requirements for *SICS* to be socially/culturally acceptable were identified on the basis of a literature review.

• Requirement 1 (Workload): *SICS* should not result in a considerable increase in workload, especially in periods where labour demand is already high.

In agriculture, working hours are generally long, considerably longer than in other professions. Therefore, it is not surprising that farmers are sensitive to an increase in workload, especially when it occurs during periods in which labour input is already high, e.g., in spring (field preparation and sowing) or summer respective autumn (harvesting) [29]. Additionally, an increase in workload not only means long working days, but also leads to higher production costs [29,30].

• Requirement 2 (Risk): In the perception of farmers, a *SICS* should not be a (too) risky practice.

A survey from north-eastern Germany showed that associated risks are among the main drivers when decisions are made to adopt new conservation measures [29]. Trujillo-Barrera et al. [31] concluded from their in-depth interviews with Dutch hog farmers that perceived risk is a barrier to the adoption of sustainable production practices.

• Requirement 3 (Reputation): Applying a *SICS* should not impair the farmer's reputation.

Much evidence exists [29,32–34], that farmers base their decision to adopt or reject conservation practices not exclusively on economic, agronomic, and ecological grounds. To be adopted, practices need to be compatible with land owners' values, and perceptions as to what makes a good farmer, or an aesthetic agricultural landscape (e.g., keeping fields nice and tidy).

Within each requirement, different questions were asked. The possible combinations of the responses and their output scores are reported in the "Metadata sheet" of File S1.

To calculate the final score of the sociocultural dimension, a specific weight was attributed to each requirement listed above (Table 2).

**Table 2.** Variables considered in the sociocultural dimension based on a questionnaire completed by the land user and their weighing factors.


Given the fact that the two topics of risk perception and workload increase are both crucial for the adoption or rejection of new farming practices, they both have double weight compared to the topic of farmer reputation. In addition, the effect on a farmer's reputation is much more difficult to grasp and verify. Therefore, it was deemed appropriate to give this topic less weight in the assessment.

Study site researchers conducted the interviews at the end of the growing season with those farmers involved in the *SICS* trials. The questionnaire was kept as simple as possible, in order to avoid any limitations related to the implementation of the questionnaire, such as an adequately trained audit team, long duration for the training and implementation.

#### 2.1.3. Assessment of the Economic Dimension

The economic dimension was assessed by evaluating the costs and benefits of the farm using a spreadsheet formatted questionnaire to ensure ease of use. This questionnaire, adapted from [35], contained the different types of costs, such as investment costs, maintenance costs, production costs, and benefits related to both the *control* and the *SICS* fields. Details on costs and benefits should refer to the same area/unit that can be defined in the Overview worksheet (refer to File S2 for more details). A summary of the costs and benefits is directly calculated and provided at the end of the questionnaire to allow the comparison between the *control* and the *SICS*. The details of each cost category are described below.

Investment costs: The assessor should list all one-off investment costs, structured according to the activities and inventorying labour, agricultural inputs, construction material, wood, earth, and other costs.

An activity refers to a defined task needed to establish the *SICS* and may consist of multiple inputs:


Maintenance costs: List of maintenance/recurrent activities and their associated costs for the *SICS* and the *control* cropping. It contains the same cost categories as listed for the investment costs above.

Production costs: List of changes pertaining to activities or inputs related to activities that have changed as a consequence of introducing a *SICS* (or a crop). Recurrent costs related to the technology itself should be recorded under maintenance costs.

Benefits: The benefits are considered at the farm level and consequently are defined as "on-site benefits". They can include: (i) Products harvested (cash and food crops, timber, fuelwood, fruits and nuts, animal fodder, etc.), (ii) Grazing/browsing, (iii) Recreation/tourism, (iv) Subsidies (e.g., for agri-environmental measures), and (v) Protection against natural hazards.

Calculation: The cost–benefit score represents the difference between the weighted (see below) relative change in benefits and in costs. A positive score means an improved cost–benefit ratio, a negative score means an impaired cost–benefit ratio. The score was calculated as follows:

$$\text{Cost} - \text{Benefit score} = \left(\Delta \text{RC}\_{\text{benefit}} \times \text{Benefit weight}\right) - \left(\Delta \text{RC}\_{\text{costs}} \times \text{Cost weight}\right) \tag{1}$$

With ΔRC*benefits* = Relative change in benefits, calculated as follows:

<sup>Δ</sup>RC*benefits* <sup>=</sup> <sup>∑</sup> *CostsSICS* ∑ *CostsControl* (2)

ΔRC*costs* = Relative change in costs, calculated as follows:

$$
\Delta \text{RC}\_{\text{Coys}} = \frac{\sum \text{Benefits}\_{SICS}}{\sum \text{Benefits}\_{\text{Control}}} \tag{3}
$$

The type of costs for both *SICS* and *Control* are: Investment costs + Maintenance costs + production costs.

The benefits of both *SICS* and *Control* include all benefits listed in File S2 and are calculated as follows: Products harvested (cash and food crops, timber, fuelwood, fruits and nuts, animal fodder, etc., +Grazing/browsing + Recreation/tourism + Subsidies (e.g., for agri-environmental measures) + Protection against natural hazards.

In both cases (i.e., change in benefits and change in costs) the relative change has been capped to +/−100%. At the cost end, the positive extreme is theoretically solid as it means that costs involved with the *control* system can be reduced to 0 and that 0 means a perfect (+1) score. A doubling of the cost (and anything above) is regarded as the most negative outcome (−1). At the benefit side, a score of −1 is attributed to any decrease.

The *Benefit weight* and *Cost weight* account for the amplitude of changes in benefits and cost. The *Benefit weight* represents the ratio between the absolute difference in benefits as compared to the absolute difference in costs. The *Cost weight* is the counterpart of the *Benefit weight* and represents the ratio between the absolute difference in costs as compared to the absolute difference in benefits. The weights are calculated as follows:

$$Benefit\ weight = \frac{|Benefit\_{SICS} - Benefit\_{Controll}|}{|Cost\_{SICS} - Costs\_{Controll}| + |Benefit\_{SICS} - Benefit\_{Controll}|} \tag{4}$$

$$Cost\ weight = 1 - benefit\ weight \tag{5}$$

This allows us to appropriately consider cases with minimal absolute changes at the cost side but large changes at the benefit side, or vice versa. For instance, the doubling of cost from EUR 10 to EUR 20 will have double the weight (0.66) compared to a doubling of benefits from EUR 5 to EUR 10 (0.33).

#### *2.2. The Case Study*

In the SoilCare project, panel meetings including scientists, farmers and other stakeholders identified a number of threats to soil quality and fertility and proposed management techniques for the mitigation of these threats [36]. At the study site in Germany, the stakeholder panel suggested focusing on conservation agriculture and to investigate, among other techniques, specifically cover crop management. Therefore, a field experiment was set up at a research farm in Tachenhausen, Germany (48.649800◦ N, 9.387500◦ E, 330 m a.s.l.). In the present study, the assessment methodology was applied to the comparison between cover cropping and bare fallow treatments.

The soil is heavy, and loess derived, with a very fine sandy loam texture. The soil profile is characterized as Cambisol (IUSS Working Group WRB, 2015) with four horizons and with a ploughing pan at 40 cm. The climate is temperate with a mean annual temperature of 8.8 ◦C and 809.3 mm precipitation (monitoring weather station Tachenhausen HfWU, 150 m from the site, 1961–1990). The field has a history of conventional agriculture, with a crop rotation consisting mainly of cereals and sugar beet and winter oilseed rape as alternate break crops. The crop rotation for the experiment was spring barley (2018)–cover crop mixture/bare fallow–silage maize (2019)–spring barley (2020). The main crop for the 2019 cropping season was Zea mays (var. Figaro) sown on 6 May 2019 with 10 plants m−2 and harvested on 17 September 2019.

The experimental layout was a randomized complete block design with eightreplicates with plots of 2.4 × 3 m. The treatments included bare fallow and cover cropped plots. Originally, the field experiment was set up as a full factorial experiment, including also two herbicide treatments. However, as no significant interaction between the cover crop and herbicide treatments could be detected, the measurements could be averaged over the two cover crop treatments. For establishment of the field experiment, a commercial cover crop mixture consisting of 55% Vicia sativa, 20% Trifolium alexandrinum, 16% Phacelia tanacetifolia and 9% Helianthus annus was sown at 25 kg ha−<sup>1</sup> in rows of 20 cm in the beginning of August. The field was tilled with a rotary harrow in a depth of 10 cm shortly before sowing the cover crops in a regime of non-inversion tillage. Mineral fertilizer was applied in 2018 at the rates of 90 kg ha−<sup>1</sup> N, 17.5 kg ha−<sup>1</sup> P, 53.1 kg ha−<sup>1</sup> K, 8.1 kg ha−<sup>1</sup> Mg and 20 kg ha−<sup>1</sup> S. The maize in 2019 was not given any fertilizer. The following spring barley received mineral N-fertilizer at a rate of 89 kg ha−<sup>1</sup> on 17 April 2020. Herbicides were applied as necessary.

The sampling and measurement of the indicators of the assessment methodology was carried out in spring after the cover crop in 2018–2019 following a monitoring plan with standardized methods for biological, physical and chemical properties of soils [26]. In the case study of Germany, the economic assessment was made possible by taking the values from publicly available tables of agricultural economics. For the calculation of the cost–benefit analysis, a sequence similar to the field experiment consisting of cereal-cover crop-silage-maize-cereal was used, but with winter wheat instead of spring barley. The sociocultural dimension was assessed by conducting semi-structured interviews based on the abovementioned questionnaires with five different farmers and a consultant of the public extension service of the region.

#### **3. Results**

#### *3.1. Environmental Dimension*

The assessment methodology was applied at the study site in Germany to compare the *SICS* integrating cover crops with the *control* treatment with bare fallow over winter. The environmental performance of the *SICS*, measured as the response of selected soil quality indicators, showed mixed results. Some indicators improved with cover crops (i.e., bulk density and soil cover) or showed a positive trend (number of earthworms) (Table 3). On the contrary, water-stable aggregates and infiltration were higher in the fallow plots, while weeds, tended to be lower than in the cover crop treatments. Mineral nitrogen tended to be lower under cover crops. Most of the other soil quality indicators showed no variation. This slight improvement indicates the positive effect of certain cover crop species on soil quality, especially on soil structure expressed by the reduced bulk density/increase in total porosity [37]. The resulting figures are presented in Supplementary Material File S3, the error bars represent the SE of the model.


**Table 3.** Impact of cover crops on the variables at least two years after the implementation.

Concerning the key topic addressed here, namely soil improving crops, there was a slight increase with an impact index of 0.10 (Table 4).

**Table 4.** Outcomes of the assessment of the environmental dimension with regard to the key topics.


#### *3.2. Economic Dimension*

In order to assess the economic dimension, the benefits of *SICS* were calculated in relation to the costs for the crop sequence of three years. Since the cereal straw was left on the field, the benefit is based on the pure grain yields, respective silage maize yield multiplied by the average market price in the respective year.

When comparing the benefits with the costs for both *control* and *SICS*, there is a loss in both cases (File S2), but less loss for the *control* than for the *SICS*.

The benefit of the *SICS* is higher than that of the *control* (Table 5). The cause of the loss is due to the higher costs for cover crop seeds and sowing that outweigh the slight increase in benefits (yield).

**Table 5.** Impact index of the economic dimension of *SICS* as compared to *control* considered in the CSS of Germany.


#### *3.3. Sociocultural Dimension*

The assessment of the sociocultural dimension shows slight positive impact due to the improvement of farmer reputation, although the moderate increase in the workload due to the short time window left after harvest to perform sowing reduces the perceived overall benefit for the farmers (Table 6). The problem with the workload at harvest time could be mitigated by using the technique of harvest–sowing, but this is only possible in some combinations of main and cover crop.

**Table 6.** Assessment of the changes of the sociocultural dimension with cover crops as soil improving cropping system. The Impact index was calculated using the responses of practitioners in structured interviews at the study region in Germany.


#### *3.4. Overall Sustainability*

The field study in Germany provides an example of the application of the tool (Table 7). In this case, the environmental and the sociocultural dimension improved slightly under cover cropping (*SICS*) compared to bare fallow (*control*). The economic dimension showed a negative scoring, because of a slight increase in costs. Further assessment in the coming years is necessary to confirm these results.

**Table 7.** Synthesis of the impact of applied *SICS*.


#### **4. Discussion and Recommendations**

#### *4.1. Outcomes of the CSS of Germany*

In order to evaluate the applicability of the assessment tool, it was applied to the dataset resulting from a field experiment at the German study site comparing cover crops and bare fallow as agricultural practices in a common crop rotation with cereals and silage maize. The proposed set of soil quality indicators was used to assess the environmental dimension. The effects on the economic dimension were evaluated by assessing the costs and benefits of the two systems, while the sociocultural dimension was studied using structured interviews with farmers. Generally, statistically observable effects of the *SICS* treatment on the measured soil properties in the field experiment were limited to a few indicators. Reports of positive effects of cover cropping on main crop yield and soil quality are abundant in the literature, but results vary [38,39]. The costs of the *SICS* with the inclusion of cover crops were slightly higher than in the conventional treatment, resulting in a slightly negative score of the economic dimension. The farmers that were interviewed for the assessment of the sociocultural dimension had consistently a positive opinion of cover crops, but also acknowledged management difficulties and a certain dependence on a favourable climate for cover crop establishment and performance.

Regarding the environmental dimension, the positive effect of cover crops on soil cover in spring was significant. Especially under conservation tillage management, cover crop litter constitutes a protective layer on the soil surface and provides important benefits for the agroecosystem, such as erosion protection, reduced evaporation and habitat for soil fauna [40]. Cover cropping had also a positive effect on bulk density, which is related to pore connectivity. As compaction is an increasingly acknowledged soil threat [41], cover crops provide an interesting opportunity to increase porosity, especially in systems with no or decreased mechanical soil loosening [42]. A previous study showed that cover crops reduced soil penetration resistance or compaction by 0–29% (average, 5%). They improve wet aggregate stability by 0–95% (average, 16%) and cumulative infiltration by 0–190% (average, 43%) [43]. In our case, soil biological properties tended to improve, with earthworm numbers showing a positive trend with cover crops although not significant (data not shown here), as well as the potential extracellular activity of β-glucosidase, an enzyme involved in the breakdown of cellulose (not shown). These results clearly showed that the micro-habitat provides more substrate and energy for microbial life under cover crops [38,44]. While most measured soil chemical attributes were not changed.

Despite these positive changes with cover crops, the *SICS* seemed to have also undesired effects on some soil variables: aggregate stability decreased significantly while bulk density decreases (or increase in total porosity). This last observation can be attributed to the dominance of the structural porosity created by earthworms. The unexpected decrease in aggregate stability indicates that positive effects of cover crops on different aspects of soil structure might require time and multiple growing cycles to develop [45]. More on the management side, weed pressure tended to be higher in the cover cropped plots compared to the bare fallow due probably to missing herbicide application, although weed suppression is another expected benefit from cover cropping. Maximising the cover crop biomass and an optimized termination and residue management can improve the weed-suppression capacity of cover crops [46]. The obtained scoring of the environmental dimension of the assessment methodology provided a quite accurate resume of the slight improvement of soil quality with cover crops compared to the bare fallow *control*, but with an uneven response of the different soil quality indicators.

Similarly, the farmers' rather positive opinion about cover crops was reflected by the improvement of the sociocultural dimension. The modest increase in workload was greatly offset by the improved reputation. This underscores the importance of prestige for decision making for practitioners [47], especially since the farmers in the region are increasingly worried about their public image, some of them even mentioned feeling attacked by media. The farmers also acknowledged potential positive benefits of the cover crops being especially interesting when considering the necessary adaption to climate change [48].

The slight negative scoring of the economic dimension matches the reality, as the adoption of *SICS* and other sustainable farming techniques frequently implies higher production and maintenance costs which are not covered economically due to the inability of the market to integrate externalities into pricing [49]. Potential benefits of cover crops were not included in the economic assessment, such as SOC increase, erosion reduction, N input by leguminous cover crops or an increased biodiversity. Nor could external benefits for society be included, such as reduced sediment runoff, C sequestration and positive effects on water quality or landscape, among others [50]. The complexity of management techniques based on (agro-)ecosystem functions means they frequently require a substantial amount of experience to yield satisfying results. Even worse, although management can alleviate many reasons for the underperformance of *SICS*, in some cases significant trade-offs between environmental performance and productivity remain and call for a paradigm shift [51]. Until then, in absence of effective market mechanisms, potential losses can be only compensated by increasing the subsidies for environmentally friendly farming practices.

The overall scoring of the SoilCare assessment methodology of cover cropping is therefore possibly partially biased by an overly negative score of the economic dimension, but seems to provide an acceptable resume of the effects of the adoption of this *SICS* for the sustainability of the system. When evaluating the assessment methodology in workshops at the German study site, the stakeholders provided a heterogeneous rating of the assessment

methodology and made some suggestions that could likely improve the applicability and power of the tool.

Regarding the measured soil properties, some farmers suggested to include methods of visual assessment of soil structure, e.g., as in a shovel test, as this method is easy to perform for practitioners and it gives relevant information for practitioners [52]. Participants with a more academic background suggested the measurement of greenhouse gases to cover this important aspect of sustainability. Another possibility would be to integrate methods to assess the soil microbial community into the tool. Creamer et al. [53] explain in detail how soil life could be integrated in the concept of multifunctionality of agricultural ecosystems. It is clear that the selection of adequate methods for judging soil microbiological properties is context dependent. The authors give in their article three different possible contexts where soil microbiological properties could have an additional value: Mechanistically understanding of multifunctionality, optimising sustainable land management and soil quality monitoring over time. Further investigations are needed to include all the above aspects in the here presented tool.

#### *4.2. Strengths and Challenges of the Assessment Tool*

This paper describes a new assessment tool to assess the overall sustainability of soil-improving cropping systems illustrated by an example from Germany. An overview of results from SoilCare study sites obtained with the tool under various conditions within SoilCare is provided in [36]. Therefore, from the outset, our intention was to develop a practical and flexible tool to assess the overall sustainability that can be adapted for other purposes and contexts.

In our assessment methodology, we included environmental/soil quality, economic and socio-cultural aspects in order to take into account all factors that are relevant for the success of *SICS*. Nevertheless, it should be realised that our assessment remains a simplification of reality. To be able to develop an assessment methodology for *SICS* that could be used in SoilCare, some assumptions were necessary, and some limitations exist:


specific, especially in 2018 occurred droughts at several study-sites, resulting in a drastic decrease in yield [41]. Moreover, all the years had high, sometimes recordbreaking, temperatures.

• Considering the existing distortions of the market and the large dependence of European agriculture on subsidies, it could be debated whether the weight given to the economic dimension in the calculation of the overall sustainability score is biased by ideology instead of a true interest in the well-being of future generations. This societal benefit effect can be captured by an extension of the indicators and extensive data collection. The semi-quantitative nature of the sustainability index would allow for an extension considering the direction of the impact of *SICS* (positive, no change, negative) on different ecosystem services for society even if valuation is not possible.

#### **5. Conclusions**

The aim of this paper was to establish a tool for the assessment of the sustainability of the *SICS* at the farm/field scale. For this purpose, a decision tree based on weights (%) was chosen because it allows simple aggregation to assess the three dimensions of sustainability, namely, environmental, economic and sociocultural, and provides flexibility. The decision tree allowed us to set up a comprehensive and standardized methodology that could be further improved and used for different purposes. The methodology was tested with data from the SoilCare Horizon 2020 study site of Tachenhausen, Germany, for the assessment of the effect of integration of cover crops into the crop rotation. The effect on the environmental indicators was slightly positive, but most assessed properties did not change during the short time of implementation (two crop seasons). Regarding the social dimension, farmers reported that the increase in workload was outweighed by an improved reputation for using cover crops. Regarding the economic dimension, the incorporation of cover crops reduced slightly the profitability, due to the costs for seeds and establishment of cover crops, which were greater than the increased income from higher yields. Further development and refinement by considering various pedo-climatic and land management conditions, as well as long-term assessments, are needed to strengthen the predictions.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/land11050729/s1, File S1: Excel Tool for the evaluation of the sustainability, File S2: Excel Tool to assess economic dimension, File S3: Results of the statistical analysis–study site in Tachenhausen, Germany.

**Author Contributions:** Conceptualization: A.A., R.H. and M.H.; methodology: A.A., R.B., F.B., I.P., R.H., M.H. and L.F.; data acquisition: C.P., E.K., M.H. and A.A.; data analysis: I.P., A.A., F.B. and C.P.; writing—review—editing: A.A., M.H., R.B., F.B., I.P., R.H., L.F. and E.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 677407 (SoilCare project). RH also received funding from Dutch Ministry of LNV, via Kennis Basis programma 34, project KB-34-008-005.

**Institutional Review Board Statement:** Not applicable in this study.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in the Supplementary Material.

**Acknowledgments:** We thank the stakeholders in the SoilCare study sites. Without their collaboration this research would not have been possible.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Do Agricultural Advisory Services in Europe Have the Capacity to Support the Transition to Healthy Soils?**

**Julie Ingram 1,\*, Jane Mills 1, Jasmine E. Black 1, Charlotte-Anne Chivers 1, José A. Aznar-Sánchez 2, Annemie Elsen 3, Magdalena Frac 4, Belén López-Felices 2, Paula Mayer-Gruner 5, Kamilla Skaalsveen 6, Jannes Stolte <sup>6</sup> and Mia Tits <sup>3</sup>**


**Abstract:** The need to provide appropriate information, technical advice and facilitation to support farmers in transitioning towards healthy soils is increasingly clear, and the role of the Agricultural Advisory Services (AAS) in this is critical. However, the transformation of AAS (plurality, commercialisation, fragmentation, decentralisation) brings new challenges for delivering advice to support soil health management. This paper asks: To what extent do agricultural advisory services have the capacity to support the transition to healthy soils across Europe? Using the 'best fit' framework, analytical characteristics of the AAS relevant to the research question (governance structures, management, organisational and individual capacities) were identified. Analysis of 18 semi-structured expert interviews across 6 case study countries in Europe, selected to represent a range of contexts, was undertaken. Capacities to provide soil health management (SHM) advice are constrained by funding arrangements, limited adviser training and professional development, adviser motivations and professional cultures, all determined by institutional conditions. This has resulted in a narrowing down of access and content of soil advice and a reduced capacity to support the transition in farming to healthy soils. The extent to which emerging policy and market drivers incentivise enhanced capacities in AAS is an important area for future research.

**Keywords:** agricultural advisory services; soil health; governance; agricultural advisers; sustainable soil management; soil policy; advice

#### **1. Introduction**

Soil health has emerged as a priority for high level and national policy makers and for agricultural communities. This is linked to the recognition of the multiple functions that soils fulfil and the soil degradation processes closely linked to agriculture: erosion, organic carbon decline, soil biodiversity decline, compaction, contamination, salinisation and acidification [1,2]. Indeed, soil health is seen as "a key solution for our big challenges" in the newly launched European Union (EU) Soil Strategy, which builds on the European Green Deal and the Farm to Fork Strategy [3,4]. For agricultural soils, soil health and managing soil sustainably are regarded as central to food system transition pathways

**Citation:** Ingram, J.; Mills, J.; Black, J.E.; Chivers, C.-A.; Aznar-Sánchez, J.A.; Elsen, A.; Frac, M.; López-Felices, B.; Mayer-Gruner, P.; Skaalsveen, K.; et al. Do Agricultural Advisory Services in Europe Have the Capacity to Support the Transition to Healthy Soils? *Land* **2022**, *11*, 599. https://doi.org/ 10.3390/land11050599

Academic Editor: Amrakh I. Mamedov

Received: 9 March 2022 Accepted: 12 April 2022 Published: 19 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

such as agroecology and regenerative agriculture, managing carbon for climate change mitigation and adaptation and mitigating pollution for human wellbeing.

However, the soil governance landscape (formal and informal institutions related to soil-related decision-making processes) continues to be highly fragmented. It is characterised by multi-level and multi-actor decision making, with no single body responsible at the EU or national levels [5] and could be described as a network mode of governance [6]. A number of public and private mechanisms are applied that influence agricultural soil management decisions (directly or indirectly), reflecting the multiple functions (provisioning, filtering of nutrients, carbon storage, flood mitigation) and private and public good that soils provide [7]. These include public cross-sectoral policy instruments (regulatory and voluntary) at the EU, national and regional levels; market-led (food assurance schemes in the supply chain); and measures which are led by the farming industry and non-governmental organisations (NGO) (voluntary initiatives, partnerships and networks).

This emphasis on soil health and its complex governance arena brings new challenges both for land managers and those that support them. The need to provide appropriate information, technical advice and facilitation to support farmers in transitioning towards sustainable soil management [8] has been identified by a number of researchers and policy makers at the international, European and national levels [8–13].

Agricultural Advisory Services (AAS) have always constituted an important part of farmer decision making with respect to soil management [14,15]. However, the increasing complexities of managing different soil functions and soil health at the farm scale [9] places new demands on these services.

Soil health has been defined as the capacity of soil to function as a vital living system [16]; however, the concept, and how it is operationalised, is still evolving [17]. Consequently, there are many understandings of what constitutes good soil health management (SHM). There are multiple practices embodied in the soil health concept, such as the use of cover crops and residues and reduced tillage; however, there is no single message or set of advice that is relevant to all contexts. Emerging interest in soil health indicators, soil biological processes and soil carbon dynamics [18] and new farming approaches (e.g., regenerative agriculture, agroecology), requires increasingly specialist knowledge and understanding (metrics, sampling techniques and analysis, interpretation) [19–21] beyond the traditional territory of soil fertility and agronomy. Meeting famers' knowledge needs, building their capacity and facilitating shared learning for SHM presents new imperatives for advisers [22]. These challenges exist against a backdrop of a changing farming population operating in a volatile, competitive marketplace negotiating multiple drivers in the agri-food system.

AAS have themselves been in transition, with privatisation and decentralisation occurring to different extents across Europe over the past 30 years [23]. The diversity of actors, intermediaries and organisations from the private (The private farm advisory sector includes profit and non-profit enterprises. Prager et al. [24] distinguishes 'private' as the status of an organisation and 'commercial' as the activities carried out by the organisation (e.g., offering advisory services for a fee)) and public sectors and NGOs engaged in some way in offering advice that influences soil management has grown. In particular, there has been an increase in the number of private advisers (These include: commercial agronomists offering services as part of farm input sales; farm management consultants; independent advisers or technicians within the supply-chain, sector or industry body or employed by farmer-owned groups) [25] and those with commercial links to farmers [26,27]. There is debate about the impact of such diversity on governance with respect to the integration and fragmentation of advice, competition and cooperation and how on access to quality advice [24,28,29]. Arguments about the advantages and disadvantages of privatisation have also been well rehearsed [30–32]. The potentially negative impact of commercialisation on public goods advice [33] and the limited investment in updating environmental knowledge for advisers has been highlighted [34]. The powerful effect of new economic actors, such as those in the supply chain, on environmental objectives has also been demonstrated [26,27,35] and noted specifically for soil management [10,36,37]. Private sector providers who support production goals can promote practices detrimental to soil health (e.g., multiple field operations with heavy machinery, a reliance on inorganic fertilizer, poor budgeting of organic inputs, harvesting in unsuitable conditions) [38]. Meanwhile, resources for public sector advice to farmers on the mitigation of soil degradation processes have also been shown to be inadequate [39].

Although there has been a requirement for all EU member states (since 2007) to establish a Farm Advisory System (FAS) (according to FAS, Regulation (EC) N◦ 73/2009) to support farmers in meeting cross-compliance requirements, including soil management though Good Agricultural and Environmental Conditions (GAEC), the singular advisory focus on compliance has been to the detriment of other soil functions and soil health outcomes [40].

The role of the individual adviser is also shifting, demanding greater professionalisation in increasingly specialised sectors, technical expertise (subject-matter knowledge), facilitation skills and awareness of a number of policy instruments, innovations, industry demands, certifications and environmental objectives. In such an environment, the assumption is that advisers will pursue different knowledge and strengthen and broaden their suite of professional practices to suit the 'new farming paradigm' [41,42]. At the same time advisers need to stay abreast of the farming community's growing informal soil knowledge networks, [43,44] and the different ways they negotiate their own microAKIS [27]. However, a body of evidence has been accumulating [10,14] suggesting a lack of specialist soil technical support and expertise in the advisory community, a poor understanding of the impact and externalities of their advice for soil, as well as varying motivations. Although studies show that farmers are deferring to advisers for their soil testing, largely in arable sectors [45,46], the lack of meaningful guidance for advisers regarding interpretation of these tests for soil health, especially for soil organic matter, and for specific farm conditions and management, is a concern [22,47]. There are a number of examples of effective soil advice across Europe [39]; however, it is clear that there are variable skill sets [11].

These insights raise questions about the capacity of advisory organisations and the constituent advisers for supporting SHM. This paper asks: To what extent do agricultural advisory services have the capacity to support the transition to healthy soils across Europe?

This addresses a recognised research gap, since understanding how the economic resources and strategies of advisory organisations determine the content of advice has received little (particularly for soil health) attention [26,30]. Equally, although soil literacy and societal engagement are central to the EU Soil Strategy and the implementation of the Mission for Soil and the European Soil Partnership, little has been done to understand the level of knowledge and expertise about soil health management in AAS.

This question is addressed using an analytical framework which positions AAS within the wider Agricultural Knowledge and Innovation System (AKIS), in which AAS are a subsystem. The framework implies that a range of organisations and stakeholders are involved in agricultural innovations along agricultural value chains, as well as agricultural research, agricultural extension and agricultural education [48]. Work conducted in the EU-funded SoilCare project underpins this analysis.

#### **2. Concepts and Framework**

AAS can be defined as sets of organisations that support and facilitate people engaged in agricultural production to solve problems and to obtain information, skills and technologies by enabling farmers to co-produce farm-level solutions by establishing service relationships with advisers [30,32,49]. AAS comprise traditional advice providers (chambers of agriculture, public bodies, research institutes), farmer-based organisations (FBOs) (unions, associations, cooperatives), non-governmental organisations (NGOs), independent consultants as well as advisers working in upstream or downstream industries, supply chains and high-tech sectors. However, these distinct categories do not fully capture the different arrangements and the new actors and roles emerging [23,26,27]. The term 'pluralistic' is used to describe the diversity of institutional options in providing and financing AAS [50]. AAS are characterised by a range of approaches, including one-to-one advice, facilitated interactive group approaches to foster peer-to-peer learning and mass dissemination.

We have adapted the 'best fit' (Birner's [49] framework was proposed for identifying modes of providing and financing advisory services that 'best fit' the specific conditions and development priorities of specific countries) framework developed by Birner, Davis, Pender, Nkonya, Anandajayasekeram, Ekboir, Mbabu, Spielman, Horna and Benin [50] to analyse the capacity of AAS for supporting SHM (Figure 1). Due to the multiple interacting dimensions of the AKIS and the AAS, it is difficult to collect data to capture the full complexity and interdependence of the system [51]. This framework provides a means of disentangling the different dimensions within the system. Here, we use selected key analytical categories in the framework relevant to the research questions. We define SHM as 'where management maintains or enhances (and does not impair) the capacity of soil to function as a vital living system, and to provide supporting, provisioning, regulating cultural services'. SHM is underpinned by the following management principles identified in the SoilCare project: integrate crop rotation, maintain continuous soil cover, build organic matter, minimise soil disturbance, prevent soil compaction, manage water for soil, use soil-friendly weed/pest control and consider landscape-scale management. These were derived by scientific review [52] and experimentation [53], as documented in this Special Issue, and have proven soil health benefits [54,55].

The focus of this study is on Characteristics of the system of agricultural advisory services (C in blue in Figure 1) and the implications for SHM. Specifically, according to Birner, Davis, Pender, Nkonya, Anandajayasekeram, Ekboir, Mbabu, Spielman, Horna and Benin [50], we include:


These were translated into characteristics relevant to SHM and framed the data collection. We understand that *Governance* (Usually defined as the systems of institutional

rules, policies and processes which govern how roles and responsibilities are delegated, managed and coordinated) *structures* enable and constrain organisational activities, in particular, the institutional options available for financing and the extent of coordination, fragmentation and integration [56]. Regarding the AAS M*management, organisational and individual capacities* to deliver SHM advice, organisational capacity is usually defined in terms of the capacity to perform effectively or to fulfil a goal (e.g., as the set of processes, management practices or attributes that assist an organisation in fulfilling its mission [57]). This capacity also affects the individual. Previous work has shown, for example, that the back-office support such as training and the organisation's knowledge management impacts advisers' capabilities [58].

Regarding individuals, AAS studies recognise that certain individual competencies are necessary for organisations and their advisers to operate effectively, where competence is: to have sufficient knowledge and skills that enable a person to act in a wide variety of situations and the ability to perform something efficiently and effectively (i.e., successfully) [59,60]. These skills include technical skills, which relate to specialist understanding (knowledge, expertise), and process skills, which are soft skills and refer to the collective skills necessary for effective performance of the individual and their organisation. Technical skills with respect to SHM are foregrounded in this study; however, we recognise that the 'soft' skills of facilitators, intermediaries and network builders are important [61].

This infrastructural approach to assessing AAS, which focuses on the presence and interaction of actors and the infrastructures that govern the behaviour of actors [62] also draws on selected criteria that Prager, Creaney and Lorenzo-Arribas [30] identified for evaluating a functional advisory system.

This paper focuses its analysis on C to address the research question, and because these characteristics can be influenced directly by policy makers. However, data for Frame conditions (A), Characteristics of farmers/land managers and their knowledge needs (B), Characteristics in Performance (D), Evidence (E) and Impact (F) (Figure 1) were also collected and analysed. Frame conditions (A) are important contextual factors in shaping the AAS, particularly as these have implications for SHM, and these include: Policy environment; Capacity of potential service providers; Farming systems and socioeconomic conditions. Equally, we acknowledge that Farmers' knowledge needs (B) (which we inserted into the framework) are important for assessing the adviser's role. It is not the intention here to follow an impact chain approach to analyse the performance of agricultural advisory services (D,E,F). Assessing (Performance, D) and the quality of advice is challenging, since its measure is the outcome for the farmer and there are multiple factors that affect this [23,50,63].

#### **3. Methods**

Countries in Europe are highly diversified in terms of their AAS and AKIS, reflecting the structure of agriculture, farming systems, soils and productivity [64] and the extent to which AKIS are embedded in national institutions, laws and cultures [27]. Six case studies (drawn from country partners in the SoilCare project) were selected to represent a range of AAS approaches and contexts: Norway, Belgium (Flanders) (Belgium (Flanders was the case study for Belgium) as Wallonia operates a different system), Spain, the UK (England) (As the UK's four countries have different political structures and agricultural policies, the focus was on England), Germany and Poland. The selection was based on three broad criteria: firstly, AAS organisations (to ensure the dominant ones were represented); secondly, characterisation of the AKIS, according to strength of national influence and level of integration/fragmentation (based on the the PROAKIS project); and thirdly, to include a range of biogeographical and pedoclimatic zones, as already determined in the SoilCare case study selection process [65].

This selection was informed by detailed AKIS descriptions for each of these countries from a range of sources, including PROAKIS [66] and i2connect [67] project country reports [23,30,32,68–70], and previous reviews for soil [39]. The dominant AAS are represented in the case studies: Spain FBO; England (Private); Germany (Public/Private/FBO); Poland (Public); Norway (Private); and Flanders (Public) according to previous studies [23]. The pedoclimatic zones represented are: Atlantic Central (Flanders, Germany, England), Nemoral/Boreal (Norway), Continental (Poland) and Mediterranean South (Spain).

Semi-structured interviews (2–4) were conducted in each case study (a total of 18). Selection procedures and interviews were carried out in each case study by project partners using standardised guidance and protocols. Respondents were selected were: representatives of decision/policy makers at national and regional level or of AAS organisations who were knowledgeable about SHM. As this was a very limited pool of experts, a purposive sampling strategy was employed. Table 1 lists the respondents and their roles and affiliation and shows the range of AAS organisations represented.

**Table 1.** The case study respondents and their roles and affiliations.


The analytical categories (Figure 1) were translated into interview questions and topics as shown in Table 2. Interviews were recoded, transcribed and translated into English, then analysed thematically using Nvivo 12. The coding structure followed the analytical categories of the interview but was extended where other themes emerged inductively. In total, 18 interviews provided in-depth analysis of AAS capacities for supporting SHM. A list of abbreviations is provided. A full interview schedule is provided as Supplementary Materials.


**Table 2.** Analytical categories translated into interview topics: example questions.

#### **Management, organisational and individual capacities**


#### **4. Results**

Where quotes are provided, the code refers to the notation in Table 1. A summary of the results is provided in Table 3.


**Table 3.** Summary of AAS characteristics for case studies.


**Table 3.** *Cont.*

#### *4.1. Framing Conditions*

The case studies represent a range of biophysical, political, socio-economic and farming contexts which determine the nature of the agri-food system, the distribution and intensity of production systems, the risk to the soil under agricultural management and public/private support. For example, in Norway, limited areas of arable land coupled with heavy rainfall, constrain timely tillage operations and has led to a national policy prioritising the reduction in the area under autumn ploughing in regions susceptible to soil erosion. In contrast, in Spain, low rainfall areas present challenges for farmers dealing with droughty soils. In Germany (Brandenburg), weather extremes mean water storage capacity, and water-saving cultivation methods are a priority. In both Flanders and Spain, specialised horticultural production systems put pressure on farmers' businesses and, consequently, the soil, while elsewhere, extensive crops such as 'soil friendly' wheat have lower profit margins. In England, arable farmers have expanded with more powerful machinery often implemented by contractors who do not always take account of soil conditions.

With respect to the political context, in Norway and England, there are, to some extent, shared goals between the government and the farming industry (farmer unions and cooperatives) which allow farms to deliver on a range of policies including food production and environmental goals (water, biodiversity, climate, soil). In Spain, a dual system results where intensive horticulture is mainly driven by commercial interests while political interests of soil conservation are more present in extensive agriculture. In Germany, public district and regional offices identified a lack of direction about soil management from the federal government.

From a socio-economic perspective, labour is expensive in Norway, and this affects farm profitability; in Flanders, seasonal land leases hamper any investment in SHM, while strong manure regulations have implications for managing organic matter. Land leasing in Poland leads to exhausted soils, while in Germany (Brandenburg), the large number of cooperatives are well managed by expert agricultural scientists, although the farms themselves are struggling with liquidity. In family-based, non-horticultural farms in Spain, traditional knowledge about and habits concerning soil management continue to be passed on through generations. These variable contextual factors act as framing conditions for AAS for SHM.

#### *4.2. Governance Structures*

#### 4.2.1. Governance Arrangements

The six case-study countries have each evolved distinct AAS (and AKIS) in response to a range of framing conditions, with a different mix of public, private and farmer-based organisations (FBOs); non-governmental organisations (NGOs); and research institutes delivering advice that influences and impacts soil management in each (summarised in Box 1 from analysis of interviewee responses). The hybrid and dynamic nature of partnering and funding arrangements is notable across all the case studies. Consequently, there is a diversity and complexity of influencers on farmers' decisions about soil management.

The role of the public support varies across case studies. For most countries, the regional or district agricultural offices have been re-oriented away from technical advice towards administration of subsidies and regulations, where the emphasis is on cross-compliance (GAEC) or supporting scheme applications. For example, in Baden-Württemberg (Germany), the soil service from district administration indirectly controls handling of soils according to the law. Other advice which directly or indirectly impacts SHM tends to be offered through a number of channels; it is often contracted out by the government to private companies, independent companies, FBOs and NGOs and focuses on aspects such as nutrient management and cross-compliance. Only the Soil Service of Belgium, an independent non-profit organisation, is specifically dedicated to soil management. Notably, a public face-to-face advisory service for soil is largely absent or very limited across the case studies. FBOs are significant in Spain, where they are linked with technical soil advice in the production of high-value crops; in Germany, farmer associations are strong, and in Norway, which has a large independent membership organisation (NRL), soil advice is demand-led.

The emergence and influence of the private sector is notable across the case studies. This encompasses a range of advisers working for input suppliers or independently. These advisers play an important on-farm role, where they support day-to-day farming operations. The powerful advisory role of private companies linked to input sales was characterised in some countries as the 'commodification of knowledge', as one Polish respondent (PL3) remarked, "*advice becomes more and more important, and knowledge becomes a commodity that can be bought or sold*". The role of FBOs and the private sector has implications for SHM advice as they respond to farmers' production-oriented needs rather than public goods per se.

#### **Box 1.** A summary of the main AAS governance arrangements relevant to SHM in each case study.

**Norway's** pluralistic advisory system comprises FBOs, public and commercial services. The Norwegian Advisory Service (Norsk landbruksrådgiving NLR) is a decentralised service which provides independent farm, phone and group advice through membership (most large cereal producers are members). NLR also receives subsidies for the organisation's regional and local units to support public good objectives such as soil management and widening access. Other advice comes from advisers working for input companies, independent private consultants and agricultural business cooperatives (input and buyers). Governmental bodies, especially at the local and county levels, have a role in supporting fertilizer plans and subsidies, but there is limited governmental support and responsibility for advisory services.

In **Flanders**, there are different forms of AAS and different sources of funding (regional and provincial funds, farmers' contributions, industry). Public support is still important through funding of regionally embedded Research Stations (RS) which focus on physical and biological soil aspects and act as practical advisory centres, with group dissemination events linking research to farmers and advisors. The Soil Service of Belgium, an independent research and advisory institution, is the main RS for soil. Advisory services subsidised by the government include the CVBB (Coordination centre for education and guidance to sustainable fertilisation), with a focus on nutrient management, now replaced by B3W (Coaching service for a better soil and water quality), with a focus on improvement of soil and water quality. Provincial and regional offices manage administrative issues. FBOs (unions and associations, cooperatives), private consulting companies, Dutch advisors and upstream and downstream industries are a main AAS component and their attention is mainly on crop nutrition and fertilizers.

In **Spain**, there are no public services that specifically provide soil advice on farm, although Agricultural and Fisheries Research and Training Centres hold field events for crop nutrition advice, and regional agricultural offices offer technical advice and training to farmers but are mainly concerned with managing subsidies. Agricultural unions, universities, RDP and operational groups are also involved in advice initiatives. The dominant type of AAS in Spain is the FBOs, the OPAs and the Agro-Food Cooperatives, which are linked to high-value crops and hire their own agricultural technicians, supply companies, certification bodies and have large and established structures. They also have innovation and development centres and provide training to farmers. Farmers with extensive low profit systems (cereals and woody crops) have less access to technical soil advice at farm level.

In **England**, the AAS is diversified and highly fragmented following privatisation. For onfarm advice, agronomists/consultants (independent or commercial) tend to dominate. Where there are commercial interests, historically the emphasis has been on fertilizer recommendations; however, consultants also provide agri-environmental services. Levy bodies (independent/FBOs) offer knowledge exchange for sector production support. Public supported advice has been linked to agri-environment schemes and catchment-based initiatives (soil management to manage diffuse pollution), where cross compliance was a key objective, delivered in partnership with government agencies, water companies and contractors through on-farm and group advice. The government is prioritising supporting public goods (with an emphasis on soil) post-Brexit. A range of NGOs have become increasingly important in facilitating initiatives relevant to all soil health functions.

In **Germany**, there is a heterogeneous and decentralised governance structure where the Federal Government and the 16 Länder take an active role. Due to limited funds, most state services are becoming privatised. These are: (i) the state agricultural offices (free public extension providers) that engage in rural development and regulatory issues, and they also attend to local soil issues; (ii) the chambers of agriculture that offer free and charged advice, education and training; (iii) private consulting and advisory companies offer fee-based advice on specialised topics such as production and business management; (iv) numerous upstream and downstream companies also contribute, as do a broad range of actors who belong to FBOs (boundaries between private organisations and FBO are often fluid). Privatised advisory companies play a key role in the eastern German states.

In **Poland**, advisory services are provided by the state (Agricultural Advice Centres (ODRs)), agricultural chambers, private advisory organisations, companies and NGOs. The ODRs are in Brwinów (centre), branches and Voivodships and are responsible for the education, certification and registration of advisers in Poland. They offer financial and economic advice, while technological advice is limited, as well as organise training courses for farmers. Private agricultural organisations operate in the scope of the publicly funded measures under RDP and other national programmes. Commercial firms, which are extensive, supply advice as part of inputs sales and interact with ODRs. There are a large number of certified individual agricultural advisers who work for various institutions, private companies and farming communities under contract. There are also a large number of active FBOs, and Poland has a long history of agricultural production cooperatives.

#### 4.2.2. Integration/Fragmentation

None of the case studies could be described as having an integrated framework for delivering soil advice. They exhibit different extents of integration and fragmentation in the AAS, which can be characterised by both cooperation and competition (for farmer clients and for project funds) between organisations.

With respect to inter-organisational cooperation, in Norway, private or FBOs cooperate and receive support from public bodies for topics relevant to SHM which do not lend themselves to commercial services. In Flanders, increasing collaboration between the CVBB and B3W advisory services provides a good example of the joint effort of several research institutes to address soil topics. Meanwhile, in England, although the AAS is horizontally fragmented, with multiple uncoordinated actors, organisations and delivery activities concerned with advice for different soil functions, there are a number of partnerships and initiatives where organisations work together towards a shared goal for SHM and water quality (for example, Catchment Sensitive Farming initiative). Synergies were identified in Poland, where the public ODRs host training events which bring together large numbers of farmers and invite private-sector companies, who are knowledgeable about the technologies or products, to participate. However, in Spain, a duality of advice was described with a clear distinction between public and private services, which has implications for soil advice.

There were different perspectives in Germany depending on experiences in the respective states: One respondent described few links between providers and competition between the different consultants and large companies. However, for another respondent (for Baden-Württemberg), the interaction between the state and private consultants at the agriculture office level was a strong point, and they agreed that synergies definitely exist, while there may be tensions between individuals.

In line with this viewpoint, the fragmented landscape and different objectives of public and private providers can have consequences for SHM at the farm level. In Spain, although most respondents did not identify tensions or conflicts in advisory service delivery, one respondent acknowledged that contradictions arise when there are commercial interests:

*The system is not fully integrated, this affects sustainable soil management negatively because conflicting advice is given, or conflicting objectives are pursued [* ... *.] when there are commercial interests, we do find contradiction.* ES2

As with Spain, in England, while advice is *"theoretically joined up"* (for example, a partnership will have shared goals), what actually matters is at an individual farm level, where farmers can be contacted by a number of advisers or projects officers. One respondent (UK2) said, *"I wouldn't say that there's contradictory advice now, but duplication",* and also noted that farmers have been advised to do things by a commercial company which are questionable with respect to SHM.

A Polish respondent (PL3) also described tension and competition between companies providing agricultural products. Although, as another respondent explained, this depends on the company:

*There are companies whose approach is to sell their products, and there are companies that act for example together with associations promoting the welfare of the natural environment recommending the use a range of suitable products.* PL2

Regarding vertical research–practice linkages in the soil context, these are considered strong for NLR in Norway which has good links with research; forexample, it is quite common for NLR and the research institute (NIBIO) to be cooperating in projects. This ensures good dissemination but also that projects are relevant to farmers. Researchers, farmers and advisers are also linked in Flanders, where research stations have strong outreach programmes, and in some states in Germany, where district agricultural office carry out practical trials with farmers. In Spain, in the horticulture cooperative sector, there are strong links from research to farmers providing a comprehensive service to these particular farmers. In England, the perception is that research and practice are disconnected, and as the respondent UK2 said, *"It's actually the translation of that [research] into current farming practices, which is where the gap is".*

#### *4.3. Advisory Services Capacity*

4.3.1. Management and Organisation for Delivering Advice on SHM

Some respondents considered that there is good organisation and management of AAS but that other limitations prevent effective delivery of SHM advice. For example, in Germany (Brandenburg), the AAS are thought to be well equipped and prepared in terms of technical capacity, with an excellent research infrastructure around Berlin, but lacking political guidance about soil from the federal government. In Spain, some respondents agreed that despite good organisation and management of advice, more information and knowledge transfer are needed for effective SHM advice to be achieved. In Norway, there was consensus from all three NRL (farmer membership organisation) respondents that they have both competence and capacity to deliver advice on SHM. Furthermore, they were optimistic that advice will improve as public funding is now available to increase the focus on soils.

The capacity of public services, where they are provided, tend to be limited by resource constraints, namely, staff and financial. In Germany, there was a sense of good capacity and resourcing in the consultation services hosted by the state agricultural office in Germany (e.g., Baden-Württemberg); however, respondents noted the staffing limitations of public provision and the need for strong personal commitment. This is reiterated later in the analysis:

*From the public side, we in the agricultural administration are mostly limited by the staff capacities. That is an aspect, which has deteriorated dramatically everywhere in recent years, so if we want to work towards [soil] sustainability, it's only possible through increased commitment beyond the actual working hours.* GR1

Furthermore, the emphasis on inspection and regulation by state bodies in Germany limits their time and scope of work with a focus on inspection. As a consequence, farmers supplement public advice with consultations by private companies.

The Polish state Agricultural Advisory Centres were described as working well to provide advisory services but not yet properly prepared to advise on soil protection, still being stuck in the *"old structures and treatments"* (PL1). They are also constrained by funding and often lose their best advisers to the private sector. The potential of private companies to fill the gap left by public services was identified in Poland. There was consensus that private companies are more visible and accessible and able to meet market demand. Referring to horticultural crops and crop- and soil-borne diseases, this respondent (PL3) remarked:

*There are private companies that have appeared in the market and provide these services at a good level [* ... ... *.].. The institutes [public] have the potential, equipment, experience and knowledge, but it seems that due to financial and personnel constraints as well as other obligations, they are unable to respond to the very high market demand, and it is very large, while possibilities for conducting research are limited. Private companies, which are more and more visible on the market, are trying to fill this gap, which is good, because such companies can provide services as part of, for example, soil or plant research projects.* PL3

However, for private services, the business model is not always commensurate with building capacity. In England, privatisation of the advisory services has introduced a profit incentive which impacts resourcing, as one respondent, who works for a consultancy, explained:

*We have to be a profitable organisation, which means that we haven't the luxury of an infinite amount of time [* ... *..] we do the very best we can with the resources we've got, but that some of the expectations of what it actually costs to deliver service are unrealistic.* UK2

This is also a factor in Germany, where dealing with new issues, such as supporting the necessary transition to new cultivation systems or meeting the state policy requirements for environmental programmes, represents an added effort for the consultation services in terms of costs, time and energy. However, adaptation is seen to be essential to ensure future services:

*And every consultation service is required to adapt, to continuously improve, and to be up to date with the latest science and technology. I think that's actually a very positive development [* ... *.]. But it is clear to them that if they do not consult their farms in the direction of sustainability, they will lose them completely in 10–20 years.* GR1

This need to build capacity for the future is reiterated by a respondent (GR3) who works with large cooperatives south of Berlin, where the long-term nature of soil health has become the focus of attention amongst the scientists who advise on the farms.

Private organisations also find that they have to compete for project funding. In Flanders, although the quality of advice is good for soil in the government research institutes and the independent Soil Service, the resourcing of activities is seen to be constrained by a reliance on short-term project funding, reducing the chance to build strong and enduring relationships with farmers. The remark "*True sustainable soil management advice does not exist to my knowledge, the Soil Service provide such integrated advices only as part of projects"* (BE3) is insightful in that it indicates poor continuity, as well as a dependence of projects for funding.

Staff recruitment and retention has implications for advisers' expertise and experience in SHM and was mentioned across a number of countries. In Norway, it can be difficult to recruit advisers who possess sufficient knowledge about soil if, for example, an experienced adviser retires. High turnover of advisers due to a lack of job satisfaction or progression and financial motivations exacerbates this. In Spain, advisers who belong to technical departments in FBOs (companies/associations) are seen to have more room for manoeuvre and are more organised and professionalised compared to commercial advisers. The absence of planning for the necessary skills and staff which may be needed in 2–4 years' time was also raised as a limitation for SHM advice in England.

Regarding an organisation's culture, there was also recognition that advisory organisations themselves have some responsibility to rethink how they advise farmers who are overburdened, face severe economic pressures and are constrained in terms of investing in new equipment, new crop rotations or new fertilization methods. In this respect, the culture of the organisation is seen to be important in Germany, where every consultation service has a specific philosophy that is shaped by the organisations' decision makers.

#### 4.3.2. Level of Advisers' Knowledge about SHM

In the pluralistic advisory systems described here, it is difficult to characterise the expertise or the quality of advice for soil overall and SHM specifically, as this can depend on the sector and systems they support. However, the following provides some insights.

#### Knowledge and Practical Experience

Practical experience is seen as indicative of good quality advice and private advisers are more likely to acquire this, compared to public advisers, due to their regular on-farm activities. For example, the quality of advice is considered high in consultation services in Germany, although the focus is limited, and wider aspects of SHM advice are not covered:

*I think, the quality of consultation is high [* ... *.] many of the consultants are running agricultural businesses themselves, so they have a certain practical background, or they have simply been working at an agricultural office for many years, so they have a very high level of knowledge [* ... *.] so far,[this] has mostly been on crop protection and, I think, especially in terms of sustainability, sustainable soil management, crop rotation, intercropping, things like that—there is still room for improvement.* GR2

Similarly, in Poland, private advisers were regarded as more effective and active than state advisers who, although knowledgeable, lack practical experience and the ability to follow up on advice to keep farmers up to date:

*A strong point of commercial services is that they have capable advisers. With regard to government institutions, their strong point is certainly their infrastructure and the preparation of speakers, i.e., advisors, who are very knowledgeable, but then somehow nothing happens. And this is the weak point, that there is a lack of continuity, on-site continuity, during on-site workshops. Often these advisers lack practical experience and [* ... *. ] are unable to keep up with these new solutions and products.* PL2

Another Polish respondent (PL3) noted that, with the loss of good quality government advisers to the private sector, their expert knowledge now only reaches farmers who are customers of private companies. This unequal distribution of quality advice (including SHM) was also identified in Spain, where technicians with a good level of specialist expertise in horticultural production support intensive crop growers, but family-based businesses with extensive systems in other sectors in Spain have limited access to good quality advice on soil. Furthermore, pockets of high-quality SHM advice were described for advisers in the organic sector, as mentioned in Poland and in Spain, and for advisers selling products related to, for example, organic or sustainable management who "*provide information about the nature of living soil, biodiversity or soil quality*" (PL2).

In Norway, governmental and public bodies, especially at the local and county levels, were described as not very competent or up to date, with a main role in supporting fertilizer plans and subsidies. However, the respondents all agreed that the standard of advice for soil management is high in the independent NRL, where the advisers are knowledgeable and have an increasing focus on soil health and environment.

#### Soil Fertility Focus

In all case studies, there was agreement that private advisers (working for input companies or as independents) are generally trained to advise on soil from the perspective of fertilization and crop nutrition and tend to look at crop management in the shorter term. This emphasis was noted by a respondent (PL3) in Poland who said, "*My impression is that most advisers focus only on the composition of the soil, on just the chemical factors, but they ignore and totally undervalue the importance of soil microbiology".*

A number of respondents called for a change in the mindset of advisers away from production-orientated to more holistic advice, with a shift in thinking from soil chemistry to a microbiological approach required, to show that *"living soil can achieve more".*

This focus on soil fertility and crop nutrition can have some negative implications. For example, in Flanders, commercial advisers were known to advise maximum fertiliser recommendations irrespective of crop requirements, which is contrary to good practice recommended by research organisations. This was also noted in Germany, where an emphasis on fertilization as part of an overall crop care package can lead to conflicts with advice for other soil functions. This respondent in Poland highlighted how some advisers are 'locked-in' by their company's commercial imperatives despite being knowledgeable:

*Many advisers are enslaved by receiving payment from the company, so they have to advise according to the company's offer, and this limits their freedom to act; they have the knowledge but they will necessarily be focused on bonuses, on a raise, on finances, and this restricts them.* PL3

However, respondents did not think commercial advisers purposely provide negative advice, although they may be slightly less inclined to look at the environment or at soil quality, soil biology, etc. In Spain, where consultants are often influenced by their employers, one respondent suggested that there is no intention to damage soil; however, they may not be aware of the externalities of their advice:

*I don't think there is one main advisory service that has either a positive or negative impact. Normally advisers have the objective of increasing overall production. The adviser does*

*not go against soil sustainability or soil quality, but [* ... *] the use of these technologies continuously without other guidelines in the end leads to an overall degradation of the system, mainly of the soil.* ES2

#### Environmental Shift

The Green Deal and the demands from supply chain companies and retailers to meet food and farming standards and gain a market advantage were considered by many respondents to be driving advisers towards SHM advice. However, there is some cynicism in Poland that advisers and input-sellers are using slogans related to environmental and soil protection issues but, fundamentally, are still largely dependent on the producers of chemical agents for their income.

In England and Germany, there has been a shift in commercial adviser activities towards supplying environmental advice (supporting agri-environment scheme applications, as well as practices for good soil management), and for agronomists linked to input sales to sell cover crop and, pollinator seeds and biosolutions. A German respondent described the growing demand:

*In recent years companies have emerged that strive towards sustainability, selling crop fortifiers, soil additives and so on. Active local consultants and some farms use these products in their cultivation. This is of course due to the fact that, in the last few years, little has been done in terms of soil fertility and sustainability on many farms. They are now reaching their limits in terms of plant cultivation, they have problems with diseases, with the soil, etc., and companies, which offer the appropriate products, have been in greater demand in recent years.* GR1

This situation is replicated in Poland where more companies are entering the market with 'natural products' aiming to meet farmer demands.

#### Heterogeneity

One common factor across all case studies was the heterogeneity in the quality of advisers with respect to soil advice, with a spectrum of very good to very bad commonly being described. In Spain, a range from very good agricultural technicians to others who do not have the necessary knowledge was linked to the number of untrained advisers emerging to meet the increasing demand for sustainability and ecological advice. Similarly, in England, a respondent (UK1) referring to agronomists said, *"I think the good are very good, but I don't think we've got many very good ones, I think a lot of us are in the category of willing triers"*. However, he acknowledged that there are excellent pockets of SHM advice amongst independent advisers and non-commercial initiatives. This range is echoed in another comment by a respondent from England who described the value of long-term experience:

*Some of them are extremely knowledgeable and interested [about soil] and have been in their post for quite a long time. Some of them are on short term contracts. And some of them who are less good than others, in terms of their understanding of the technicalities of what they're talking about, and what they're being asked to do.* UK2

The same sentiment was expressed by respondents in other case studies, where advisers develop a very good reputation because they have been in the profession for many years. A range of abilities and interests was also described in Germany, where the ease of substituting SHM principles with agrochemicals was blamed on a lack of attention to soil by some advisers:

*There are consultation services, or even individual consultants on the part of the industry, who attach importance to the topic [soil]. But there are also people who have never bothered with the subject, because it is still possible to achieve good yields with the use of mineral fertilizers or chemical-synthetic pesticides.* GR1

The distinction between the role of the advisers as generalists or experts was widely discussed. There are very few agricultural advisers across the case studies who focus specifically on the soil or get the opportunity to become experts. Some take the view that soil experts can be consulted when necessary, but that wider skills are needed at farm level, as this respondent explained:

*Rather than being experts on that particular aspect, we reflect the farming community, in the sense that we are people with a wide range of skills, but an expert in nothing. An expert—he's talking purely about the soil, and the health of the soil, we will be talking about it on the profitability of the rotation, the control of various injurious weeds, diseases and pests, and then looking at a rotation that is sustainable, which then comes back to the soil. However, we know where to go to get expert [soil] advice.* UK1

Differing perspectives on the value of experts versus generalists were picked up in the Spanish interviews. One respondent agreed that a historical focus on supporting production has led to fragmentation where an agricultural technician may know a lot about tillage or agricultural equipment but does not have a general vision of sustainable soil management. The other two respondents in Spain, however, argued that advice to farmers on soil management is too general and the level of expertise low; one (ES4) identified "*A strong need for the participation of people who are soil specialists—soil scientists, biotechnologists with application to soil microbiology".*

#### 4.3.3. Advisers' Training for Delivering SHM Advice

There are a number of opportunities through multiple talks and events for all advisers to expand and update their SHM knowledge, mentioned for all case studies. In Flanders and England, for example, large numbers of advisers reportedly attend dissemination events and demonstration days, and for many, this is important for networking. In addition, there is now comprehensive information about soil topics on the internet and social media and opportunities for peer-to-peer learning and exchange. However, as noted already, poor attention by advisers to SHM has been attributed to the absence of good training.

Time and resources for soil training are a concern for some. As one respondent (N2) in Norway noted, "*unfortunately we have to prioritise covering our hourly rate as employees, so that can affect how much time we have to educate ourselves, go to conferences, seminars*", illustrating the fact that advisers (from all organisations) are often under financial targets and pressures to the detriment of their training and upskilling in SHM. This imperative steers organisations' decisions about training as well. In Flanders, for example, obtaining a certificate (Flemish Land Agency certification or other quality control procedures) is costly both in terms of time and money, and as a consequence, certification is profitable for only a few advisory institutes/services.

There are a large number of options for in-service training in Poland with ODRs taking on a key role for farmers and advisers. Advisers within the commercial sector in Poland are also considered well trained but only within the sphere of their operations and products:

*It seems to me that every commercial business tries to train its advisers so that they do at least have this information as regards their own products, how they affect the soil and therefore they must have prior knowledge or learn about the soil, its quality, the processes that take place in the soil environment.* PL3

In Spain, the nature of skills and training depends on the type of agricultural technicians (cooperative, input company or independent). Most respondents agreed that the level of SHM training of technical advisers in Spain is low overall, as one remarked (ES2): *"Advisers do not have sufficient skills and experience to give advice on sustainable soil management [,* ... *] because they have not had sufficient training during their studies".* As such, these agricultural technicians need to seek out further training to enable them to meet changing demands. These points were reiterated for Poland, where the notion of continuing education was raised:

*Every adviser needs to participate in continuing education, as the knowledge gained when graduating from college is not enough [* ... *.]. It is necessary to educate, educate and* *again educate advisers and farmers, and to provide this new knowledge about sustainable soil management, which is completely different from the information provided before.* PL1

Another respondent from Poland (PL3) supported this, remarking that studies are only the basis and a good adviser has to train for the 'rest of their life', otherwise, they quickly lose touch with reality. In Spain, respondents noted that there is no unified certification validating the agricultural technicians' knowledge. In Germany, large differences in the range of training courses were described, with more being offered in recent years. In England, there is an established continuing professional development (CPD) scheme for advisers (FACTS, BASIS) which offers courses on soil and water management. While acknowledged to be outstanding compared to other European countries, a respondent pointed to the inadequacy of these courses in terms of SHM:

*In terms of sustainability, I think they're both useless. They've evolved out of a commercial requirement. So it wasn't evolved to deliver good, independent, impartial information [* ... *] they do provide a level of professionalism.* UK1

Another respondent from England thought that a FACTS-qualified adviser would understand about nutrient management but argued that BASIS is too technical and academic and that the modular training does not prepare advisers to deliver integrated advice, considering soils, nutrients, water management together, nor help them understand the underlying principles of SHM:

*So as far as, is the training fit for purpose for the next generation of advisors? One of the problems that we and the whole industry has got to know that there's plenty of advisers who are qualified, but not necessarily have a good understanding of the principles[* ... *] you need to be able to understand what you're doing. And why are you doing it.* UK2

There was also agreement that capabilities need to be expanded to beyond a focus on production objectives and soil fertility and crop nutrition advice, to meet new demands, reinforcing the points made earlier. This respondent from Spain noted that this was a key limitation for SHM advice:

*From my point of view, there is enough organisation to provide advice on sustainable soil management and there are enough people capable of providing basic guidelines for sustainable soil management but there is a lack of general training on what is the true nature of soil quality beyond nutrient fertility.* ES3

The extent of informal learning through adviser networking was mentioned by some respondents but did not emerge as a particularly strong aspect in the interviews.

4.3.4. Attitudes and Motivations of Different Advisers and AAS

Positive adviser attitudes towards the soil were described by a number of respondents, however, there is still a range of attitudes linked to economic motivations. Fundamental differences in motivations between advisers were identified in Spain and this aligned to their organisations' objectives:

*An adviser who belongs to a trade union or a regional agricultural office has a different vocation than an adviser who belongs to a commercial company or to a research centre; their motivations are very different, which means that their inclinations are also very different.* ES1

This can have implications for advisers' reputation and credibility. According to respondents, for example, in the horticulture sector in Almeria, agricultural technicians do not always have high social value and may even start to have a bad reputation. This is echoed in Poland, where the balance between commercial advantage and gaining farmers' respect was seen to be important: *"There is no doubt that an advisor's motivation is influenced both by economics and by the desire to be respected by farmers, it really depends on the person"* (PL1). Many agreed that farmers are able to quickly discern any 'shortermism' and the commitment and motivation of advisers.

For many respondents the different motivations and attitudes of the individual adviser were regarded as more important than the type of organisation they belong to. The high level of personal commitment required by some advisers to pursue their interests in, and deliver, SHM advice was mentioned by respondents. Consultants in Spain, for example, are often limited by specific short-term projects or task forces. When these are finished, if they want to continue with the topic, this has to be done in their own time. Similarly, in Germany, personal effort is linked to quality advice:

*Yes, well, there are advisers who are just all-around good advisers who really give their best and try to constantly educate themselves in order to be able to provide the best possible consultation to the farmers* ... *I think most of the vocational counsellors actually—and yes, I think that of the ones that I know, most really put their full effort into it.* GR2

#### **5. Discussion**

#### *5.1. Governance Structures*

This analysis confirms the picture painted by previous researchers of considerable AAS diversity between, and pluralism within, European countries [71]. This translates into a diverse landscape for SHM advice with different governance, funding and delivery mechanisms and no evidence of any integrated advisory frameworks for delivering advice for soil management. The analysis shows that institutional options available for financing and the level of coordination are limited with respect to delivering advice for soil management, as observed elsewhere for AAS more generally [50,56]. A reduced central organisational role of government agencies in AAS and an emerging 'knowledge market' [33] has led to a continued decline in the public sector's role in delivering on-farm soil advice for all case studies, with the diversion of their resources and staff towards compliance regulation and scheme/grant administration. Conversely, the prominent role of the private sector, independent organisations, FBOs and NGOs is apparent in filling the gap in delivering on-farm advice that influences and impacts soil management, either through contracts (projects) to fulfil government objectives (e.g., FAS, grants) or commercially in a more market-led environment, as described in other AAS studies [33]. When state and private advisors are incentivised to administer regulations and grant applications, this narrows down choices and limits broader understanding of 'know-why' soil processes [14,72]. New services are also emerging, and overall, the number of advisers with commercial links to economic actors (input suppliers, consultants) is increasing [26].

Fragmentation means competition for clients and project funding, and soil advice at farm level can be compromised by conflicting delivery or duplicating advice in multipartner approaches, as reported by others [73]. However, many hybrid and dynamic arrangements for partnering and funding for delivering SHM advice are notable. These 'creative alliances' provide opportunities for the effective integration of delivery of soil advice at programme level. This ability of pluralistic advisory services to overcome constraints (shortages in funding, staffing, etc.) through increased cooperation, collaboration and partnerships has been observed elsewhere [29,71,74,75]. Individual relationships of both competition and cooperation, described by Compagnone and Simon [24], were not shown in this analysis.

#### *5.2. Advisory Services Capacity*

These governance arrangements provide a backdrop to understanding different organisational arrangements and capacity to provide SHM.

#### 5.2.1. Management and Organisation Capacities for SHM Advice

The analysis identified organisational constraints in resources, funding and staffing, notably in the public services, which are not always able to meet demands, and this impacts the capacity to deliver SHM advice. There are inherent frustrations concerning reliance on short-term project funding for developing and continuing with advice streams, as previously described for environmental advice [31,76]. This often means only committed

advisers continue with the SHM advice when the project ends. Poor staff retention [24], with the loss of advisers' knowledgeable about SHM to the private sector, reduces farmer access to SHM advice. Although farmers might look outside of formal advice in such circumstances [72], their options for benefiting from high-quality soil advice are diminished. Other commentators have noted that commercialisation threatens the extension capacity of government agencies [77], however, technical expertise has not been considered.

Investment in staff capacity for SHM advice (training and field days) is restricted in both public and private sector organisations by limited time and resources and the competitive business environment. Small firms also struggle to meet new environmental requirements, corresponding with previous observations [34,42]. Furthermore, in some commercial organisations, economic drivers can lead to an organisational culture that values input sales over expertise in SHM.

These organisational capacities affect individual advisers' capacity to operate effectively, their objectives and motivations, their professional culture and the support they are offered to deliver SHM. As observed by Klerkx and Jansen [34], this wider set of institutional conditions, and the relationship with the 'back-offices' which supports them professionally, is critical for enabling advisers to develop and deliver specialist and professional advice. Furthermore, maintaining a stable or increasing workforce as well as diversifying the expertise and increasing the competence of staff are seen to be critical for AAS [29].

#### 5.2.2. Individual Capacities for SHM Advice

Individual capacity results from a combination of attributes: quality of advice; training; and motivations in relation to SHM. Firstly, regarding advice quality, heterogeneity in levels of advisers' soil knowledge was observed across all cases and across all AAS types, with few advisers considered to be delivering all-round high-quality advice to support SHM. This adds to the emerging body of evidence showing that advice on soil management is suboptimal. What constitutes 'good quality advice' with respect to soil management was understood differently due to advisers' varying goals and their clients' needs. It was generally characterised by, not only extensive on-farm practical experience and a good level of subject-matter knowledge or expertise [67], but also critically by an understanding of soil chemical, biological and physical processes and principles [78]. Private advisers (commercial consultants, technicians and agronomists), while being credible with respect to providing high-quality technical advice, are limited in scope to soil fertility and crop nutrition. This observation is supported by studies showing the predominance of advice based on nutrient testing and interpretation to support farmers' short-term production decisions, e.g., [45]. This limits opportunities to incorporate soil health perspectives into advice, which are critical to understanding the capacity of soil to function as a vital living system [16,17]. Only very few advisers are taking a holistic approach, accounting for non-linear mechanistic relationships between various physical, chemical and biological soil properties considered important for soil health [19].

The significance attributed to practical experience, however, should not be overlooked. This allows advisers to provide localised advice and meet the fine resolution of soil information and data that farmers require [9]. This highlights the value of experiential learning (and co-learning with farmers), which has a particular significance for soil management due to the in-field observations and sensory experiences required [79,80] and is highly appreciated by peers and the practitioner community [78].

Equally, whilst expertise in soil science and management (demonstrated by some individual advisers) is valued, the role for the generalist agronomic adviser who takes a whole farm perspective is seen as important. Interestingly, advisers have been shown to be capable systems thinkers [74] and positioning SHM within the wider farm business and environment is in itself an important skill. Further specialisation, in, for example, soil microbiology was called for by some respondents, in line with emerging farmer interest in soil health, but how such specialists would position themselves in the AAS

landscape was not elaborated. Landini [41] suggested that not all individual advisers can hold the same knowledge and capabilities but instead can act in groups to enrich their work. This professional distribution of advisers' SHM roles, skills and specialisms and the way they interact, complement and learn from each other, is an interesting area for future research [24,79]. Furthermore, the changing role of the technical 'expert' needs consideration [81].

Secondly, with respect to training, poor investment in training and particularly in continuing education in SHM in both private and public spheres was seen to be a key reason for the limited scope of advisers' expertise. Training and professional development courses on soil topics, whilst considered to be at a high standard in certain countries, do not always provide an understanding of the principles and processes of SHM. A number of studies have shown that advisers are increasingly relying on each other for sharing soil expertise through professional networking [82]; however, this was not identified in this analysis as an alternative to training. These findings are inconsistent with previous studies [30], although the focus was not SHM.

Thirdly, regarding motivations, personal intrinsic interest in soil was a further facet demonstrated by a few public and independent advisers. The economic motivations of private sector advisers' (linked to input sales) observed here are widely reported in studies that concern soil [10,36,83]. The image of advisers as 'locked into' supporting intensive agriculture pathways has been also described for high-input production systems [26,84], as has the power of supply chain actors [36]. However, analysis here suggests a more nuanced picture, with many private advisers balancing economic motivations with the need to retain respect, social value and trust in the farming community. This loyalty dilemma between private good (what the farmer demands and pays for) and public good (issues of broader importance for society as a whole) [29], may need to be re-examined in a future context when incentives for providing SHM become available (e.g., carbon farming, Environmental Land Management Schemes in England). Organisations are already responding to the market and offering a range of environmental services, and supporting sales of 'natural' biological products. However, the depth of understanding and commitment that accompanies these was queried, and there were calls for a more fundamental shift in advisers' mindsets.

Professional culture is closely connected to individual advisers' motivations and mindsets, accepted norms and values, how they perceive and execute their tasks [34], and their performance rationale and economic strategies [26]. However, adviser roles are not set: Nettle, Crawford and Brightling [42] describe the fluid nature of adviser professional identities and opportunities for evaluating their roles through reflective practice [41,85], which, if organisations were more flexible, could lead to reorientation of soil management advice.

#### 5.2.3. Narrowing Down

Although it was not the intention here to assess the performance characteristic of the framework, some observations can be made. The needs and opportunities, which characterise performance [50] that have been steering advice in relation to soil are: policy (cross-compliance regulation and grant administration support) and markets (farmer demands for crop production advice). As a result, there has been a narrowing down of soil advice, both with respect to content and access, as depicted in Figure 2.

However, the increasing interest in soil health from both farmers, in part due to the recognition of soil degradation [18,86], and policy makers, will provide the new drive and opportunities to widen the scope of advice to cover physical and biological, as well as chemical, processes. To achieve this, AAS organisations will need to invest in adviser training and capacity building and aim to shift professional cultures and mindsets at organisation and individual level. This will require incentivisation, and Dhiab, Labarthe and Laurent [26] identified a need for public policy intervention to support this. This could be through, for example, strengthening national FAS with requirements for member states to provide standardised and certified adviser training and continuing professional development in SHM. Ultimately, however, AAS are shaped by the framing conditions, the priorities within the agricultural sector strategies (high-value commodities or environmental sustainability) which are beyond the direct influence of policy makers and advisory services managers [50]. In turn, these determine the governance structures and the relative capacities of public, private or NGO AAS and the services offered. As Knierim, Labarthe, Laurent, Prager, Kania, Madureira and Ndah [23] point out, the historically grown, pathdependent institutions and institutional constellations in each EU member state play an important role in AAS.

**Figure 2.** This figure shows how the collection of capacities act to narrow down the nature and extent of advice for soil. Changing needs and opportunities will open up the scope of advice for delivering SHM.

There have been calls for capacity building in knowledge systems at individual, organisational and AKIS levels [42]. This encompasses adviser training and professional development and more back-office support [28,85] as well as the need to understand the varying roles of professional advisers [87]. However, the focus has often been on process skills, the (new) intermediary, advisory styles and facilitatory skills that advisers should master to support and empower farmers in networks of interactive learning [88]. Adviser technical or specialist roles have received less attention, notably for soil, despite the growing demands placed on them for understanding and supporting land managers in the management of complex soil functions.

#### **6. Conclusions**

The framework employed allows the collective capacities (governance structures; organisational and individual capacities) of AAS for SHM advice to be revealed. It shows that advisers' competences and skills should not be seen in isolation. As such, the recommendations for expanding the scope of content and access to SHM advice include addressing deficiencies in training and capacity building, shifting professional culture as well as addressing more deep-seated institutional conditions and governance structures. Incentivising such changes will require changes in both policy and market drivers. These insights show that AAS can play a central role in the transformation of food systems more widely [89].

The method based on in-depth interviews (18 experts) provides insights for a crosssection of European countries offering a range of perspectives, as well as common themes with respect to capacities which affect the nature and extent of SHM advice. However, the results can only be indicative for Europe as a whole and further qualitative and quantitative research will be needed to provide a more comprehensive picture. In particular the results

show how different advisory services that influence and impact soil evolve in specific country contexts. This suggests that the model of identifying systems that best fit contextspecific conditions is suitable for future support of national AAS with respect to SHM. Critically, the methodology did not explore the complexities of the relationship between advisers and farmers/land managers, nor capacities in terms of the soft skills required for co-producing technical soil knowledge or the changing mature of the 'expert' role of advisers.

With the accelerated move towards the integration of soil health issues in a number of European Commission strategies and the actions and ambitious targets set for soil health within the Soil Mission 'A Soil Deal for Europe', the requirements for building capacities and a knowledge base for soil health enhancing practices in agriculture will increase [13]. This will require member states to significantly enhance their AAS capacities to achieve this desired transition, with implications for both European and national level policies.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/land11050599/s1, SoilCare advice review Interview schedule.

**Author Contributions:** Conceptualization, J.I., J.M. and C.-A.C.; formal analysis, J.I. and C.-A.C.; investigation, J.M., J.E.B., C.-A.C., J.A.A.-S., A.E., M.F., B.L.-F., P.M.-G., K.S., J.S. and M.T.; methodology, J.I. and J.M.; project administration, J.M.; resources, J.M.; supervision, J.M.; writing—original draft, J.I.; writing—review and editing, J.I., J.M., J.E.B., C.-A.C., J.A.A.-S., A.E., M.F., B.L.-F., P.M.-G., K.S., J.S. and M.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was part of SoilCare (Soil care for profitable and sustainable crop production in Europe). Grant Agreement 677407 funded by the European Union's Horizon 2020 research and innovation programme. www.Soilcare-project.eu (accessed on 20 September 2021). 2016–2021. This funder had no role to play in the design, data collection, analysis or interpretation of results presented here.

**Informed Consent Statement:** Ethical standards and guidelines have been applied to the collection, processing and storage of data about persons, in accordance with the agreed project ethical statement. All subjects gave their informed consent for inclusion before they participated in the study.

**Data Availability Statement:** Data are confidential and not available.

**Acknowledgments:** We acknowledge the support given by other partners and stakeholders in the SoilCare project and in particular the interview respondents.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **Abbreviations**


UK United Kingdom

CPD Continuing Professional Development

#### **References**


## *Article* **An Overview of Sustainability Assessment Frameworks in Agriculture**

**Abdallah Alaoui 1,\*,†, Lúcia Barão 2,†, Carla S. S. Ferreira 3,4,5 and Rudi Hessel <sup>6</sup>**


**Abstract:** Recent research established a link between environmental alterations due to agriculture intensification, social damage and the loss of economic growth. Thus, the integration of environmental and social dimensions is key for economic development. In recent years, several frameworks have been proposed to assess the overall sustainability of farms. Nevertheless, the myriad of existing frameworks and the variety of indicators result in difficulties in selecting the most appropriate framework for study site application. This manuscript aims to: (i) understand the criteria to select appropriate frameworks and summarize the range of those being used to assess sustainability; (ii) identify the available frameworks to assess agricultural sustainability; and (iii) analyze the strengths, weaknesses and applicability of each framework. Six frameworks, namely SAFA, RISE, MASC, LADA, SMART and public goods (PG), were identified. Results show that SMART is the framework that considers, in a balanced way, the environmental, sociocultural and economic dimensions of sustainability, whereas others focused on the environmental (RISE), environmental and economic (PG) and sociocultural (SAFA) dimension. However, depending on the scale assessment, sector of application and the sustainability completeness intended, all frameworks are suitable for the assessment. We present a decision tree to help future users understand the best option for their objective.

**Keywords:** agriculture; sustainability frameworks; socio-economic and environmental indicators; soil land management

#### **1. Introduction**

Agricultural land covers over a third of the earth's surface [1] and 41% of the European Union's 28 member states [2]. Agriculture uses and affects natural resources, such as soil and water, shaping the landscape and contributing to establishing and maintaining semi-natural habitats [3]. Over the last decades, agricultural management practices have changed considerably to enhance crop yields and productivity to ensure food security [4]. This has been achieved through (i) technological developments, particularly by improving and adapting machinery to different management requirements, the genetic improvement of seeds and development of new agrochemicals [5], (ii) the plantation of extensive areas of monocultures [6] and (iii) the high use of mineral fertilizers and phytopharmaceuticals (e.g., pesticides and herbicides) [7–9].

The pressure on the agriculture sector will continue to rise due to global challenges, such as an increasing population and food requirements, and climate change [10]. To

**Citation:** Alaoui, A.; Barão, L.; Ferreira, C.S.S.; Hessel, R. An Overview of Sustainability Assessment Frameworks in Agriculture. *Land* **2022**, *11*, 537. https://doi.org/10.3390/ land11040537

Academic Editors: Guido Wyseure, Julián Cuevas González and Jean Poesen

Received: 16 March 2022 Accepted: 1 April 2022 Published: 7 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

meet the world's projected food demands by 2050, food production must increase by 60–100% [11]. Furthermore, global agricultural production will be affected by increasing competition with certain non-food crops for several economic sectors (e.g., energy for bio-fuels production), a reduction in market prices due to globalization and limited natural resources driven by, e.g., land degradation and water scarcity [12,13] exacerbated by climate change [14].

Agricultural intensification is often associated with environmental degradation, including soil erosion, water, and soil contamination, and biodiversity loss [15–18]. By the end of the 20th century, the consequences of the intensive agriculture approach, especially in developed countries, were thoroughly investigated and frequently reported. As a result, agriculture had been highlighted as one of the main activities worldwide contributing to water depletion [19], soil degradation/pollution [20,21], biodiversity loss [22] and climate change [23]. According to the EU Soil Thematic Strategy [24], the erosion and loss of organic matter are some of the major soil threats affecting agricultural areas, along with compaction, contamination, salinization and loss of soil biodiversity.

Besides environmental problems, intensive agriculture also causes social damage and the loss of economic growth itself in the medium/long term [25]. Thus, the integration of environmental and social dimensions is key for economic development itself, and sustainable agriculture is therefore seen as the only approach towards a successful future [26]. When assessing the sustainability of different agricultural land-uses and land management practices, it is therefore important to consider not only the immediate economic benefit but also how they compromise the overall environmental quality and affect the rural communities, since these factors are relevant to sustaining future economic growth in the short and long-term [27].

As stated in the literature "Sustainability is a multidimensional concept [28] of a dignified life for the present without compromising a dignified life in the future or endangering the natural environment and ecosystem services" [29],. Its evaluation process plays an important role in the development and promotion of sustainable agricultural systems [30]. To investigate the transition towards more sustainable production, various frameworks have been proposed to gain knowledge about the sustainability performance of such production systems [31,32]. Some of these frameworks are based on indicators, whereas others are based on indices (e.g., [33]). Indicator-based sustainability assessment frameworks combining environmental, economic and social issues require the processing of a wide range of information (qualitative vs. quantitative), parameters and uncertainties [34]. They also differ in scope, target audience, indicator selection, aggregation, weighting and scoring methods, as well as the time required to complete the assessment [35]. Although many frameworks emphasize the necessity of including socio-economic and environmental aspects in sustainability assessment, many others focus only on environmental indicators to investigate the short- and long-term effects of different agricultural management practices [36] or are applied to a specific context [37]. In addition, existing assessment methodologies to investigate agricultural sustainability are scattered, focusing on single, complicated and demanding aspects regarding time, cost and required skills.

The main aim of this paper is to identify and summarize the indicators and frameworks used to assess sustainability in agricultural areas. The specific aims are (i) understanding the criteria to select appropriate frameworks and summarize the range of those being used to assess the environmental and the socio-economic themes of agricultural sustainability; (ii) identifying the frameworks available to assess agricultural sustainability; and (iii) understanding the methodological approach and analyzing the strengths, weaknesses and applicability of each framework.

#### **2. General Considerations**

The following section summarizes the general considerations about the indicators' importance, and selection criteria to set the context for those commonly used in the selected frameworks to assess sustainability in agriculture.

#### *2.1. Criteria for Selecting Sustainability Indicators*

Indicators are set to monitor and highlight the current conditions and enable stakeholders (e.g., farmers, businesses, policymakers) to identify trends and compare performances among specific places, such as farms, regions or countries, concerning their sustainability performance [38]. They should present the results in a way that is understandable by people with different occupations and sociocultural and educational backgrounds, since they are a powerful public communication tool [39].

The selection of indicators is crucial since it influences conclusions. Thus, the purpose of the assessment, the system boundaries (e.g., aims, scope and temporal and spatial scales) and the end-users should be clearly identified [40]. The assessment should also establish a baseline or reference value (starting point to measure change from a certain state or date) or target (usually established by policymakers). The comparison and contextualization helps to understand the current state or trend [41] and to support the interpretation/significance of the results [39]. Criteria to select indicators include: (i) coverage of environmental, economic and sociocultural dimensions of sustainability [1]; (ii) practicability and simplicity considering field measurements and data availability (e.g., historical data), which should consider spatial and temporal data coverage, reliability, accuracy and consistency [38,42]; (iii) the meaningful use of the indicators to take into consideration the differences in culture and geography to match them to locally relevant problems [39]; (iv) the system's sensitivity to both anthropogenic and natural stresses [1]; (v) meaningfulness to end-users in order to respond to stakeholders' expectations and support policy decisions [40]; and (vi) costeffectiveness, since the costs to produce the information should justify the benefits of the knowledge produced [40].

Selected indicators can be assessed by qualitative or quantitative techniques [41]. Qualitative techniques are typically based on visual evaluations applied at the field scale and have been increasingly used to evaluate the soil quality (e.g., soil structure and texture, rooting depth and slope) and farm management information [42]. Ball et al. [43] summarized the visual assessment techniques that can be used to monitor soil structure, soil quality and fertility as impacted by land management. Quantitative techniques include: (i) direct measurements via field data collection (e.g., crop yields); (ii) a compilation of secondary data based on a literature review; (iii) statistical correlations of the existing data (e.g., soil compaction); (iv) modeling approaches based on empirical models (e.g., biophysical and economic); or (v) sensing approaches, such as spectroscopic techniques and remote sensing [1].

#### *2.2. Indicators Typically Used*

Table 1 summarizes chronologically some relevant studies assessing the sustainability of different agricultural practices using indicators. These studies acknowledge the need for a coherent and consistent methodology to successfully evaluate the agricultural management practices and the adoption of three-dimensional indicators. They demonstrate that an oversimplification of the evaluation does not provide a comprehensive overview of the sustainability potential of the different farming practices. These studies also show the myriad of indicators/methodologies that can be used when assessing agriculture sustainability, namely when different farming systems, practices and geographical locations are considered.

Due to the growing concern for environmental issues, numerous indicators devoted to the environmental dimension have been used, and relatively little integration of social and economic aspects on farm assessments has been considered [40]. Environmental indicators reflect the complex interaction between agriculture and environment, providing a causeand-effect relationship. They tend to include the number and type of crops in the farm, since it links to agricultural biodiversity; soil cover, which is linked to soil erosion; water use; nutrient balance (particularly of nitrogen); and the use of pesticides [44,45], given their toxicity to the environment.

Since soil is nowadays seen as one of the most valuable resources on Earth, given its essential elements to sustain and maintain life, it has received increasing attention under the environmental indicators, and typically includes physical, chemical and biological aspects. Bünemann et al. [46] identified the most frequently proposed soil quality indicators and summarized the measured soil properties that have been used for assessing the environmental dimension in agricultural land uses from 65 soil quality assessment approaches.

Economic indicators aim to address the economic context, focusing on the economic viability defined by profitability, stability, liquidity and productivity, based on input and output prices and yields [47].

Profitability is calculated by cost and revenue, and includes variable and fixed costs (e.g., land rent), whereas liquidity measures the ability of an enterprise/farm to meet short- and long-term obligations and stability is determined by the equity share and equity development [39]. Another important indicator is productivity, which measures the ability of production systems to generate output [48]. Typical economic indicators also consider public subsidies for the farmers, since they provide protection regarding their agricultural activities. GDP is sometimes considered as an indicator of the difference between producers' income and transfers to other economy sectors (variable costs, subsidies) [44].

Most social indicators focus on the following: (i) the sustainability of the farming community, which involves the welfare of the relevant actors and communities; and (ii) the sustainability of society as a whole. The first type of indicators focuses mainly on working conditions, education and the quality of life defined by physical well-being and psychological well-being [40]. Social sustainability is linked to society's demands, with regards to its values and concerns [49], and may be grouped in: multifunctionality (e.g., quality of rural life, contribution to local employment and to ecosystem services) [50], sustainable agricultural practices (e.g., animal welfare, environmental impacts) and product quality (e.g., quality processes, food safety) [40]. These indicators also tend to measure the socio-economic implications of agriculture in the rural income, and may be measured by the total labor generated, as well as by the seasonal variations linked to individual crop requirements, often associated with peaks in agricultural employment (e.g., sowing and harvesting) [44]. Measuring indicators of a sociocultural dimension is challenging, since they are based on a qualitative assessment and are therefore subjective. Farm-communitybased indicators are usually based on farmers' self-evaluation gained from surveys or interviews [40].

**Table 1.** Most relevant papers on sustainability providing relevant information on the indicators and their relevance for case study applications and different conditions. The papers in relation to the frameworks considered in this manuscript are not reported, except those comparing different frameworks.


#### **Table 1.** *Cont.*


#### **Table 1.** *Cont.*


#### **3. Methodology**

During the past 20 years, various approaches and tools have been proposed for assessing the overall sustainability in the agricultural production system and food sector [31,74,75]. However, these methods have many limitations. As an example, life cycle assessment tools quantify many aspects of the environmental dimension in a narrow way, need a high amount of data and do not consider the impacts on soil quality and biodiversity [76] and economic and socio-cultural impacts [77], or can only be applied to agricultural enterprises [32]. Eco-management and audit schemes, as well as sustainability reporting systems, include procedures accounting for the sustainability of a company, but do not

enable comparison between the outcomes of different ones since they are not science-based assessments [78].

In this study, we selected indicators and frameworks based on the following criteria: (1) went through a peer review process, (2) have a farm assessment level, (3) cover universal agricultural sectors, (4) include the three dimensions of sustainability, (5) suitable for Europe and countries worldwide and (6) present transparency of information allowing for an informed assessment as well as solid cultural and value-based elements.

For the search of the frameworks, we considered literature including at least one peer reviewed publication, reports and presentations available online by searching on scientific web platforms.

Each framework selected was therefore described by stating information on the type of tool used (software, database, etc.) and where it can be found available, requisites for running the tool, type of input data required, time needed for the assessment and number and description of indicators (environmental, socio-cultural and economic) used.

The six sustainability assessment frameworks were also compared according to their ability to cover the main themes of environmental, economic and sociocultural dimensions, and their themes were reported. We compared their strengths and weaknesses and developed a decision tree based on possible scales, sectors of applicability and the completeness of sustainability dimensions required to help stakeholders decide which framework is the most suitable for their sustainable assessment purposes.

#### **4. Results**

Based on selected criteria, the following frameworks were identified: SAFA, RISE, MASC, LADA, SMART and PG. Below, each framework is briefly described, as are the environmental (Table 2), sociocultural (Table 3) and economic (Table 4) indicators included in each one of them. In the next section, their strengths and weaknesses are highlighted individually.

**Table 2.** Environmental themes, sustainability objectives, indicators and measured parameters for each framework considered in this study.



#### **Table 2.** *Cont.*


#### **Table 2.** *Cont.*



<sup>1</sup> Originally "Plant protection" in the RISE framework.

**Table 3.** Social themes, sustainability objectives, indicators and measured parameters for each framework considered in this study.



#### **Table 3.** *Cont.*

**Table 4.** Economic themes, sustainability objectives, indicators and measured parameters for each framework considered in his study.



**Table 4.** *Cont.*

<sup>1</sup> Originally "guaranty of production levels" in the SAFA framework. <sup>2</sup> Originally "liquidity" in the RISE framework. <sup>3</sup> Addressed by operational management with the indicators: goals, strategy and implementation, information availability, risk management and sustainable relationships. SMART has a 4th dimension "Good Governance", with the following themes: corporate ethics, accountability, participation, rule of law and holistic management (not included here).

#### *4.1. SAFA*

The Sustainability Assessment of Food and Agriculture systems (SAFA) is a framework developed and proposed by FAO to assess the environmental and social impacts of food and agricultural operations [79]. It offers a comprehensive reference framework for assessing sustainability in agricultural, forestry and fishery chain systems. The framework is designed hierarchically starting with four dimensions: environmental integrity, social well-being, economic resilience and good governance [72].

The available software (https://www.fao.org/nr/sustainability/sustainabilityassessments-safa/safa-tool/en/) (accessed on 4 March 2022) calculates 116 indicators that target the principles of sustainable development. Measured and/or calculated data from production sites with defined unit processes of a system include a wide diversity of sources, including literature or available databases, and public and other independent

sources of information. Additionally, interviews are carried out with local employees in the sector considered. Data analyses should be conducted by an expert in sustainability. SAFA-Tool assists users with setting system boundaries and scoring ranges, and selecting targets, practices or performance indicators from qualitative or quantitative information. The latest software version 2.4.1 allows the user to add their own indicators. Depending on the complexity level of the analysis, determined by the choices made by the user, data collection may range from ±2 h to weeks, and the total assessment from 0.5 days to months [69].

Environmental indicators established in SAFA cover a broad range of themes including water use, wastewater quality, soil quality, air quality, species conservation practices and ecosystem diversity, energy-saving practices, material consumption and reduction practices, energy use and animal welfare, all linked to the food and agriculture processes (Table 2). The social angle of the evaluation process is also very well represented in SAFA, with the rating of indicators covering themes such as employment contracts, the wage level of employees, safety and health environment, job satisfaction, gender equality, cultural diversity or even transparency in the labeling, safety for the consumer and the impact of using a regional workforce (Table 3).

Finally, economic indicators figuring in SAFA cover both profitability and vulnerability topics, such as the net income, production cost and risk and stability of the market or risk management, among others. It also includes indicators related to accountability, such as the existence of system traceability, the investment potential and the will to invest in local economy (Table 4).

LADA data are extracted from the LADA indicators' toolbox developed for LADA (see [80]); the indicators of LADA are divided into two types: those describing the state of the resources+ and those describing direct pressure on the resources++; thus, the indicators used are those that indicate the degradation type

#### *4.2. RISE*

The framework RISE (Response-Inducing Sustainability Evaluation) was developed by Hafel, in Switzerland, for evaluating the environmental, sociocultural and economic sustainability of farm operations [80]. Currently, the RISE version 3.0 software can be found online (RISE 3.0 - Software Manual (bfh.ch)) (accessed on 4 March 2022) or offline (Microsoft SilverlightTM plug-in required) to analyze the data. It includes a total of 50 indicators addressing environmental, social, economic and land management aspects. The data are collected with a questionnaire-based methodology, where farmers are interviewed for 3 to 5 h, which, with the additional time for data computation, requires a total assessment time of 5–9 h [80]. The framework should be used by agronomists or specialists in agricultural advisory. The results are thoroughly discussed with farmers and used to support the continuing improvement of farm sustainability. The environmental indicators included are mainly related to water use and plant protection (Table 2), whereas the social dimension is focused on the workload and the economic dimension mainly tackles the business vulnerability by assessing the financial liquidity (Table 4).

#### *4.3. MASC*

INRA (Institut National de la Recherche Agronomique) developed MASC (Multiattribute Assessment of Sustainability of Cropping Systems) to assess how cropping systems contribute to sustainability at the farm level [13]. The tool that is currently available (http://wiki.inra.fr/wiki/deximasc/Main/) (accessed on 4 March 2022) uses a decision tree to break down the sustainability assessment decisional issue into 32 input criteria. Indicators used to assess these basic input criteria can be chosen by the user depending on their accuracy and the context of their study, as well as the available data [63].

Qualitative and quantitative information is collected through questionnaires and reported results. Methods such as MASC that are suited for the analysis of qualitative data may be more relevant for sorting and categorizing technical solutions when considering a wide range of performances [13,81]. The tool should be managed by a researcher/professional, who then interprets the results obtained.

The indicators included in this framework deal with the evaluation of environmental aspects such as water use, biodiversity and energy use through indicators of water dependency, number of pesticides doses and energy conservation (Table 2). Social indicators are also included, especially targeting the safety and health trainings of employees and the priority to employ a regional workforce. The economic dimension is assessed through indicators of net income and financial liquidity (Table 4).

#### *4.4. LADA*

The LADA tool (Land Degradation Assessment in Drylands) framework was developed by FAO (Food and Agriculture Organization of the United Nations) for assessing and quantifying the nature, severity, impact and extent of land degradation on ecosystem services across different spatial and temporal scales. In order to support policy decisions to combat land degradation, the framework aims to identify hotspots and bright spots [82]. It is available as a tool-kit (https://www.fao.org/nr/kagera/tools-and-methods/lada-locallevel-assessment-manuals/en/) (accessed on 4 March 2022) that identifies the state of the land resources through different indicators, the pressures and driving forces that caused this status and the impacts on ecosystem services and on livelihoods. The data required are collected through agricultural and other national surveys and censuses and maps of soil and natural resources, as well as digital and computer-assisted methods.

LADA environmental indicators focus on water quality and water use, soil quality and the soil degradation status. It includes an assessment of the irrigation area and the over-exploitation of water resources, as well as the salinization process, and includes indicators focused on general soil threats, including erosion, compaction and nutrient loss. Biodiversity is also tackled through indicators of land cover (Table 2). Additionally, LADA also includes economic indicators related to the economic risk caused by land degradation problems, through the assessment of land loss by fires, urbanization and livestock pressure, among others (Table 4). The sociocultural dimension is represented by the pressures on the resources that will impact society as a whole. The change in land users' life is not investigated. The LADA framework considers climate components illustrated by climate resources and climate extreme events.

#### *4.5. SMART*

The SMART (Sustainability Monitoring and Assessment RouTine) framework was developed by FiBL (Research Institute of Organic Agriculture) to assist farms and enterprises in the food sector for assessing their sustainability level in a credible and transparent manner [83]. The specific software (https://www.fibl.org/en/themes/smart-en/ smart-method) (accessed on 4 March 2022) is used to compute context-specific indicators (up to 200) that are compiled individually for each case study. Data needed for the assessment are semi-quantitative and collected using a standardized interview procedure [84]. The time for data collection is 2–3 h [64]. The software should be handled by scientists and/or field practitioners. The extensive list of indicators includes transversal environmental topics from water pollution to soil quality and degradation, air quality, fertilizer consumption, biodiversity, energy use and even animal welfare. Examples of the broad list of environmental indicators in the framework include pesticide presence in water, greenhouse gas emissions, phosphorus crops content, conservation of species and the use of renewable energy (Table 2). Social indicators are also included in the framework, assessing employees' rights and their wage level for a dignified life. The social dimension also includes gender equality and non-discrimination, cultural diversity, health coverage and access to medical care (. Finally, economic indicators cover a set of themes, from profitability to vulnerability, accountability, the resilience of the investment and the value of local economy (Table 4).

#### *4.6. PG*

PG (public goods) is a framework developed by the Organic Research Centre in the United Kingdom for assessing the provision of a broad range of public goods from farming activities [84]. It is based on the premise that agriculture produces many by-products that are deemed public goods [85].

Information related to the farming activity is gathered and computed in an excel sheet (https://www.organicresearchcentre.com/our-research/research-project-library/ public-goods-tool/) (accessed on 4 March 2022), where 11 individual public goods are scored. Information is collected using questionnaires with several key "activities" and includes qualitative and quantitative data. The analysis is normally undertaken by famers and/or sustainability experts. The time of data collection varies between 2 and4h[84].

Environmental indicators from PG framework include water management and soil quality through the assessment of the irrigation method used, flooding defenses implemented and the existence of water and nutrients management plans, cultivation types and cropland and livestock diversity. Biodiversity and energy use are also tackled extensively through the screening of conservation plans, the presence of habitats and rare species, GHG emissions, energy balance and the correct disposal of farm waste. The animal welfare is accounted through parameters such as housing, biosecurity and their ability to behave naturally (Table 2). The social indicators are basically represented in the job satisfaction through the skills and knowledge of the employees and the contribution to local/regional employment assessed by the level of community engagement. Economic indicators range from financial viability and farm resilience to others, such as accountability by food quality certification, the local economy value through assessing the production of local products and the economic risk by checking landscape features and the management of boundaries (Table 4).

#### **5. Discussion**

#### *5.1. Strengths and Weaknesses of the Frameworks*

5.1.1. SAFA

The study by Landert et al. [83] aimed to transform intensive livestock farming in 15 European countries with a high impact on the environment, society and economy in sustainable livestock farming, which reduces emissions and the costs associated with this. The authors showed that farms with an optimized governance component can improve sustainability in general and that the farmers should learn about this context and improve their production and economic performance within each individual farm. In this context, SAFA is an important tool to provide recommendations for future actions to support achieving sustainability [86].

The study by [87] in the central Sicily Mountains showed that a growing economy would also require more resources to reduce environmental impacts, modernize animal shelters and use renewable energy sources to make them more sustainable. It illustrates how, on the one hand, the sustainability areas that are discussed in SAFA are interconnected, and, on the other, that there are many open pathways for Sicilian organic farms to improve their performance. Although SAFA is a valid asset for addressing the sustainability potential of food in urban system contexts, two main weaknesses related to some subthemes have been pointed out by Landert et al. [83]: (i) the subtheme Remedy, Restoration and Prevention would need a specific adaptation to become food-focused, and (ii) the subtheme Rights of Suppliers does not include the full web of existing relations and processes normally present in these systems. In addition, the subthemes Long-Ranging Investment, Profitability, Stability of Supply, Stability of Market and Liquidity are not flexible for use in this system [65]. In this context, by setting the boundaries of the system, the majority of the indicators became less responsive to drivers or pressures. In turn, this led to poorer analyses of the cost-effectiveness and political and societal acceptance.

#### 5.1.2. RISE

Grenz et al. [88] showed that RISE is an effective tool for field production since it measures fertilizer application relative to soil nutrients and crop requirements for optimum crop growth and calculates the non-renewable energy percentage, as well as the farm financial security (e.g., diversifying income sources, securing access to land, maintaince of infrastructure). Röös et al. [68] observed the potential of this framework to integrate the social dimension of the farm, although some modifications would be necessary to enhance its relevance for the specific context of the study. The authors perceived the results of RISE as highly solid because they are based on quantitative data input and integrate experts on the subject.

RISE becomes complex due to complicated calculations and the elevated number of data required. However, regarding the tool, farmers consider it as relatively simple to understand [68] because of the language adopted compared to the more general one found in in SAFA (e.g., rule of law) [88] and IDEA (e.g., organization of space) [68]. Regarding the relevance of RISE, the farmers recognize that the obtained outcomes reflect the positive and negative points of their farming activities well. Therefore, in comparison to other frameworks (e.g., IDEA, PG), farmers consider RISE as one of the most appropriate frameworks to use [72]. However, it was shown that the time investment and time required for learning RISE are relatively long in comparison to other frameworks [68], while also not being highly transparent as other frameworks due to the complexity of the calculations that complicates the computation rationale behind it [64]. In addition, using standardized quantitative measures makes it hard to capture the specific situation (e.g., farmers' financial situation and working situation), since farming activities will always endorse high variability from one case study to another [68].

Havardi-Burger et al. [72] showed that the process of selecting indicators in RISE becomes difficult since, on the one hand, one must include all of the significant indicators that represent the system well, but, on the other, the number of indicators cannot be too high otherwise it compromises the application of the tool. This aspect is observed for all frameworks except for SAFA, which includes a relatively high number of indicators. In describing this difficulty, Binder et al. [31] refer to parsimony as a principle in order to strive for the system representation under consideration and the sufficiency to address its complexity. Overcoming this difficulty by setting different indicators from different sustainable dimensions and themes is not an easy task since one becomes easily lost on what is actually under study. One possible example is the indicator stability used in RISE to address how financially stable a farm is (e.g., farm infrastructure, long-term access to land, the number of customers and main source of income). The authors showed that covering more aspects would be a benefit, as also shown in the indicator liquidity combining two SAFA indicators (safety nets and net cash flow). This allows the adoption of concrete measures to improve the business performance, even when under financial stress [31].

#### 5.1.3. MASC

MASC can be described as an objective and broad tool. Its ability to incorporate qualitative data in addition to its ease-of-use in terms of the necessary input becomes very helpful for real situations and enables a high comprehensibility of the outputs. Quantitative values can be processed as qualitative information by simply using thresholds, and, thus, MASC integrates both measurements (e.g., yields), calculated data (e.g., semi-net margin) and empirical knowledge (e.g., physical difficulties of crop interventions) into the indicators. This ensures that the best available information is used and that there is a high participation approach, since, as an example, the users' point of view can be integrated in the framework, since normally it would be difficult to address them by using quantitative indicators [63].

Graheix et al. [62] applied MASC to evaluate 31 cropping systems previously chosen to study different management practices, from conventional tillage systems to other systems where conservation agriculture principles were incorporated. In this study, the integrative approach of the MASC framework provided a benefit for the understanding of how the different cropping systems behave when considering, at the same time, (i) the multiple objectives of the dimensions (economic, social and environmental); (ii) various time scales and (iii) the objective worries and goals of the farmers, and generally also the society, raised by different stakeholder groups with various interests. While the results of many studies have highlighted advantages of MASC for adapting cropping systems through conservation agriculture [63], they also identify a weakness in terms of MASCs' inability to properly evaluate the agronomic effects of biodiversity (e.g., normally, a higher biodiversity is an advantage, but decreasing the soil tillage may also contribute to a higher diversity of pests and weeds) from a simple description of the practices employed. The diversification systems may have many advantages (e.g., lower GHG emissions) in comparison with the conventional reference system. They may improve both the air and water quality and contribute to a higher biodiversity [70]. The indicators were initially determined based on scientific knowledge and the context available at the time of the development of MASC, with the aim of keeping its use relatively simple [25]. This probably led to a too generalized meaning of the indicators that cannot highlight the specific context found in different pedoclimatic conditions and under different agricultural management practices of the different studies. As reported by Médière et al. [89], "we still have little scientific information concerning the responses of biological process to agricultural practices in a given pedoclimatic context". The balance between benefits from the services provided and the negative effects that are often observed when tillage is reduced is still unknown, and crop rotation is included, which results in a higher biodiversity [90]. Al Shamsi et al. [91] showed that the best practice reduces the need for off-farm inputs while increasing the product range. However, it is also reported that this diversification can cause negative impacts, i.e., NO3 leaching, NH3 volatilization or pesticide use [70]. When assessing the effect of a combination of different practices in one single indicator, some complexity is added, since this will also be dependent on the pedoclimatic conditions, the intrinsic performance of the system and the goals set for the sustainability performance [70]. Thus, using such frameworks and interpreting its results should be carried out carefully, since there is a high level of subjectivity that cannot be erased [25].

#### 5.1.4. LADA

LADA is a framework that is focused on the following items: biomass production, yearly biomass increments, soil health, water quality and quantity, biodiversity, economic value of the land use and social services of the land and its use [82]. It is also very solid in providing baseline data for improving the land degradation status, offering valid assets to plant, prioritizing and monitoring [92]. The cost-effectiveness is reasonable, i.e., the mapping activity, which includes the land use systems classification, costs approximately USD 250,000 for a country the size of South Africa [92]. This framework also operates with both local and national scales when assessing the land degradation and sustainable land management, cooperating with different stakeholders and proving applicable in at least 18 countries [93]. This is seen as a strength, since the contribution given by different stakeholders (locally and/or nationally) contributes significantly to equilibrated responses and results. For instance, the same status of a land may be classified differently depending on the stakeholder value system [82]. The LADA framework differs from others in its integration of climate factors, which may account for the long-term performance under climate change conditions.

The use of the framework, however, is still rather limited to people with multi-sectoral expertise [92]. This is linked to the need to build a comprehensive database to store both the quantitative and qualitative data obtained during the assessment operations. The assessment should provide a fixed baseline to monitor future changes and trends, and to feed more in-depth knowledge and understanding into the findings of the national assessment for the area in question [94,95]. Reed et al. [93] also states that, in this framework, land degradation assessment and the impact of the soil management practices that could be applicable in each specific situation should be tighter.

#### 5.1.5. SMART

The tool has the advantage of having a high number of indicators to assess the trade-off and synergy analysis. It operationalizes the SAFA guidelines by including indicators that are based on scientific procedures and extensive literature revision. SMART is distinguished from all sustainable assessment frameworks studied by Landert et al. [83] because it integrates the contribution of the stakeholders in its development, which strengthens the acceptance by the end-users while also being specific to local situations [94,96], whereas the others typically involve stakeholders in the application of the framework, but only partly in its development [31]. Therefore, there is a compromise in the intended global applicability of the sustainable assessment tools and the incorporation of a local context.

SMART can be combined with other available tools to improve items such as the system boundary definitions and cut-off criteria when assessing farming activities. The study by Landert et al. [83] used three tools when assessing farm sustainability: COMPAS (an economic farm assessment tool); Cool Farm Tool (a greenhouse gas inventory, water footprint and biodiversity assessment tool, CFT); and the SMART Farm Tool. The results showed that SMART results can be used in combination with quantitative data from COMPAS and CFT. This study was a pioneer in showing the sustainability outcome for 15 different farms in Europe at different stages of their agro-ecological transition. The interdisciplinary tune of this research is characterized by its quantitative contributions and the plurality of view [96]. However, this framework proved to be too time consuming for all of the stakeholders involved, as well as for the interviewers. The combination of SMART with different tools and an improved standard method to incorporate data between the frameworks would facilitate this in the future.

Ssebunya et al. [97] used SMART to assess the sustainability performance of certified organic and fair-trade coffee when compared to non-certified in Uganda. The farm scores were included in the study, which enable analyses of synergies and the trade-off between different sustainable themes. Results showed a link between the certification and the improvement of the sustainable performance of the coffee farms. The framework was also used to enhance the governance objectives by suggesting alterations in group organizations and collective capacities, which, circularly, would also impact other sustainable dimensions. The authors pointed out three main limitations and specific requirements for credible and more consistent outcomes. One of these limitations is related to the comprehensiveness, which is related to the necessary trade-offs for the analysis specificity of some sun-themes. For example, 'Energy Use' and 'Greenhouse Gases' might be more accurately quantified through life cycle assessment methods. Profitability can also be calculated from detailed data from farm incomes and expenditures, whereas this is impossible for other sub-themes. Another limitation is related to the implementation, since the use of SMART requires an adequately trained audit team, involving very time-consuming practice activities to properly understand the functioning of the framework, its indicators and application range. Finally, the team also requires an expertise background on agronomy.

#### 5.1.6. PG

PG is a user-friendly tool, with scores of the indicators coming directly from farmers' answers. One of the strengths of this framework is, therefore, its ease of application. On the one hand, data needed to compute the sustainability assessment are easy to obtain from simple interviews with farmers [85], and the questions include accessible data from the farm accounts and management. On the other, the framework was specifically designed to be simple, which means that input data requirements are modest, and are easily translated in the calculation methods and results [84]. This also implies that relatively little time is required for an assessment, since both manuals are simple to use and questions and calculations are easy to follow. This framework was specifically developed for agri-environmental schemes, making it the best option for policy makers wanting to address questions on whether suggested schemes/subsidies will significantly impact the different sustainability

dimensions. Famers also have a direct answer on the impact that future improvements will have on the provision of public goods [84].

Other strengths of PG include the high level of transparency and the opportunity to transform the results directly into understandable outcomes of public goods in agriculture. Additionally, its user friendliness integrates better farmers and provides a useful tool for them to gain awareness on their sustainability farming activities, which is the first step to adopt better practices [96]. The main weaknesses, however, are also related to the simplicity of the tool, based on qualitative data collection and the lack of quantitative indicators, which allows for subjectivity in the scoring and results. Other more minor weaknesses are related to the presence of terminology related to nature conservation, which can be unfamiliar to farmers, the lack of the possibility to select indicators and the impossibility of including indicators within the framework [96].

#### Scoring the Frameworks

For the environment dimension, RISE, SMART and SAFA show a higher number of indicators covered (seven of eight themes), whereas MASC includes only three themes (water, soil and biodiversity). PG and LADA cover six and four themes, respectively, with water, soil and biodiversity as common themes (Table 5). Although an important subject, climate change seems to be missing in most of the frameworks studied, except in the LADA framework.

In the sociocultural dimension, SAFA is the strongest framework, including nine indicators of a total of twelve themes, followed by SMART covering seven, whereas RISE, MASC and PG cover two themes. SMART and SAFA cover the most important aspects of the sociocultural dimension, whereas RISE assesses only two (workload and wages) and LADA does not assess the sociocultural dimension at the individual level, but rather through land degradation that affects the society as whole (Table 3). In addition to the sociocultural advantages of SMART and SAFA, they enable us to engage stakeholders in different steps in order to increase their acceptance by end-users.

In the economic dimension, SMART, SAFA and PG all cover five themes out of six, followed by RISE and MASC with two themes each (profitability and vulnerability), and LADA with one theme (economic risk). SMART and SAFA assess all themes of the economic dimension except economic risks, whereas PG excludes only the investment theme. Despite the low number of economic themes included, farmers perceive RISE and SMART as the most indicated frameworks for understanding the level of sustainability achieved in their farm because they are based on quantitative data, which are then used for specific contexts [63,64].

In summary, SAFA is the framework with more focus on sociocultural aspects, while still covering some environmental and economic themes. SMART is also homogenous, and covers all three dimensions, but with fewer themes in each one in comparison to SAFA. In contrast, LADA does not include the sociocultural dimension at the individual level and is focused on the environmental dimension. The same is true to some extent for RISE and PG, which include few themes of the sociocultural dimension, while being focused on the environment and /or economy, respectively (Figure 1).

#### *5.2. Which Frameworks Should Farmers Select?*

To help stakeholders decide which framework is the most suitable for their sustainable assessment, we have developed a decision tree based on possible scales, sectors of applicability and the completeness of sustainability dimensions required (Table 6). For global assessments, there are both SAFA and LADA, but SAFA differs from LADA in assessing food systems in addition to land degradation. In addition, SAFA covers all dimensions, whereas LADA excludes the sociocultural dimension at the individual level, and it includes only a few economic themes.


**Figure 1.** Total number of environmental, sociocultural and economic indicators used in each

framework under study: RISE, MASC, LADA, SMART, SAFA and PG.

**Table 6.** Decision tree according to the framework scale assessment (global/local), sector of application (cropping system, livestock system, forestry system, urban system and food sector) and completeness of sustainability assessment (environmental, economic and sociocultural dimensions). Icons in black represent a higher number of themes whereas grey represent a lower number of themes in each dimension. Strengths and weaknesses related to the user-friendliness of the tool and the use of qualitative/quantitative data are also mentioned.


When the stakeholder intends to perform a sustainability assessment on a farm level, he/she has four choices: RISE, PG, MASC and SMART, but the latest only covers the cropping sector and is rather limited in the number of themes covered. The other three frameworks include cropping and livestock systems, whereas SMART also includes the food sector, which is the only possible choice if that is the user's goal. The selection between RISE, PG and SMART depends on the level of completeness intended for the analysis. SMART covers all dimensions, but with fewer themes in each dimension, whereas the other two include more themes in the environment and economy, respectively. However, MASC and RISE are more complex frameworks, whereas PG is the most user friendly and accessible for farmers.

#### **6. Summary and Conclusions**

The comparison between the six sustainability assessment frameworks (SAFA, RISE, MASC, LADA, SMART and PG) showed that they have different characteristics with regard to their assessment methodologies, time and data requirements to operate, and different outcomes with a different accuracy and level of complexity. Balancing all of these aspects in the development of the sustainability frameworks in order to meet the expectations of the main actors has proven to be a challenging task.

The high variety of characteristics of each sustainability frameworks derives from the fact that they were developed to serve different end-users: (i) farmers for assessing their farm performance; (ii) advisories and technicians for advising farmers on how they can improve their sustainability; (iii) researchers who conduct comprehensive regional and local assessments adaptable for context-specific conditions by combining, for example, different indicators from different frameworks.

The six sustainability assessment frameworks were compared according to their ability to cover the main themes of environmental, economic and sociocultural dimensions, and their themes were reported. We have also developed a decision tree based on possible scales, sectors of applicability and the completeness of sustainability dimensions required to help stakeholders decide which framework is the most suitable for their sustainable assessment purposes.

This overview study reveals that a multi-actor approach is necessary to enable the acceptance of the outcomes and their adoption by the main actors (i.e., farmers). When a value judgement is incorporated into a framework without involving farmers (e.g., assuming that organic farming will be more sustainable), the results may become irrelevant and are not considered useful by them [58,98,99].

It might be difficult to include alterations occurring in climatic, environmental, socioeconomic or technological dimensions, in both the short- and/or long-term in the agricultural and societal aspects, but it may also offer new opportunities for more sustainable development [100]. Therefore, assessing the long-term performance under climate change conditions should be addressed further while assessing agricultural sustainability. For this purpose, realistic climate scenarios should be included.

**Author Contributions:** Conceptualization and writing—original draft preparation, A.A., L.B. and C.S.S.F.; review and editing, R.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is a part of the EU—H2020 project entitled 'Soil Care for profitable and sustainable crop production in Europe' (SoilCare), contract number 677407, which aims to identify and evaluate promising soil improving CS and agronomic techniques, increasing profitability and sustainability across scales in Europe.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We would like to thank the handling editor as well as the three reviewers for their constructive comments which considerably improved this paper.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Opportunities for Mitigating Soil Compaction in Europe—Case Studies from the SoilCare Project Using Soil-Improving Cropping Systems**

**Ilaria Piccoli <sup>1</sup> , Till Seehusen 2,\*, Jenny Bussell 3, Olga Vizitu 4, Irina Calciu 4, Antonio Berti 1, Gunnar Börjesson 5, Holger Kirchmann 5, Thomas Kätterer 6, Felice Sartori 1, Chris Stoate 3, Felicity Crotty 7, Ioanna S. Panagea <sup>8</sup> , Abdallah Alaoui <sup>9</sup> and Martin A. Bolinder <sup>6</sup>**


**Abstract:** Soil compaction (SC) is a major threat for agriculture in Europe that affects many ecosystem functions, such as water and air circulation in soils, root growth, and crop production. Our objective was to present the results from five short-term (<5 years) case studies located along the north–south and east–west gradients and conducted within the SoilCare project using soil-improving cropping systems (SICSs) for mitigating topsoil and subsoil SC. Two study sites (SSs) focused on natural subsoil (>25 cm) compaction using subsoiling tillage treatments to depths of 35 cm (Sweden) and 60 cm (Romania). The other SSs addressed both topsoil and subsoil SC (>25 cm, Norway and United Kingdom; >30 cm, Italy) using deep-rooted bio-drilling crops and different tillage types or a combination of both. Each SS evaluated the effectiveness of the SICSs by measuring the soil physical properties, and we calculated SC indices. The SICSs showed promising results—for example, alfalfa in Norway showed good potential for alleviating SC (the subsoil density decreased from 1.69 to 1.45 g cm−1) and subsoiling at the Swedish SS improved root penetration into the subsoil by about 10 cm—but the effects of SICSs on yields were generally small. These case studies also reflected difficulties in implementing SICSs, some of which are under development, and we discuss methodological issues for measuring their effectiveness. There is a need for refining these SICSs and for evaluating their longer-term effect under a wider range of pedoclimatic conditions.

**Keywords:** degree of compaction; soil penetration resistance; relative normalised density; air-filled porosity; tillage; straw incorporation; bio-drilling crops; subsoiling; crop productivity

**Citation:** Piccoli, I.; Seehusen, T.; Bussell, J.; Vizitu, O.; Calciu, I.; Berti, A.; Börjesson, G.; Kirchmann, H.; Kätterer, T.; Sartori, F.; et al. Opportunities for Mitigating Soil Compaction in Europe—Case Studies from the SoilCare Project Using Soil-Improving Cropping Systems. *Land* **2022**, *11*, 223. https:// doi.org/10.3390/land11020223

Academic Editor: Amrakh I. Mamedov

Received: 16 December 2021 Accepted: 31 January 2022 Published: 2 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Soil compaction (SC) is a form of physical degradation due to the disruption of soil micro- and macro-aggregates, which are deformed, reduced in volume, or destroyed under pressure. Compaction is a "hidden" threat that occurs belowground and is one of eight European soil threats [1], affecting as much as 18 to 36% of croplands [2,3]. There are several consequences of SC because of its influence on many important soil functions. For example, it can negatively affect physical soil properties, such as gas permeability and water infiltration and storage [4,5]. This hampers the ecological function of the soil, leading to reduced soil fertility and crop production [6–8]. Furthermore, SC problems can reduce water infiltration and, in addition to causing problems with runoff and erosion, the soil workability may be reduced due to high water content, and the crops may not be able to explore the entire growing season (e.g., delayed seeding date) [9–11]. In this regard, the climate and the expected climate changes are important; for example, Northern Europe may be subject to increasing precipitations and wetter conditions during the growing season [12]. Indeed, soil compaction is one of the main reasons for stagnating yields [10,13]. A study also showed that even if SC does not necessarily lead to a reduction in yields, it can cause considerable amounts of extra costs not only for the farmers but also for society [14].

There are several common reasons for SC in most European countries. Compaction may occur in both the topsoil (i.e., arable layer) and subsoil layers (i.e., below the arable layer) due to pressure from the passage of machinery and repeated trampling of grazing animals, or occur naturally from previous geological periods during the initial ground formation under land ice. Subsoil compaction is also associated with in-furrow ploughing, during which tractor wheels that are in direct contact with the subsoil transmit the pressure to deeper soil horizons, especially when using heavy machinery under wet and suboptimal soil conditions [15]. Unlike topsoils, subsoils are not loosened annually, and compaction may become cumulative [16,17]. Another feature regarding the SC of subsoil is the formation of a plough pan layer that results from repeated ploughing and is less permeable for roots and limits water flow and gaseous exchange. Ruser et al. [18] report that compaction can become almost irreversible once it reaches the threshold of the preconsolidation stress (i.e., the index of soil load-bearing capacity).

Even though certain climatic conditions and processes (i.e., drying/wetting or freezing/thawing and shrinking cycles) can be effective in counteracting the SC of clayey soils [19,20], these processes are mostly absent on silty soils, making them especially susceptible to subsoil compaction [21]. While ploughing is effective for loosening up compaction of the upper soil layers, there is a lack of measures for persistently loosening up the subsoil [22]. There is a need for developing strategies to avoid subsoil SC and to stabilise and improve subsoil structure. For example, plant roots can be effective for loosening up subsoil, a strategy referred to as "bio-drilling" where roots modify the soil structure by pushing aside soil particles, thereby creating large pores that improve both hydraulic conductivity and gas flow [23–26]. Cresswell and Kirkegaard [27] defined bio-drilling as the creation of bio-pores by deeply penetrating taproots as low-resistance pathways for the roots of a succeeding crop. For this purpose, alfalfa, forage radish, or oilseed crops, which are known for having deep taproot systems, may be efficient for improving the soil structure even deeper in the soil profile [24,28,29]. However, the understanding of optimising the effect of bio-drilling crops through appropriate management remains limited, and their effects on crop yields vary with climatic conditions [29].

Mechanical subsoil loosening, referred to as deep loosening, deep ripping, or subsoiling, is a common practice to loosen up dense soil layers below the topsoil [30,31]. Subsoil loosening can decrease penetration resistance and bulk density [32] and increase infiltration [33], root development [34], and crop yield [35–37]. There is a need for loosening subsoil under optimal soil moisture conditions. When the soil is too wet and loose, the soil might be smeared and compacted [38,39]. When the soil is too dry, thick clods are formed [39]. Furthermore, the benefits of subsoiling are often not long-lasting due to re-compaction by the overburden topsoil and field operations [40–42]. However, when

combining mechanical subsoil loosening with the addition of organic materials into the subsoil, loosening may last for several years [43,44].

Great efforts have been made to quantify SC, which is needed both for identifying SC problems and for evaluating the effectiveness of mitigating strategies. For instance, Huber et al. [45] suggested the following indicators: soil bulk density, air and water permeability, mechanical resistance, and a visual assessment of the soil structure and rooting. The proposed indicators involve several common measurements, such as bulk density and penetration resistance. However, suitable definitions of critical limit values linked to crop impairment are difficult to define. A number of penetration resistance threshold values above which rootability is impaired [46] can be found. They range between 1 and 2 MPa or higher [47–54] and are strictly linked to pedoclimatic conditions and soil management (e.g., tillage vs. no-tillage). Similarly, for SC characterization, Håkansson [8] suggested an index of the degree of compaction (DC). The DC index represents the bulk density-to-reference density ratio and is considered detrimental for crop development when it exceeds 87% [8]. Although the DC is a fast and easy index, two issues have recently been raised—the identification of the correct reference bulk density is not obvious and the 87% threshold seems to not be applicable for all pedoclimatic conditions [55].

Compaction is one of the threats included in the EU "SoilCare" project (soil care for profitable and sustainable crop production in Europe). This project addressed the use of different soil-improving cropping systems (SICS) involving both the crop type and rotation, as well as specific management techniques aiming to improve soil quality and functions (http://soilcare-project.eu/, accessed on 1 January 2022). In this paper, we present the main outcomes from five case studies within the SoilCare project using different SICSs to counteract compaction. The study sites (SSs) were located in five European countries, where we investigated different innovative strategies for mitigating SC under various soil and climatic conditions. The SICSs involved different types of tillage, including subsoiling and various deep-rooted bio-drilling crops.

#### **2. Materials and Methods**

All SICSs had a common objective—to counteract soil compaction. They were located in five countries along the north-to-south and east-to-west gradients from Norway to Romania (Figure 1 and Supplementary Materials, Table S1).

**Figure 1.** Location of the five study sites involved in the present study.

The SICSs examined in each country for alleviating topsoil and subsoil SC comprised the use of various deep-rooting crops and different types of tillage operations, including subsoiling (Table 1). At all SSs, the SICSs were compared with a reference standard practice, and both topsoil and subsoil samplings were made at different depths according to the soil characteristics for each of the experiments (Supplementary Materials, Table S2). Although exactly the same measurements were not performed at all SSs (as detailed below), some were similar for all SSs. This allowed us to make a generic analysis and identify the relationships with soil properties using the three SC indices described in Section 2.4.

**Table 1.** Soil-improving cropping systems (SICS) applied at the five study sites (SSs) and the reference standard practice at each site.


#### *2.1. Norway*

#### 2.1.1. Experimental Design

The Norwegian SS investigated soil compaction alleviation by using bio-drilling crops. The soil was characterised by poor natural drainage and medium erosion risk. This field has been under cultivation for several decades and the site was drained. In the early summer of 2015, a multiple wheel-by-wheel approach was used for establishing the initial compaction with a tractor and trailer combination passing across the plots ten times, with a total weight of 17 Mg and resulting in a wheel load of 2.8 Mg for the trailer tandem axles (compacted "C" plot). This is a typical wheel load for small- and medium-sized farms in Norway and representative of other machinery, such as a combine harvester. There was little precipitation the days before the compaction treatment and none during it, resulting in workable conditions and higher soil moisture tension in the topsoil and subsoil (−25 and −63 kPa, respectively) than assumed at the field capacity (−10 kPa) while wheeling. The site was used for yield studies until 2017; for further details, please see the work of Seehusen et al. [56]. Thereafter, four different rotation treatments were applied during a 4-year period—(1 and 2) oilseed (*Brassica rapa* L. ssp. Oleifera) and barley (*Hordeum vulgare* L.) rotation, (3) barley monoculture, (4) alfalfa (*Medicago sativa* L.) monoculture. The experimental design was a split plot with two replicates, with the compaction level as the splitting factor (compacted "C" vs. reference "R" plot) and the rotation treatment (1 to 4) as the main plot factor. Crops were grown in 5 × 1.5 m plots for a total of 16 plots, that is, four rotations × two compaction levels (compacted vs. uncompacted) × two replicates.

All plots were subject to spring ploughing at 25 cm beginning in 2015 (after the compaction) except for the plots with perennial alfalfa. The ploughing was assumed to be effective for alleviating compaction and aligning the root effects, and therefore, only the topsoils from Treatments 3 and 4 were sampled in 2020. Management practices (seeding, fertiliser, and tillage) were done in the same way as the surrounding fields.

#### 2.1.2. Soil Sampling, Field Measurements, and Laboratory Analysis

Undisturbed cylinder cores (100 cm−3) were collected at both 10–20 and 40–50 cm depth in 2015 (*n* = 4–5 per depth) and 2020 (*n* = 4 per treatment and depth) for soil physical analysis. The soil bulk density (BD) was determined gravimetrically by weighing the soil before and after drying for 24 h at 105 ◦C. In 2015, the BD represented the field conditions at sampling, while in 2020, it represented the BD at −3 kPa. The water retention was studied in both years by first saturating the samples and then draining them at different matric potentials (−3, −50, and −1500 kPa in 2015, and −2, −10, −100, and −150 kPa in 2020). In the latter year, the wilting point (−1500 kPa) was calculated using a pedotransfer function [57]. The pore size distribution was derived from the water retention curves (details in Seehusen et al. [56]). The air capacity was measured assuming a field capacity of −10 kPa by measuring the airflow through the soil samples at a pressure of 2 kPa [58]. Saturated hydraulic conductivity (Ks) was determined with the hood permeameter method [59] on saturated soil samples in 2015, while in 2020, it was derived from the air permeability according to Riley's pedotransfer functions [57] (for further details, see Seehusen et al. [60]).

In 2015, the data from the compaction trial were analysed using the R statistical software package (2014) (details in Seehusen et al. [56]). In 2020, the data for the Norwegian SS were analysed using general linear models and Fisher tests in Minitab 19. Comparisons between the values from 2015 and 2020 were done by two-sample t-tests and confidence intervals. The results after Treatments 1–4 from 2020 of the reference plot (no compaction) were compared with the reference values from 2015, and the results after Treatments 1–4 from 2020 of the earlier compacted plot were compared with the values of the compacted plot from 2015.

#### *2.2. Sweden*

#### 2.2.1. Experimental Design

The SS in Sweden is located at a farm in southern Sweden, where the subsoil is naturally compacted (1.7–1.9 g cm−3) due to its formation under land ice and the root growth of crops is restricted to the topsoil, with hardly any roots below 30 cm [61,62]. The site has been under cultivation for at least a century and is tile-drained. The experiment consisted of a pilot study starting in September 2018 that investigated the possibility of improving the upper subsoil through the supply of undecomposed organic material in combination with a mechanical subsoil loosening. A randomised block design (12 plots, 6 × 20 m) with four replicates was established, involving three treatments—(a) a control treatment, (b) loosening of the subsoil (to a depth of about 35 cm) without the incorporation of organic material, and (c) loosening of the subsoil with the incorporation of undecomposed straw pellets at amounts of about 25 Mg ha<sup>−</sup>1. Subsoiling and straw incorporation were performed using adapted HE-VA sub-tiller equipment at a speed of 1 km per hour to 24–35 cm depth. Straw pellets were pumped from a tank mounted on the front of the tractor and injected under pressure into the upper subsoil through oval openings in metal pipes welded behind each vertical bill. The loosening of the subsoil and the addition of straw pellets was performed only once (in 2018). Thereafter, normal tillage practices, including mouldboard ploughing to 25 cm, were applied in all plots. Crop fertilisation followed the local recommendations. Winter wheat (*Triticum aestivum* L.) and sugar beet (*Beta vulgaris* L.) yields were recorded in the 2019 and 2020 growing seasons, respectively.

#### 2.2.2. Soil Sampling, Field Measurements, and Laboratory Analysis

In 2019, at the end of the winter wheat heading (growth stage Z60 according to the Zadoks scale), a soil profile description was conducted in one plot per treatment. The portions of the upper subsoil (24–35 cm) volume and surface affected by subsoiling and the presence of roots were visually evaluated. A more detailed soil sampling was done in 2020 about six weeks before harvest within a small area in the middle of each plot that was kept free from sugar beet plants starting around mid-summer. In this area, a soil pit 65–75 cm long and 25 cm wide was dug, and six undisturbed soil cylinders (7.2 cm diameter, 5 cm height) were taken at 10–15 cm depth, as well as six at 28–33 cm depth, by placing the cylinders one after the other in a row at a spacing of about 5 cm between each. Before removing the cylinders from the 28–33 cm depths, six penetration resistance (PR) tests were collected along the row with cylinders. On the same occasion, a soil profile for the control plots (only) was obtained using an auger sampling with depths divided into increments of 0–20, 20–22.5, 22.5–25, 25–27.5, 27.5–30, 30–35, and 35–40 cm. Each increment was analysed for the total C and N, and the pH and soil texture were measured in the 0–20 cm layer and in a combined sample for the 25–40 cm depth.

The soil moisture content and dry soil bulk density were determined from each of the cylinders. Each of the cylinder samples was also passed through a 2-mm sieve, and the occurrence of gravel and small stones (i.e., particles >2 mm) was determined by measuring both their weight and volume fraction. Thereafter, the total C and N concentrations were measured by dry combustion, and the pH (water) was determined for each sample from the 28–33 cm depth; only a pooled sub-sample for the 0–15 cm depth was retained for these analyses.

All statistical analyses were done with the GLM Procedure in SAS software (SAS Institute, Cary, NC, USA). The means were compared using Fisher's least significant difference (LSD) when the F-value in ANOVA was statistically significant (*p* < 0.05).

#### *2.3. United Kingdom*

#### 2.3.1. Experimental Design

The United Kingdom SS is located at the Allerton Project—a 300 ha mixed arable and livestock research, demonstration, and education Farm. The experiment aimed to examine the alleviation of compaction by using tillage. The experiment started in October 2017 at the Allerton Project. This SS historically used a wheat–rape (*Brassica napus* L.) rotation with a "break" spring crop, and over the last ten years, had a reduction in tillage, going from a plough-based system to direct drilling. Soil compaction was artificially created by driving a tractor (Massey Ferguson 7720, approx. 8 tons total weight) across the area, ensuring a tractor tyre was running over the whole plot twice. Directly afterwards, measurements with a penetrometer verified the degree of compaction, showing the average compaction was 15% higher to a depth of 45 cm, with the highest compaction (+32%) occurring at 7.5 cm depth. The experimental design was a randomised complete block design with 3 replicates involving a total of 6 plots (9 m wide and 40 m long). The ploughing system (20 cm depth) was compared with a no-cultivation direct-drilled control treatment. Following the fall cultivations in 2017, winter barley grew across all plots and was harvested in July 2018. The compaction and treatments were repeated in October 2018, keeping the same plot structure, and faba beans (*Vicia faba* L.) were planted across all plots and harvested in September 2019. In March 2020, spring wheat was planted across all plots and harvested in October 2020.

#### 2.3.2. Soil Sampling, Field Measurements, and Laboratory Analysis

The measurements of BD and penetration resistance (PR) were split into topsoil (0–25 cm), which was within the cultivation depth of the plough, and subsoil (>25 cm), which was below the depth of cultivation. PR measurements were conducted in 2020 after crop drilling in May using a field penetrometer (Field Scout, SC900) to a depth of 45 cm, with 10 measurements taken per plot and averaged. The bulk density was also measured in May 2020 using a soil cylinder (196 cm3) in the topsoil and subsoil layers. The soil was dried for 48 h at 105 ◦C and weighed to calculate the bulk density. Soil samples were also collected from the topsoil layer and the particle size distribution and soil organic carbon were analysed. Infiltration was measured using the double ring method (outer ring diameter of 53 cm, inner ring diameter of 28 cm diameter, water depth of 24 cm). Both rings were partially buried in the soil and the outer ring was kept topped up with water to prevent lateral leaking. Once the water loss reached a stable rate, the water loss from the inner ring was recorded over time and converted to saturated hydraulic conductivity (Ks). The crop yield was measured at harvest each year by taking a reading from the combine after each plot was harvested.

Differences between treatments were analysed using Genstat version 18. A general linear model was used, with blocking treated as a random effect in all analyses. Where topsoil and subsoil measurements were both included in the analysis, a split-plot design was used, with the sample depths of the split-plot and treatment (plough vs. direct-drill) as the main plot effects.

#### *2.4. Italy*

#### 2.4.1. Experimental Design

The Italian SS aimed to prevent soil SC by combining no-tillage and cover crops. In the area, the shallow water table ranged from about 0.5–1.5 m in late winter to early spring to 1–2 m in summer. The experiment has been ongoing since 2018 and has a splitplot design (12 plots in total, 12 m wide × 85 m long) with two replicates, two levels of tillage intensity (main plot), and three levels of soil cover (sub-plot). The no-tillage (NT) system based on sod seeding was compared with the conventional practice (CT) based on mouldboard ploughing to 30 cm, followed by disk-harrowing to 15 cm. The main crop was maize (*Zea mays* L.), while during fall, the soil remained bare (BS) or was covered with cover crops, for example, winter wheat (WW) or tillage radish (*Raphanus sativus* L.) (TR), which are characterised by fibrous and taproot root systems, respectively. Subsurface band fertilisation was applied at sowing in NT, while side-dressing fertiliser was followed by hoeing in the CT treatment. Pesticide applications depend on the crop requirements but were the same for all the plots.

#### 2.4.2. Soil Sampling, Field Measurements and Laboratory Analysis

Soil samples were collected before seedbed preparation in spring 2020 from the topsoil (i.e., tilled layer) and subsoil (i.e., below the tilled layer), as reported in the Supplementary Materials, Table S2. Undisturbed soil cores (7 cm in diameter, 60 cm in height) were collected with a hydraulic sampler and cut to extract the 0–20 and 40–60 cm soil layers. Remoulded soil samplings were collected at the same depth for chemical–physical analysis. PR measurements were performed up to 60 cm depth before tillage operation (at the end of February), with a digital cone penetrometer (Eijkelkamp, Giesbeek, The Netherlands) with a base area of 2 cm<sup>2</sup> and an apex angle of 30◦. Undisturbed soil cores were oven-dried at 105 ◦C for 24 h to calculate the volumetric water content (VWC) and BD using the core method [63]. Remoulded soil samples were air-dried, sieved at 2 mm, and analysed for particle size distribution according to the methods by Bittelli et al. [64] and the soil organic carbon concentration (SOC). On-field soil hydraulic properties were measured inside each plot using a double-ring infiltrometer (inner ring diameter of 60 cm, outer ring diameter of 80 cm) according to the methods by Parr and Bertrand [65]. The hydraulic conductivity (Ks) and sorptivity were calculated by applying Philip's infiltration equations [66]. At the end of the growing season, the maize grain yield was collected at the commercial moisture content from four representative areas (2 m2) in each plot and then dried at 65 ◦C until a constant weight was obtained to determine the dry weight.

The data were analysed by applying a linear mixed-effect model based on the restricted maximum likelihood estimation method considering tillage, soil covering, and their interaction as fixed and block as random factor. Post-hoc pairwise comparisons of the least-squares means were performed using the Tukey method to adjust for multiple comparisons (*p* < 0.05). Statistical analyses were performed with SAS software (SAS Institute Inc. Cary, NC, USA), 5.1 version.

#### *2.5. Romania*

#### 2.5.1. Experimental Design

The Romania SS is located in an area characterised by natural subsoil compaction. The experiment consisted of a pilot study established in March 2018, and its aim was to mitigate natural SC by tillage. The experimental design was a split plot (36 plots, 6 × 33 m) with three blocks and involving four treatments—(TR1) mouldboard ploughing with furrow inversion to 25 cm depth, (TR2) subsoiling to 60 cm by ripping and disking to 12 cm depth, (TR3) a control treatment with 2-times disking, and (TR4) chiselling to 25 cm depth with furrow inversion. All treatments were repeated every year. The testing of tillage treatments

also involved three rotations with deep-rooting leguminous crops. Only the main effect of the tillage treatments on the soil physical properties is reported here.

#### 2.5.2. Soil Sampling, Field Measurements and Laboratory Analysis

Soil physical and chemical parameters were measured in all plots during the three years of the experiment. For this, disturbed soil samples were collected in autumn after crop harvesting for soil water-stable aggregates (WSA) >250 um, and undisturbed soil cores (100 cm3 volume) were sampled at 10–20 cm and 40–50 cm depths for soil physical analyses (Ks and BD).

The content of water-stable aggregates (in % g/g) was measured by the Henin– Feodoroff method based on wet sieving (SR EN ISO 10930:2012). The Ks was determined according to the steady-state falling head method (Romanian standard: STAS 7184/15–91). The BD was gravimetrically determined by weighing the soil core samples before and after drying for 24 h at 105 ◦C (SR EN ISO 11272:2017).

The data obtained for the soil properties measured at the Romanian SS were analysed by one-way repeated measure ANOVA considering either the soil tillage or year as the tested factor. Post-hoc pairwise comparisons of the least-squares means were performed using the Tukey method to adjust for multiple comparisons (*p* < 0.05). All statistical analyses were performed with OriginLab 6.1 software (Origin Lab Corporation, Northampton MA, USA).

#### *2.6. Soil Compactions Indices*

The effects of the SICSs across the different SSs were investigated using three soil compaction indices—degree of compaction (DC), relative normalised density (RND), and air-filled porosity (AFP). The DC was calculated as follows:

$$\text{DC} = \text{BD/BD}\_{\text{ref}} \times 100 \tag{1}$$

where BD is the bulk density and BDref is the reference bulk density. The BDref was calculated according to Equation (12) reported by Keller and Håkansson [67], as follows:

$$\begin{array}{l} \text{BD}\_{\text{ref}} = 1.308 + 0.0119 \text{ clay} + 0.0103 \text{ sand} + 0.00018 \text{clay}^2 - 0.00008 \text{sand}^2 \\ \quad - 0.00062 \text{siltOM} - 0:00059 \text{sand} \text{OM} \end{array} \tag{2}$$

where OM is the soil organic matter. The RND index was derived from the ratio between the BD and the critical bulk density (BDcrit), the latter being 1.6 g cm−<sup>3</sup> for soils with clay < 16.7% or calculated with the following equation for soils with clay > 16.7% [68]:

$$\text{BD}\_{\text{crit}} = 1.75 - 0.0009 \times \text{clay} \tag{3}$$

The air-filled porosity (AFP) at the sampling was calculated as the difference between the total porosity and the volumetric water content.

#### **3. Results and Discussion**

#### *3.1. Norway*

In the topsoil, there were no significant differences in the BD, TPV, or AC between the treatments in 2020 or between years, with average values of 1.35 g cm<sup>−</sup>3, 47.6%, and 10.8%, respectively. Similarly, there were no significant differences between the treatments for Ks and air permeability in 2020. However, for both treatments, the Ks and air permeability were significantly lower in 2020 compared to 2015 (Supplementary Table S3).

In the subsoil, multiple wheeling in 2015 led to a significant increase in BD, with 1.69 g cm−<sup>3</sup> in C as compared to 1.59 g cm−<sup>3</sup> in the R plots (Table 2). Five years after the compaction event, the BD was still significantly higher in the C than in the R plots in 2020, with the exception of Treatments 2 and 4. In the uncompacted R plot, the BD in 2020 was significantly decreased compared to 2015 for all treatments, from 1.59 to 1.45 g cm<sup>−</sup>3. In the

C plot, the BD was also significantly reduced after 5 years for Treatment 4 (1.45 g cm−3), reaching the same level as Treatment 4 in the uncompacted R plot (1.44 g cm−3). There was also a trend towards a reduction in the BD after 5 years for Treatment 2, from 1.69 to 1.55 g cm−3. Compared to the topsoil (Supplementary Materials, Table S3), the BDs in the subsoil for the R plot were about 20 and 10% higher in 2015 and 2020, respectively (Table 2). The subsoil TPV significantly decreased by about 6% in 2015 following the compaction event (Table 2). In 2020, there were no significant treatment effects on the TPV, which was 45.5% on average. However, both Treatment 1 (+5.1%) and Treatment 4 (+7.0%) led to a significant increase in the TPV on the C plots in 2020 compared to 2015.

**Table 2.** Subsoil (30–40 cm) bulk density (BD), total pore volume (TPV), air capacity (AC), saturated hydraulic conductivity (Ks), and air permeability (Air perm) in uncompacted reference and compacted plots at the Norwegian study site in 2015 and for the different treatments in these plots in 2020. Treatments 1, 2: oilseed rape–barley rotation; Treatment 3: barley monoculture; Treatment 4: alfalfa monoculture.


Mean ± standard deviation (2015 *n* = 5, 2020 *n* = 4). For 2015, values followed by different letters are significantly different. For 2020, different letters after values indicate significant differences between the treatments within reference (R) and compacted (C) plots. ns= not significant at *p* < 0.05. ¶ indicates a significant difference between 2020 and the value in 2015 for each treatment (i.e., in R or C plots).

In contrast, multiple wheeling in 2015 had no significant effect on the subsoil AC. In the uncompacted R plot, the AC was an average of 5.81% after 5 years and there were no significant differences between the treatments, while in the C plot, Treatment 3 presented the lowest increase of all treatments in 2020 (from 3.38 to 5.26%). With the exception of Treatment 3 on the R plot and Treatment 1 on the C plot, the AC values significantly increased during the research period, from 3.4 to 5.9% on average across treatments.

Similarly to the TPV, soil compaction in 2015 did not lead to a significant reduction in either the saturated hydraulic conductivity or air permeability (Table 2). In 2020, Ks followed a similar pattern to air permeability since it was estimated using a pedotransfer function based on air permeability. In both the R and C plots, there were no significant differences between the treatments in 2020 in either the Ks or air permeability, which were 0.15 m day−<sup>1</sup> and 1.2 um<sup>2</sup> on average, respectively. Compared to 2015, there was a significant reduction in Ks by 5.40 ×10−<sup>3</sup> m day−<sup>1</sup> in the C plots for Treatment 3, while there was a significant reduction in the air permeability for Treatments 1–3 in the C plot, with an average of 23.9 um2. Compared to the topsoil (Table 2), there were very large differences regarding both the Ks and air permeability in the subsoil for the R plot in 2015 and 2020.

There was only a significant difference between the same treatment in the R and C plots in 2020 for BD (Treatments 1 and 3) and not for any of the other soil physical properties (data not shown).

#### *3.2. Sweden*

The soil visual assessment showed that the straw was not mixed with the subsoil in rows but located at the bottom of the subsoil rows created by the bills in the subsoiling + straw treatment (Figure 2a). Indeed, the two subsoiling treatments forced the topsoil into the subsoil, forming distinct rows, while the subsoil moved into the topsoil irregularly (Figure 2a). However, subsoiling affected only a portion of the upper subsoil layer (24–35 cm) below the topsoil. We evaluated that the volume of the subsoil affected by the subsoiling treatments varied between 36 and 40% and that the surface of the subsoil affected varied between 42 and 49%. Analysis of the soil profile samples for the control plots that characterised the experimental site more precisely showed that the sand, silt, and clay contents were 62, 27, and 11%, and 64, 27, and 9% in the topsoil and subsoil layers, respectively.

**Figure 2.** (**a**) Illustration (top) and photo (bottom) of a Swedish soil profile used for evaluating the effects of the subsoiling + straw treatment. (**b**) Changes in the penetration resistance with depth in 2020, a metric used for evaluating the effects of subsoil loosening and loosening + straw incorporation treatments at the Swedish study site. The vertical line (2.5 MPa) indicates the critical limit for root penetration. Data are mean values of six measurements made across treatment stripes covering a width of about 40 cm in each experimental plot.

As shown by the visual assessment for the presence of roots in 2019, which was done by counting the number of roots along a 10-cm line at two depths in the topsoil (10 and 20 cm) and in the subsoil (30 cm), there were more roots present in the subsoiling treatments at the 30 cm depth. Meanwhile, there were almost no roots present in the subsoil for the control treatment, and the subsoiling + straw treatment also appeared to improve the number of roots at all three depths compared to the control (Supplementary Materials, Table S4). The mean maximum root penetrations into the subsoil (>24 cm) were about 4 cm in the control and 11 cm in the subsoiling treatments. The maximum penetrations were more variable for the subsoiling treatments since among the six measurements made within each plot, some presented values similar to those for the control, but some values were much deeper, indicating the measurements were sometimes penetrating the subsoil rows created by the bills (data not shown). The measurements in the control plots almost never exceeded 6 cm. The changes in the soil penetration resistance with depth in 2020 showed a mean maximum (i.e., exceeding the 2.5 MPa critical limit for root penetration) rooting depth of about 28 cm in the control, almost 30 cm in the subsoiling alone, but much deeper at around 40 cm for the subsoiling + straw treatments (Figure 2b).

There were no significant differences between the SOCs in the topsoil (10–15 cm) and the subsoil (28–33 cm) cylinder soil samples (Supplementary Materials, Table S4). The soil total C/N ratios, as well as the pH values in the top- and subsoils, were also not significantly different between treatments at around 10.0 and 6.0, respectively.

Compared to the subsoiling + straw treatment, the BD was significantly higher in the topsoil in the subsoiling treatment. It was higher also in the subsoil compared to the control (Supplementary Materials, Table S4). However, when correcting the BD for the presence of gravel and stones [69], which varied between 6.1 and 8.3%, there were no significant differences between the treatments in either the top- or subsoils. Since this site had a naturally compacted subsoil with high soil densities, we were restricted to using smaller cylinders than usual (i.e., 204 vs. 408 cm<sup>−</sup>3), which provided less precise and more variable measurements. The experimental site was also heterogeneous, and there was a negative correlation between the SOC contents and the BDs (Supplementary Materials, Figure S1). This may have contributed to the differences because the subsoil SOC content in the control was slightly higher than for the subsoiling treatments, even if not significant.

#### *3.3. United Kingdom*

The soil BD was not affected by the compaction alleviation treatment and showed no significant difference between the treatments in the topsoil or subsoil layers. Nevertheless, a trend of lower BDs was observed under the direct-drilling treatment in both the topsoil (1.46 g cm−<sup>3</sup> ± 0.073 vs. 1.52 g cm−<sup>3</sup> ± 0.086) and the subsoil (1.43 g cm−<sup>3</sup> ± 0.17 vs. 1.64 g cm−<sup>3</sup> ± 0.029). A significant (*<sup>p</sup>* = 0.007) sample depth × compaction alleviation treatment interaction was observed for the PR results, with the treatments ranked as follows: plough topsoil < direct-drilling topsoil < plough subsoil < direct-drilling subsoil (Figure 3). Overall, the PR was significantly lower in the plough plots (*p* < 0.001) and significantly higher in the subsoil compared to the topsoil (*p* < 0.001). In the subsoil, the PR exceeded the 2.5 MPa limit in about 30% of the measurements but did not exceed it in any in the topsoil.

**Figure 3.** Penetration resistance (PR) in topsoil (0–25 cm) and subsoil (25–45 cm) in ploughed and control direct-drilled (dd) plots at the UK study site.

The measurements of the SOCs in the topsoil showed no significant differences between treatments (plough: 2.85% ± 0.70; control dd: 2.85% ± 0.80). The measurements of Ks also showed no significant differences between treatments (plough: 1.28 × <sup>10</sup>−<sup>2</sup> m s−<sup>1</sup> ± 5.22 × <sup>10</sup><sup>−</sup>3, control dd: 1.53 × <sup>10</sup>−<sup>2</sup> m s−<sup>1</sup> ± 6.25 × <sup>10</sup><sup>−</sup>3).

#### *3.4. Italy*

The BD was not affected by agronomic management neither in the topsoil nor in the subsoil despite a tendency for denser topsoil being observed in the NT compared to the CT plot (1.43 vs. 1.35 g cm−3). The topsoil PR was affected by both tillage and soil covering, where the NT was 5% higher than the CT plot (1.05 vs. 0.82 MPa) (Figure 4). At the same depth, the PR had a lower value in the BS (0.85 MPa) compared to the WW (1.06 MPa). Contrarily, in the subsoil, the PR was not affected by the agronomic management. PR observations were always < 2.5 MPa for the topsoil, while 33% of the measurements exceeded that limit for the subsoil. No significant treatment effects were found.

**Figure 4.** Topsoil (0–20 cm) penetration resistance (PR) as affected by soil cover (**a**) and tillage (**b**) at the Italian study site. Different letters indicate a significant difference according to the Tukey test at *p* < 0.05. BS: bare soil; TR: tillage radish; WW: winter wheat; CT: conventional tillage; NT: no-tillage in VWC, SOC content, or stock.

Tillage affected the hydraulic parameters. Indeed, the sorptivity (*p* = 0.05) increased almost five-fold under NT management. The Ks, despite not significant with a *p* < 0.05, showed a three-fold value for the NT compared to the CT (Supplementary Materials, Figure S2). For further details, see Supplementary Materials, Table S5.

#### *3.5. Romania*

For the BD in the topsoil (10–20 cm), throughout all three years of the Romanian experiment, the mean value in the subsoiling treatment (TR2) ranged from 1.28 to 1.32 g cm−<sup>3</sup> and was always significantly different from the other treatments, which had higher values between 1.42 to 1.48 g cm−<sup>3</sup> (Figure 5a). With the exception of the control treatment (TR3), where no significant differences in BD occurred between the years, the BD was significantly lower in 2020 compared to 2018 for all treatments (Supplementary Materials, Figure S3a).

For the BD in the subsoil (40–50 cm), the results follow the same trend as for the topsoil regarding the subsoiling treatment and always had significantly lower values over all three years (Figure 5b). The control treatment also presented a significantly higher BD during each year of the study compared to the other three treatments. The bulk density in the subsoil was significantly lower in 2020 compared to 2018 for both the subsoiling and control treatments, but no significant differences were observed for the other two treatments (Supplementary Materials, Figure S3b).

**Figure 5.** Bulk density (BD) in topsoil (10–20 cm) (**a**) and subsoil (40–50 cm) (**b**) as affected by different tillage systems during the three years of the Romanian experiment. Different letters represent statistically significant differences according to the Tukey post-hoc test at *p* < 0.05. TR1: mouldboard ploughing with furrow inversion to 25 cm depth; TR2: subsoiling to 60 cm + disking to 12 cm depth; TR3: control treatment with 2-times disking; TR4: chiselling to 25 cm depth with furrow inversion.

The tillage treatments significantly affected the WSA in the topsoil throughout the 3 years of the experiment (Supplementary Materials, Figure S4a), with subsoiling (TR2) always exhibiting the highest WSA (23.9, 28.9, and 29.3% for 2018, 2019, and 2020, respectively) with respect to all other treatments (mean values across treatments of 17.0, 17.9, and 17.7% for 2018, 2019, and 2020, respectively). Except for 2018, the control (T3) always had a significantly lower percentage of WSA. The WSA remained the same during the study period for all treatments except for subsoiling, where the aggregation was significantly higher in 2019 and 2020 (Supplementary Materials, Figure S5a). The tillage treatments also significantly affected the Ks over all three studied years and were always 4 to 5 times higher for the subsoiling treatment (Supplementary Materials, Figure S4b). Differences in the Ks for other treatments only occurred in 2018, where TR2 (202 ×=10−<sup>8</sup> m s−1) >TR4 (74 × <sup>10</sup>−<sup>8</sup> m s−1) >TR3 and TR1 (60 × <sup>10</sup>−<sup>8</sup> m s−1). Only TR1 and TR2 differed between the years, with lower values of Ks in 2018 compared to the other years (Supplementary Materials, Figure S5b).

#### *3.6. Soil Compaction Indices and Crop Yield*

Generally, SC is considered to impair crop performance [70]. In the present study, the crop yield was only affected by the adopted SICS for the Romanian SS, a site predisposed to natural subsoil compaction as well as having a plough pan at 30 cm, which restricted the rooting depth. Compared to the other tillage types, the main effect of conducting subsoiling every year always gave the best crop performances after 3 years, with yields of 5.8, 1.6, and 3.4 Mg of dry matter ha−<sup>1</sup> for maize, soybean, and spring barley, respectively. There were no significant effects of the SICSs on crop yields at the other SSs (data not shown).

In contrast to the Romanian SS, the Swedish SS only applied the subsoiling operation once, and there were no significant differences in yields between the treatments throughout the experimental period. Subsoiling did not affect the whole area but only a portion of it (i.e., distinct subsoil rows), in which roots would theoretically be able to grow deeper and take up more water and nutrients by exploring a greater volume of soil. This trial differed in this respect from the other types of experimental treatments that affected the whole area. Thus, the measured yields of the whole field represent a weighted mean value of the treated and untreated subsoil volumes. Conceptually, calculating yields as the weighted mean of the affected and unaffected subsoil may be a more reasonable indicator of the effect of subsoil loosening. To illustrate this, we recalculated the measured relative winter wheat yield (2019) of the whole area compared to the control for the subsoiling treatments (Figure 6). This was done by scaling the two subsoiling treatments by factors of 100/38 and 100/45, considering they affected either 38% of the subsoil volume or 45% of the subsoil surface, respectively.

**Figure 6.** The measured relative winter wheat yield for the subsoiling (B) and subsoiling + straw (C) treatments compared to the control (A) in 2019 (left columns) at the Swedish site. Assuming that the whole (100%) subsoil was affected, and not only a portion of the subsoil surface (45%) or subsoil volume (38%), the potential yield increase is proportionally higher (middle and right columns).

To predict the possible effect of SC on crop performances, a few indices can be adopted [8,67,71,72]. A combined indicator of critical conditions in the soil (e.g., PR, porosity, and gas exchange) is the degree of compactness (DC), defined as the ratio of bulk density-to-reference density [8]. A threshold of 87% has been suggested as critical for root growth and crop development [8,67]. At the five SSs, the DC ranged from a minimum of 56% to a maximum of 124% (Figure 7a). The Norwegian SS exhibited the highest DC, which always exceeded the 87% threshold in both the topsoil and subsoil. On average, values in the subsoil were higher in the compacted compared to the reference plots (113 vs. 101%) (Figure 8b). At the Swedish SS, the DC was always <87%, averaging at 75 and 81% for the topsoil and subsoil, respectively (Figure 7a). For the UK SS, the DC averaged at 81%, with small variation between the topsoil and subsoil (Figure 7a). The DC limit exceeded the threshold of 87% for about one-third of the observations at the UK SS. At the Italian SS, the DC was higher for the no-tillage plots compared to the conventionally tilled plots (88 vs. 81% in the topsoil and 103 vs. 100% in the subsoil) (Figure 8a,b). At the Romanian SS, 47% (topsoil) and 25% (subsoil) of the measurements exceeded the DC limit but with lower magnitudes, where the maximum recorded DC was 90% (Figure 7a). At this site, a DC of >87% was frequently found in the topsoil under ploughing, chiselling, and disking, while this was only the case for the subsoil under disking (Figure 8a,b).

**Figure 7.** Box and whisker plots of degree of compactness (DC) (**a**), relative normalised density (RND) (**b**), and air-filled porosity (AFP) (**c**) in topsoil and subsoil at the five study sites. N: Norway; SE: Sweden; UK: United Kingdom; IT: Italy; RO: Romania. The box delimits values from low to high (from the 25th percentile to 75th percentile). Inside the box, the median and mean are indicated by a line and an X, respectively. The whiskers represent the minimum and maximum values in the range. \* data not shown for the UK topsoil due to high frequencies of zero values.

**Figure 8.** Degree of compaction (**a**,**b**), relative normalised density (**c**,**d**), and air-filled porosity (**e**,**f**) across the study sites (N: Norway; SE: Sweden; UK: United Kingdom; IT: Italy; RO: Romania) for topsoil (**a**,**c**,**e**) and subsoil (**b**,**d**,**f**). The dotted horizontal lines indicate an 87% degree of compactness, relative normalised density = 1, and air-filled porosity = 0.15, which represent suggested limits for good crop growth. The values shown are the mean and standard error. Rot 1, 2: oilseed rape–barley rotation; Rot 3: barley monoculture; Rot 4: alfalfa monoculture (N). CTRL: control treatment; Sub: subsoiling; Sub + straw: subsoiling + straw (SE). Control: direct-drilled treatment; Plough: ploughing system (UK). CT: conventional tillage; NT: no-tillage; BS: bare soil; TR: tillage radish; WW: winter wheat (IT). TR1: mouldboard ploughing with furrow inversion to 25 cm depth; TR2: subsoiling to 60 cm + disking to 12 cm depth; TR3: control treatment with 2-times disking; TR4: chiselling to 25 cm depth with furrow inversion (RO). For further details on adopted treatments, see Section 2. Materials and Methods.

The relative normalised density (RND), sometimes referred to as the degree of "overcompaction", is a texture-modified expression of density that might be useful to compare the state of compactness across differently textured soils [68]. Soil is defined as compacted when the RND > 1. In this study, the RND ranged from 0.65 to 1.09 across all the SSs (Figure 7b). The Norwegian SS always exhibited an RND of < 1.0 in the subsoil, with a higher value for the artificially compacted (0.90) plots compared to the reference plot (0.83) (Figure 8c,d). Similarly, the Swedish SS showed RND values in the subsoil below the suggested limit, being lower in the subsoiling + straw plot (0.78) compared to the subsoiling plot alone (0.84) (Figure 8c,d). At the UK SS, the RND ranged from a minimum of 0.65 to a maximum of 0.99 (Figure 7b). At the Italian SS, the RND was lower in the topsoil than in the subsoil (0.80 vs. 0.96), and the RND in the subsoil was only above 1 in the treatment with a winter wheat cover crop in the no-tillage system (Figure 8c,d). At the Romanian SS, the RND was always <1, with higher values in the topsoil associated with disking, chiselling, and ploughing (0.84, on average) compared to subsoiling (0.75), while the RND values in the subsoil of the different treatments ranked as follows: disking (0.87) < chiselling (0.81) < ploughing (0.79) < subsoiling (0.73) (Figure 8c,d).

In compacted soils, the soil–root contact may be very close, and reduced porosity could result in reduced soil aeration [73]. Therefore, the AFP may also be a useful index to estimate the compaction impact on crop growth. Except for tolerant crops, the ideal AFP is in the 10–15% range [74]. In the present study, the AFP ranged from values close to zero in the UK and some of the Norwegian treatments to a maximum of 0.32 found at both the Italian and Romanian SSs (Figure 7c). Higher topsoil values compared to the subsoil were found at both the Norwegian (0.06 vs. 0.01) and Italian SSs (0.25 vs. 0.09), while small differences between the soil layers were found at the Swedish (0.18 on average) and Romanian SSs (0.25 on average) (Figure 7c). At the latter SS, the AFP was affected by the tillage treatments in both the topsoil and subsoil, following the opposite trend observed for the RND (Figure 8e,f). Lower values in the topsoil were associated with disking, chiselling, and ploughing (0.25 on average) compared to subsoiling (0.30), while the AFP values for the subsoil of the different treatments ranked as follows: subsoiling (0.29) > ploughing (0.25) > chiselling (0.22) > disking (0.18) (Figure 8e,f).

Although roots may benefit from soil cracks and pre-existing bio-macro-pores [75], to fully exploit the soil, matrix roots must be able to explore the intra-aggregate space [76]. It is generally recognised that a root can either penetrate a soil aggregate or be deflected along its surface depending on the soil strength [72]. A total root growth decrease and impaired crop yield are observed when the PR exceeds a soil-specific limit, which typically ranges from 1 MPa [51] to 2 MPa or greater [47–53]. In this study, only the Swedish, UK, and Italian SSs directly measured the PR in the field. At the Swedish SS, the PR measurements in the subsoiling + straw incorporation treatment showed values below 2.5 MPa down to about 40 cm depth, while the control and subsoiling alone treatments showed values >2.5 MPa higher up in the soil profile (Figure 2b). Indeed, visual assessments for the presence of roots and maximum penetration (Supplementary Materials, Table S4) indicated that subsoiling had a positive impact on both the root growth and rooting depth at this site. At the UK and Italian SSs, soils under no-tillage presented higher PR values than the ploughed treatments, with at least 30% of the subsoils exceeding 2.5 MPa, which might impair root-growing conditions.

With the exception of the Romanian SS, there was no relationship (data not shown) between the crop yield and the SC indicators (i.e., DC, RND, AFP, and PR). At this SS, we found a 2% yield reduction for every percentage of DC increase or every unit of AFP decrease (Figure 9). This implies that in passing from a DC of 83% (average of Romanian soils) to 87% (DC limit for crop growth according to Håkansson [8]), a 7% reduction in the crop yield might be a possible scenario for this SS, irrespective of the crop type. The response of the crop yield to different levels of SC is usually considered parabolic, with low production in loosened soil, high yields at an optimal degree of soil compaction, and lower yields for compacted soils [77]. Only the descending part of this parabolic relationship may

have been observed at the Romanian SS, and it is possible that the optimal DC for crop production in its fine-textured soil might be located at a DC lower than 87% [6].

#### *3.7. Soil Compaction and SICS with Tillage*

The relationship between SC and tillage has been thoroughly investigated, especially in northern [16,78,79] compared to central [80] and southern Europe [81]. It is generally recognised that topsoil compaction might be mitigated using annual soil loosening with conventional tillage practices such as mouldboard ploughing, while subsoil compaction has proven to be persistent and difficult to recover and sometimes requires more specialised tillage operations [82].

In this study, all countries except Norway involved different intensities of tillage for mitigating topsoil and subsoil compaction. Inversion tillage through mouldboard ploughing was adopted in the UK (to 20 cm depth), Italy (to 30 cm depth), and Romania (to 25 cm depth). The UK and Italian SSs also included no-tillage (i.e., direct-drilling) treatments. Subsoiling treatments were used at the Romanian SS (to 60 cm depth) and in Sweden (to 35 cm depth), the latter with and without the injection of organic materials. Lower-intensity tillage (i.e., reduced or no-tillage) is considered to improve soil structural stability and, therefore, theoretically, tillage practices prevent SC [83]. On the contrary, highintensity tillage might produce an unstable soil structure more prone to soil compaction. Disking is one of the less conservative soil aggregate tillage practices, often resulting in a greater proportion of micro-aggregates (2–250 μm) but a lower proportion of macroaggregates (>250 μm) [84]. At the Romanian SS, the treatment with 2-times disking (TR3) decreased the topsoil BD but showed the lowest proportions of WSA and Ks together with the highest DC and RND and the lowest AFP. These findings suggest that despite providing suitable conditions for crop establishment, disking can make the soil more prone to SC due to greater soil structure instability. Mouldboard ploughing inversion tillage is considered responsible for soil aggregate fragmentation [85], although this negative effect on soil structure may be counteracted by increasing organic C inputs to the soil from crop residues or the incorporation of organic amendments [86]. At the Romanian SS, the results suggest that mouldboard ploughing inversion tillage had a less negative effect on the WSA compared to 2-times disking (Supplementary Materials, Figure S4a).

No-tillage or direct drilling is usually considered a more sustainable agronomic practice [87,88] because it is thought to be less harmful to soil biota, and by keeping the crop residues at the soil surface, it reduces the risk of soil erosion [89]. Verhulst et al. [90] found a greater proportion of large macro-aggregates (>2000 μm) and macro-aggregates (250–2000 μm) under no-tillage compared to conventionally-tilled soils, confirming both the positive effect of tillage absence and crop residue retention. The UK and Italian SSs show somewhat opposite results when comparing the effects of no-tillage and directdrilling against inversion tillage with mouldboard ploughing, with both the DC and RND being higher under no-tillage and direct drilling in Italy but not in the United Kingdom. Derpsch [91] identified four phases after the adoption of no-tillage—an "initial phase" (0–5 years), when crop residues are expected to be low due to lower yields and with no measurable changes in the SOC while the soil starts rebuilding aggregates; a "transition phase" after 5 to 10 years, when crop residues and SOCs are expected to increase, although these changes are accompanied by higher SC; improvements are expected only after 10–20 years during the "consolidation phase" followed by the "maintenance phase", characterised by stabilised agro-ecological conditions. Six et al. [92] found that for drier climates before a positive trend occurs, no-tillage could even have a negative effect on the SOC during the first 5–10 years. According to this classification, the Italian soil was in the initial phase and experienced SC, while the UK soil was reaching the end of the transition phase (around 10 years after the first adoption), showing improved soil conditions. Different soil types and their interactions with agronomic practices may explain the differences between these two SSs. The Italian SS is mainly formed from Calcisols and Cambisols (WRB, 2006), with low SOC content (<1.0%) and far from equilibrium, having a silty texture and poor aggregate stability [93]. Piccoli et al. [81,94] previously postulated that the limited amount of non-complexed SOC available for interaction with clay minerals and the low clay-to-silt ratio could prevent the formation of a resilient structure that goes beyond the adopted agronomic management. In contrast, the higher clay and SOC (2.88%) contents at the UK SS might have fostered an improved soil structure by ensuring high stability of the macro-pores [95], better exploiting the benefits related to no-tillage.

Subsoiling is primarily aimed at counteracting subsoil SC and does not disturb the soil surface unless it is associated with another tillage operation. At the Romanian SS, subsoiling was associated with shallow disking (to 12 cm depth), while the plots with the subsoiling and subsoiling + straw treatments at the Swedish SS were subjected to the same conventional tillage as for the control plots (i.e., mouldboard ploughing to 25 cm depth and normal seedbed preparations). The hypothesis for the Swedish SS, which is naturally compacted, was that the incorporation of organic material, in addition to subsoiling alone, would stimulate biological activity and lead to the stabilisation of the soil structure at a lower density, enabling roots to grow deeper. The results show a positive impact on root growth and rooting depth, particularly for the subsoiling + straw treatment, and partly confirm this hypothesis. However, when corrected for gravel and stones, there was no significant effect on the BD (Supplementary Materials, Table S4). The subsoiling treatment at the Romanian SS was aimed at counteracting the natural compaction and preventing the formation of a plough pan layer. At this site, both the topsoil and subsoil BD were significantly improved with subsoiling, as also reflected in the SC indices, and subsoiling had a positive effect on the topsoil WSA and Ks (Figure 5 and Supplementary Materials, Figure S4).

A stronger response to subsoiling at the Romanian SS compared to the Swedish SS was probably related to the frequency of subsoiling. It occurred only once in Sweden, whereas it was repeated every year during the Romanian experiment. However, it may also relate to the clay, silt, and SOC, as discussed in a meta-analysis by Schneider et al. [89]. They suggested that for many soils with a clay-to-silt ratio of <0.3, subsoiling might result in a complete collapse of the natural soil structure and SC instead of loosening, while for soils with a clay content of >20%, subsoiling may have a better possibility of lowering the BD and increasing the macro-porosity. The clay content was twice as high at the Romanian SS as at the Swedish site and the clay-to-silt ratio was 1.5, while this ratio was 0.34 in the Swedish subsoil. In fact, the response to the SICSs was faster at the UK and Romanian SSs compared to those in Italy and Sweden. This may be related to the high clay contents at the former sites (31 and 44%, respectively), which were much lower at the latter SSs (18 and 10%, respectively). Furthermore, high silt (58%) and sand (63%) contents characterise the Italian and Swedish SSs, and these inherent soil properties may partly explain the lower responsiveness to SICSs. The dynamics also differ between the topsoil and the subsoil; the former is more often subject to external factors (e.g., meteorological conditions), while the latter mostly follows natural dynamics (e.g., pedofauna activity). The results from the Italian SS agree with this reasoning, as there was a relationship between the PR (as an index of soil strength) and the fine silt + clay particles (0–20 μm), relating to the SOC protection against microbial degradation [96] only in the subsoil (*p* < 0.01 and 0.65 R2) and not in the topsoil (Figure 10). In this subsoil, a 10% increase in soil fines reflected a PR reduction of 0.6 MPa among the studied range.

**Figure 10.** Scatterplot of fines 20 (% of particles below 20 μm) against penetration resistance (PR) at the Italian study site. Linear interpolation equations and their coefficient of determination (R2) are also indicated for topsoil (0–20 cm) and subsoil (40–60 cm).

#### *3.8. Soil Compaction and SICS with Deep-Rooted Bio-Drilling Crops*

Another possibility for mitigating SC at deeper horizons or under no-tillage management is the adoption of deep-rooted crops, which may be either cash (e.g., alfalfa) or cover crops (e.g., tillage radish, mustard). Two SSs tested the effect of crops with a taproot apparatus—Norway (oilseed and alfalfa) and Italy (tillage radish).

Despite a relatively low machinery weight, multiple wheeling in 2015 led to subsoil compaction at the Norwegian SS [56], with increased BD and decreased TPV compared to the uncompacted reference plot (Table 2). In the topsoil, the presence of alfalfa (plots not ploughed) shows comparable results to Treatment 3 (barley monoculture), which was ploughed each year (Supplementary Materials, Table S3), suggesting that the alfalfa was equally effective at loosening the topsoil compared to ploughing. In 2015, the BD in the subsoil often exceeded 1.5–1.6 g cm−3, which represents a threshold for root growth [97]. All subsoil BD observations in 2015 were classified as "very compact" according to Pagliai et al. [98], while in 2020, all values could be classified as "compact", suggesting that the SC mitigation occurred during the five years after the compaction event. In particular, this was evident in Treatment 4 (alfalfa), where the Ks was improved compared to 2015. More specifically, the subsoil Ks under the alfalfa treatment was frequently higher than the proposed limit for good soil functioning (e.g., 0.10 m day−1) [22]. On the contrary, the same Ks threshold was undercut for all the other plots except for Treatment 2 in the compacted plots, confirming how subsoil compaction may be long-lasting [99]. Alfalfa was efficient at reducing the subsoil BD and restoring the TPV in the compacted plots, which were on the same level in 2020 as the uncompacted reference plot in 2015. These results are similar to other studies (e.g., [24,100]), showing that alfalfa, especially if grown over several years, is efficient for restoring soil structure. Effects on other parameters, such as air permeability and water infiltration, were less clear due to both higher data variability and a methodological issue. In fact, during the sampling operations, the

alfalfa was still growing, and the living roots blocked the bio-pores (data not shown). Consequently, positive effects on the soil structure (e.g., improvements in AC and water and airflow) might be more recognisable over time, after the roots have decomposed [101].

Furthermore, oilseed crops are known to have deep-growing taproots that efficiently loosen up the soil structure [29], but contrary to that observed for alfalfa, the oilseed established only poorly at the Norwegian SS mostly due to a short growing season at this high latitude. Therefore, the root system was not well established, and this crop was not effective at loosening the soil or mitigating SC.

At the Italian SS, using the deep-rooted tillage radish cover crop during the winter season was not reflected in soil improvements, either in terms of SC mitigation (e.g., BD, DC, and RND) or in terms of soil functioning (e.g., Ks). Only the PR test suggested a trend for higher soil strength in changing from bare to covered soil (Figure 4a). As mentioned for Norway, a methodological issue might have impaired the results for the tillage radish crop in Italy due to incomplete taproot decomposition during sampling since it was necessary to take measurements for the PR and Ks prior to the following field operations (i.e., in early spring before tillage and seedbed preparation for the main crop, maize). The higher temperature during the subsequent maize growing season may have promoted complete root degradation and, in turn, fostered improved soil functioning. Indeed, in higher density soils, the presence of a few vertical macro-pores may dominate structure dynamics and soil functions (e.g., water infiltration and gas exchange) [94,102] and possibly counteract the negative effects of increased BD and soil strength. Moreover, the bio-macro-pores left by tillage radish provide low resistance paths for the subsequent cash crop roots [103]. Bio-drilling with cover crops was previously demonstrated to be more effective for topsoil under no-tillage management than with conventional tillage because bio-pores can persist and function for a longer time without tillage [104]. Nevertheless, in subsoil below the tillage depth, root-derived bio-pores might also persist even if shallower tillage occurred.

Beyond its correlation with compaction, the overall impact of bio-drilling on cash crop yields varies with climate conditions [29], improving crop performances under highly rainy climates (e.g., tropical) [105] and reducing yields in semiarid environments [106]. The response of crop yields to bio-drilling might also be dependent on the number of years since its first adoption [29]. The first year of bio-drilling adoption may not result in a boosted crop yield, while after several years, a more positive effect can be expected [107,108]. Finally, bio-drilling crops may contribute to SOC formation by providing more above- and belowground C inputs to the soil, in addition to the crop residues from the main crop [109]. Particularly because they have an important and deep root system, and compared to aboveground biomass, root-derived C is about twice as efficient in the C input conversion into stable SOC. However, changes in the SOC occur slowly and become measurable only after longer periods (>5 to 10 years) [110,111].

#### **4. Future Prospects and Conclusions**

Strategies to avoid SC, stabilise soil structure, and loosen up compacted layers are clearly needed. Many conventional tillage practices are effective in loosening topsoil SC, but measures to counteract subsoil SC are scarce [22]. The use of deep-rooted bio-drilling cash and cover crops showed potential for being applicable in European countries. At the Norwegian SS, a cash crop such as alfalfa had good potential for mitigating both topsoil and subsoil SC. However, in using a cover crop such as tillage radish, the Italian SS did not obtain the expected positive outcome for SC. Both SSs experienced methodological difficulties in evaluating the effectiveness of these mitigating strategies. For practical reasons, some measurements could not be conducted at the optimum time and the presence of actively growing bio-drilling crop roots hampered the evaluation of water infiltration and hydraulic conductivity. Further studies are needed for investigating and identifying suitable crop varieties, as shown at the Norwegian SS, where it was difficult to establish the oilseed bio-drilling crop because of a short growing season at this high latitude. This highlights the need for optimising the management and crop growth of bio-drilling species

for different pedoclimatic conditions. There is also a need for policymakers to address the economical dimension, as farmers need financial support for adopting deep-rooted bio-drilling cash or cover crops. For example, alfalfa involves low production costs, but it is difficult to find a profitable market, and all types of relevant cover crops are not necessarily covered by current subsidies.

The two case studies with SICSs involving subsoiling were found to be efficient for improving both SC and crop yields only at the Romanian SS. However, it was also found that applying subsoiling every year was time and energy consuming, and the financial benefit for farmers is questionable. There is a need for further evaluating if the subsoiling at this SS could be done only periodically, such as every 3 or 4 years. At the Swedish SS, subsoiling was only done once and using pilot-scale equipment that is still under development. Although there was an improved rooting depth with subsoiling, and a lowering of the PR with the subsoiling treatment with the incorporation of organic material into the upper subsoil, there was no significant effect on crop production. There is a need for long-time studies with this equipment on other crop and soil types, to test other sources of organic materials and, in particular, to examine the effects of repeated subsoiling treatments over time.

The responsiveness of the SICSs investigated in these case studies appeared to be at least partially influenced by inherent soil properties, such as texture, as illustrated by the different responses to no-tillage observed at the UK and Italian SSs. The effect of climate was not evaluated directly since exactly the same SICS was not present at a sufficiently large number of SSs. However, the effect of climate is not negligible. Furthermore, the effect of future climate change might vary between European regions. In northern Europe, greater precipitation is expected during the growing season [12,82], which will lead to a reduction in workable days for field operations [20,112]. The use of heavy machinery under future sub-optimal conditions may further increase the risk for SC, especially in the subsoil [15]. Soil compaction may reduce water infiltration [6,113], which may shorten the growing season and thereby increase the risk of leaching and erosion [9–11]. It is also expected that wetter growing seasons might give greater yield reductions due to subsoil compaction than the drier seasons [114].

In southern Europe, a higher frequency of dry days during the growing season is predicted [82]. Since compacted soil may suffer from poor rooting conditions during drought [114], this could increase the demand for freshwater for irrigation [115]. Deep tillage can be an effective measure to mitigate drought stress and improve the resilience of crops under climate change scenarios in soils by creating a more stable soil structure and alleviating root-restricting layers [89].

The case studies on different SICSs for mitigating SC showed encouraging results as well as several difficulties relating to their implementation and evaluation. Some were pilot studies and need more technical development, and all were short-term studies. More research is needed to refine these SICSs and evaluate their long-term effects at more SSs covering a wider range of pedoclimatic conditions.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/land11020223/s1, Table S1: study sites description, Table S2: sampling depth, Table S3: topsoil results at the Norwegian study site, Table S4: results at the Swedish study site, Table S5: statistics at the Italian study site, Figure S1: relationship between soil organic carbon and dry soil bulk density at the Swedish study site Figure S2: saturated hydraulic conductivity at the Italian study site, Figure S3: bulk density at the Romanian study site, Figure S4: water-stable aggregates and saturated hydraulic conductivity by treatment at the Romanian study site, Figure S5: water-stable aggregates and saturated hydraulic conductivity by year at the Romanian study site.

**Author Contributions:** Conceptualization, I.P. and M.A.B.; formal analysis, I.P., T.S., J.B., O.V., I.C., A.B., G.B., H.K., T.K., F.S., C.S., F.C. and M.A.B.; investigation, I.P., T.S., J.B., O.V., I.C., A.B., G.B., H.K., T.K., F.S., C.S., F.C. and M.A.B.; data curation, I.P., T.S., J.B., O.V., I.C., A.B., G.B., H.K., T.K., F.S., C.S., F.C. and M.A.B.; writing—original draft preparation, I.P. and M.A.B.; writing—review and editing, I.P., T.S., J.B., O.V., I.C., A.B., G.B., H.K., T.K., F.S., C.S., F.C., I.S.P., A.A. and M.A.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research leading to these results received funding from the European Union HORI-ZON2020 Programme for Research, Technological Development, and Demonstration under grant agreement No. 677407 (SOILCARE Project). Martin A. Bolinder also received financial support provided by the Swedish Farmers' Foundation for Agricultural Research, grant number O-18-23-141.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author upon reasonable request.

**Acknowledgments:** Sølve Stiauren, Norwegian Agricultural Extension Service (NLR Øst), Kirkenær Norway took care of all the Norwegian practical fieldwork (except soil sampling) in the research field. The Italian study site wishes to acknowledge Riccardo Polese for his contribution to the field experiment. We wish to thank T. Ingelsson (farmer and consultant) for his contribution in adapting the subsoiling equipment used for the Swedish SS, and the Swedish Rural Economy and Agricultural Society for their help in managing the experiment. We acknowledge the technical and administrative personnel, and the many work package members and stakeholders for their efforts and insights. In particular, we thank Rudi Hessel (coordinator) for his leadership throughout the SoilCare project. More information on other SSs and SICSs, methodologies, and fact sheets as well as on other aspects, such as socio-cultural and economic dimensions, can be found at https://www.soilcare-project.eu, last accessed on 1 January 2022.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


#### *Article*
