Intensive Agriculture vs. Invertebrate Biodiversity: A Case Study of Woodland Islets in a Matrix of Arable Land

Abstract

Increasing areas of arable land, which is often heavily managed, negatively affect biological diversity in many ways, decreasing species richness and abundance. There is a substantial social demand for implementing agricultural management practices to preserve biological diversity locally. Here, we present the results of studies on the invertebrate diversity of woodland islets, which are small areas of forests surrounded by arable fields. Studies on invertebrate taxa show high values of diversity indices within such forest remnants, with a predominant presence of hymenopterans, collembolans, beetles, and spiders, which serve many ecosystem services, such as pollination or predation. A low abundance of herbivores and a high abundance of agile predators make such small woods a potential habitat for natural pest enemies. The results indicate a potential role for such woodland islets isolated from each other by a distance of ca. 1 km as an efficient substitute for extensive field management since they allow the maintenance of invertebrate diversity in the agricultural landscape.

1. Introduction

The increased intensity of agricultural management, which manifested several decades after World War II, has caused an enormous loss of biological diversity. Drivers linked to food production cause 70% of terrestrial biodiversity loss [1]. The vast number of studies carried out worldwide show a significant decrease in the number of species and their abundance in areas affected by intense farming [2,3,4,5,6]. The reasons for such a rapid decrease are the predominantly unsustainable agricultural practices, like the extensive use of pesticides, herbicides, and inorganic fertilizers around crop fields, as well as soil tilling, which collectively alter the environment to such a degree that it becomes inhabitable for many species. Nowadays, ca. 52% of agricultural land is degraded [1]. According to the Living Planet Report [7], between 1900 and 2016, the share of agricultural and build-up areas doubled. Meanwhile, it is predicted that in 2050, if our dietary habits and food production methods have not substantially changed, the estimated area of arable land will increase by 10–25% compared to 2005. However, it is expected that by 2050, the meat demand will grow by 200% compared to 2005. Given the negative implications of human activities for biodiversity, attempts to inhibit and reverse a global loss of species, especially in an intensely managed environment, have recently become the priorities of environmental protection goals [8,9].

A high proportion of the agricultural landscape leads to the fragmentation of natural habitats, especially forests, and consequently to significant habitat loss and increased spatial isolation of the remaining patches suitable for forest species [10]. According to the IPBES report [1], agriculture accounts for 80% of global deforestation. Large distances between the remaining woodland islands, which once lost connectivity, pose plant and animal dispersal difficulties. They may inhibit their migration and the successful establishment of their populations in woods, delivering habitats potentially suitable for colonization [11]. Thus, vast areas covered by crop monocultures are almost biological deserts and serious barriers for organisms living in adjacent habitats [12,13] once there are no linear structures that could act as effective migration corridors for forest species [14].

Nowadays, more than 70% of the world’s forests are located within a radius of 1 km of a forest edge, and nearly 20% are within 100 m of an edge, near agricultural, urban, or other environments heavily modified by humans. Thus, the fragmentation process of forest areas, with an altered microclimate and degraded structure, can significantly affect the species composition [15]. Once only small woodland islands remain in the agricultural landscape, the number of sites and the abundance of populations of forest species are distinctively reduced. Such a spatial structure of forests has tremendous implications for the long-term persistence of forest species, especially those with poor dispersal rates, in the era of global climate change since it may hamper their movement to find a new climatic optimum [14,16]. In this respect, special attention should be paid to the small patches of natural habitats remaining in intensively managed agricultural landscapes, like hedgerows and woodlands.

Temperate zones, especially forests, are of great interest due to their predominant association with species diversity across many taxa [17,18,19]. Two-thirds of all terrestrial species are associated with forests [20], with most vascular plants occurring in the forest herb layer [21]. Woodland islands are often a subject of research interest, especially when they are remnants of ancient forests, since forests of such an origin, if not heavily degraded, often maintain the well-preserved plant species composition typical of natural forests [22,23,24,25]. The long-term existence of such islets and the subsequent planting of new ones are often postulated [14] and are part of ongoing plans to restore biological diversity in agricultural landscapes [26].

Woodland islets and hedgerows in a landscape dominated by agriculture may serve as an archipelago of natural patches and as stepping stones in ecological corridors leading from one natural habitat to another across an agricultural matrix hostile to most species associated with woodlands [14]. The awareness of the importance of such habitats in the European landscapes heavily transformed by humans is progressively increasing among EU countries. Both the EU Biodiversity Strategy for 2030 and the Nature Restoration Law documents emphasize [27,28], among other things, the need for either their maintenance or restoration to combat the dramatic global biodiversity loss and to “put Europe’s biodiversity on the path to recovery by 2030 for the benefit of people, climate and the planet”.

Numerous studies concentrate on the role of habitat islands in the diversity of forest herb layer species [23,29,30,31], whereas investigations focusing on invertebrates are less frequent [11,32,33,34]. For these reasons, our study aimed to assess the extent to which woodland islets may serve as suitable habitats for invertebrates in the agricultural landscape despite the heavy use of fertilizers, pesticides, and herbicides in their surrounding and determine the diversity of invertebrates (diversity indices) in these woods. Additionally, due to the high potential in terms of multiple ecosystem services shown by invertebrates in an agricultural landscape, we investigated the spectrum of ecological groups in reference to their dispersal abilities (one may expect that some groups face difficulties in migration due to their low dispersal capacity) and feeding adaptations.

2. Material and Methods

Studies were conducted in seven broadleaved woodlands isolated from other forested areas and surrounded by the matrix of the intensively managed agricultural landscape of the Opolska Plain, southern Poland (Figure 1 and Figure 2).

Due to the fertile soils, the effects of the deforestation of this region, which has mostly been converted into agriculture and human settlements, were already seen on the first cartographic source, the mid-18th century (years 1767–1787) map of Schmettau and von Schulenbug-Kehnert, where six of the studied woodland islands were already present. In contrast, one of them (site V) appeared in the 20th century. Thus, six of the woods are remnants of ancient forests and nowadays represent a well-preserved oak–hornbeam Tilio cordatae–Carpinetum betuli Trach. 1962 community, whereas the remaining one can be treated as a recent, post-agricultural sensu Peterken [22], and its herb layer composition is dominated by meadow grasses, with a small admixture of forest species. The area of the woods ranged between 0.42 and 4.06 ha, and the mean distance to the nearest woodland area was 1.024 km (SD ± 0.756). Detailed information about the location of the studied woodland islets (sites), their area, and the distance to the nearest wood of a similar size is presented in

Table 1. The location and the area of the studied woodland islets and their distance (from the edge) to the nearest woodland area of at least the same size (km).

			Distance
Site No.	Location	Area (ha)	N	S	W	E
I	N: 50.4707 E: 18.4731	2.638	1.610	2.450	0.697	2.181
II	N: 50.4857 E: 18.4255	4.057	0.846	0.367	0.970	0.571
III	N: 50.4777 E: 18.4718	1.639	0.928	1.100	1.006	2.630
IV	N: 50.4773 E: 18.3081	0.956	0.515	1.380	0.753	2.920
V	N: 50.4783 E: 18.3022	0.421	0.324	1.540	0.276	0.229
VI	N: 50.4783 E: 18.3061	0.713	0.395	1.470	0.571	0.128
VII	N: 50.4751 E: 18.2912	2.870	0.268	0.871	0.611	1.059

.

Pitfall traps of 6 cm in diameter were used with a mixture of a 30% water solution of ethylene glycol, 10% glycerol, and a drop of scentless detergent to collect invertebrates. They were set three times for one week (168 h of constant trapping) on the following dates: 6–13 April 2023, 11–17 May 2023, and 11–17 June 2023. The pitfall traps were placed in the ground, evenly with the soil level, at least 1 m from each other. In each woodland islet (site), traps were placed in the following pattern: ten near the northern edge, ten near the southern edge (at a distance ranging from 0 to 15 m), and four in the woodland interior (location the farthest from the edge, representing the best, shady, forest interior conditions available within a wood) except for site no. V, where the setup was 8 + 8 + 4 due to the very small size of the wood, not allowing for the same sampling scheme. A total of 164 (24 × 6 sites + 20) traps were placed per sampling date, which gives 492 traps (164 × 3 sampling weeks, a total of 504 h of collection). Then, for most analyses, we compiled the data on the identified invertebrates from the S and N edges and the interior into one group per site (wood).

The collected specimens of animals were counted and identified to the higher taxonomical levels of type, class (Arachnida, Myriapoda), or order (Insecta). Due to the applied collection method, some airborne groups may be underestimated. The taxa were then assigned to functional groups based on the literature data (e.g., [35,36,37]) and general characteristics of groups existing in taxonomic and faunistic resources. Thus, in order to reflect the animals’ diverse dispersal capacities, playing a pivotal role in their chances of effective movement in the matrix of arable land with fragmented, isolated woods, species were classified according to their movement abilities and divided into the following functional groups: crawlers, climbers, walkers, runners, and fliers. Some taxa were assigned to a single group, e.g., snails and slugs to crawlers, dipterans to fliers. Others, for example, ants, were assigned to walkers and runners, as various species in this family represent both modes. Additionally, to emphasize the fact that diverse feeding strategies reflect the wide range of sources of food available in these woods and, on the other hand, play multiple environmental roles, the invertebrates were grouped according to their feeding modes and requirements into omnivores, detritivores, carnivores, phloem feeders, foliage feeders, and pollinators. Some of the taxa were simultaneously assigned to a few groups, e.g., Hymenoptera were assigned to omnivores, carnivores, and pollinators since they play multiple roles in their environments.

The abundance of particular taxa was used to calculate the diversity indices: Shannon’s, Simpson’s, and Brillouin’s, as well as the evenness indices for these. These indices were calculated in two ways: separately for each woodland (site), after compiling the N and S edges and interior samples into a single one, and separately for the northern and southern edges and woods’ interiors, regardless of the site (grouping all the N locations into one group, all the S locations into a second group and all the wood interiors into a third group). Then, the diversity indices values were checked for statistical differences using Kruskal–Wallis tests, performed in the STATISTICA ver.13.3 package (StatSoft). All the collected material was preserved in 70% ethanol and deposited in the collection of the Institute of Biology, Biotechnology and Environmental Protection, Faculty of Natural Sciences, University of Silesia in Katowice, for further identification and for verification purposes, if required.

3. Results

In the three sampling periods, 13,164 specimens of invertebrates belonging to 24 distinguished taxa were caught. The general results summarized for all the studied woods and given for each woodland separately are presented in

Table 2. Detailed results of the abundance of particular taxa in the studied sites.

	Gastropoda	Nematoda	Oligochaeta	Isopoda	Acari	Opilliones	Araneae	Pseudoscorpiones	Diplopoda	Chilopoda	Collembola	Coleoptera	Dermaptera	Diptera	Hymenoptera without Formicidae	Formicidae	Hemiptera	Lepidoptera	Psocoptera	Blattodea	Orthoptera	Thysanoptera	Mecoptera	Siphonaptera	Total
Site I	226	0	8	56	150	30	85	1	21	6	240	434	43	141	139	179	26	42	3	2	0	19	0	0	1851
Site II	66	0	10	8	268	44	110	1	9	1	213	161	19	80	177	127	149	14	4	0	0	122	0	0	1583
Site III	93	0	8	79	107	122	125	0	102	1	318	328	30	91	164	166	32	31	0	0	3	93	1	3	1897
Site IV	35	1	8	17	111	26	188	0	47	2	363	309	31	93	154	169	47	29	2	1	0	12	0	4	1649
Site V	35	1	36	12	42	15	283	0	22	0	256	233	167	62	505	563	39	76	2	2	1	18	0	2	2372
Site VI	2	12	11	20	68	210	339	0	68	0	356	203	5	132	173	195	68	36	4	2	0	11	0	1	1916
Site VII	39	1	13	23	99	117	121	0	12	0	707	309	26	107	81	136	67	8	1	0	0	29	0	0	1896
total	496	15	94	215	845	564	1251	2	281	10	2453	1977	321	706	1393	1535	428	236	16	7	4	304	1	10	13,164

. The most abundant were representatives of Hymenoptera (2928 specimens), followed by Collembola (2453) and Coleoptera (1977). The least abundant were Mecoptera (1 specimen), Pseudoscorpiones (2), and Orthoptera (4). Some of the taxa had very dissimilar abundance between the woods, e.g., there were 226 specimens of Gastropoda in wood I but only two in wood VI; similarly, Dermaptera—167 in wood V and only five in wood VI. On the other hand, some other taxa were relatively evenly distributed among the woodlands, e.g., Araneae, Collembola, Coleoptera, Diptera, and Hymenoptera (Table 2).

The diversity indicators reached high values in each studied site (

Table 3. Diversity indices of each of the studied woodlands.

Site No.	No. of Taxa	Simpson’s Indicator	Evenness	Shannon’s Indicator	Evenness	Brillouin’s Indicator	Evenness
I	19	0.81	0.90	1.96	0.84	1.65	0.84
II	18	0.81	0.90	1.93	0.85	1.64	0.85
III	19	0.81	0.90	2.01	0.84	1.74	0.84
IV	20	0.81	0.91	1.96	0.85	1.67	0.84
V	20	0.78	0.87	1.90	0.80	1.63	0.80
VI	19	0.82	0.91	1.95	0.86	1.68	0.86
VII	17	0.79	0.89	1.90	0.84	1.59	0.83

), and the number of recorded taxa ranged between 17 and 20 in each woodland site. When analyzed separately for the northern and southern edges and woodland interiors, the diversity indicator values were always the highest within the woodland interior (Figure 3). These differences were statistically significant for Simpson’s and Shannon’s coefficients (p < 0.05), but not for Brillouin’s coefficient (with p > 0.1 in Kruskal–Wallis test). However, Brillouin’s evenness was also statistically significant (p < 0.005).

Among the dispersal modes, the climbers and walkers dominated the studied woods, while the crawlers were the least abundant functional group of invertebrates (

Table 4. Shares [%] of functional groups in the studied woodland islets.

Functional Groups with Reference to Feeding Modes
Detritivores	Carnivores	Omnivores	Foliage Feeders	Phloem Feeders	Pollinators
25.22%	20.48%	29.72%	4.76%	4.73%	15.09%
Functional groups with reference to movement abilities
Climbers	Crawlers	Walkers	Runners	Fliers
36.54%	2.15%	26.76%	11.18%	23.37%

).

In terms of the feeding modes, the shares of particular groups were very similar, with omnivores having 30% of the abundance, detritivores and herbivores (phloem feeders, foliage feeders, and pollinators) having 25% each, and carnivores reaching 20% (Table 4).

4. Discussion

The results indicate a relatively high diversity of invertebrate taxa in woodland islets despite being surrounded by intensively managed agricultural landscapes. Although some of the taxa are poorly represented (usually by single specimens), possibly due to the applied collection method, there are no taxa with dominant abundance, and the evenness of the taxa abundance is very high, which is typical of natural and seminatural habitats. The much higher values of the diversity indices in the woodland interiors compared to their edges signify the similarity of the habitat condition of the inner parts of these patches to the forest interiors, despite agricultural pressure and the small areas of these woods. The influence of crop field management, including unsustainable practices (high loads of fertilizers and pesticides), might explain the lower values of these indices on the edges compared to the islets’ interiors. Significant for the study sites is the meager share of crawlers, mainly gastropods, which are undoubtedly connected to significant dispersal capabilities through the agricultural landscape. High sun exposure, pesticides, and tillage may be the main barriers for snails attempting to penetrate new areas, like the woods, through crop fields. The vast majority of recorded taxa are actively moving animals, capable of effective, quick migration over huge distances, e.g., spiders (via silk threads) or flying insects. For these groups, the colonization and settlement in woodland islets surrounded by heavily transformed agricultural areas seems not to be problematic, regardless of the distance from the nearest woodland (up to 2.92 km, usually ca. 1 km) (Table 1). Similar research on bats also proved that the presence of woodland islands is more beneficial for species with high mobility [38]. Also, flightless species of beetles or species living underground seem to be most vulnerable to a decline in agriculture [39]. Considering the mean surface of the studied woodland islets and the surrounding landscape within 1 km of the nearest woodland patch, islets constitute, on average, 0.62% of the anthropogenic crop field surface. This result accords with the suggested <1% of the field surface [26] needed to retain biological diversity. The observed number and abundance of high-level taxa and the high diversity indices values align with the observed multiservice potential of woodland islands [40]. Also, the observed distance lies between the distance values recognized as beneficial for moths migrating between fragmented woodland patches [41], so it may be regarded as the compromise distance between forest patches in the agricultural landscape, although not for species with low dispersal potential.

Regarding a woodland islet’s area, although not supported by statistical tests, due to the sample size being too small, there was no distinctive correlation between its area, the number of taxa, and the abundance of collected specimens. The lowest abundance of invertebrates was recorded in the most prominent woodlands, i.e., in the biggest one (site II), while the highest was in the smallest one (site V) (Table 1 vs. Table 2). Other studies provide contradictory results on species richness, e.g., Carabidae and Araneae were significantly correlated with the woodland area [42]. However, such trends were not observed in the case of Myriapoda, Opiliones, and Isopoda [42,43]. Such divergent results indicate that not only the area of the woodland islets and distance from other habitat patches are important but also particular species’ requirements as well as the vegetation type, age, and plant species composition [44,45].

The role of woodland islets, constituting less than 1% of the agricultural landscape, may not only be substantial due to the maintenance of biodiversity in the agricultural landscape but also for crop productivity. The islets may serve as a source of potentially beneficial insects, especially pollinators and carnivorous but also flying predators, which, by penetration of the surrounding fields, may also diminish the population of crop pests [46,47,48]. We observed a 15% share of pollinators, predominantly Hymenoptera but also Diptera and Coleoptera, and 20% of carnivores (mainly spiders, but also Hymenoptera and Coleoptera) and a relatively low number of potential pests—phloem and foliage feeders. Among the phloem feeders, aphids predominate and tend to be monophagous species in natural habitats. In contrast, most aphid pests are oligophagous to polyphagous species—a similar case concerned foliage feeders—lepidopterans. Therefore, woodland islets cannot be regarded as an essential reservoir of invertebrate species potentially destructive to cultivated plants but as the refugia for species with diverse feeding modes, thus playing multiple environmental roles in impoverished, intensively managed agricultural landscapes. Even tiny woods may serve as a habitat or nesting site for pest-feeding birds [49].

Consequently, woodland islets should be rather regarded as an essential part of natural modes of pest control, as the increasing heterogeneity of agricultural landscape positively affects trade-offs in biological control of crop pests [50] and hosts more diversity than simple, large-scale crop monocultures [51,52]. Based on the obtained results, one may presume that in the homogenized landscapes dominated by intensively managed arable land, even tiny woods, but with relatively well-preserved vegetation (mature, broadleaved stands, species-rich, well-developed woodland vertical structure) and diverse habitat conditions may effectively serve as refugia for multiple invertebrate taxa in such a simplified landscape. However, when considering the risk that in the case of invertebrates, the extinction depth has yet to be fully reached in response to former disturbances, one cannot predict how long these small woods may play the role of species refugia in this region. There is a risk that their invertebrate diversity will gradually decrease with time before it reaches a new equilibrium. Furthermore, we do not know how the diversity of the studied woods relates to the diversity of a comparable forest patch area inside a more extensive forest since we did not find such undisturbed forests in this region for comparison. Despite these uncertainties, the long-term maintenance of such woods in arable fields should be a priority for local managers. These attempts should be coupled with efforts to improve the connectivity of wooded areas in the agricultural landscape of this region. Undoubtedly, the presence of such woods, well connected via landscape corridors, would help achieve the goals of the EU Biodiversity Strategy 2030 and the Nature Restoration Law to benefit biodiversity, climate, and people.