**Effects of Vermicompost Leachate versus Inorganic Fertilizer on Morphology and Microbial Traits in the Early Development Growth Stage in Mint (***Mentha spicata* **L.) And Rosemary (***Rosmarinus officinalis* **L.) Plants under Closed Hydroponic System**

**Abraham Loera-Muro 1, Enrique Troyo-Diéguez 2, Bernardo Murillo-Amador 2, Aarón Barraza 1, Goretty Caamal-Chan 1, Gregorio Lucero-Vega <sup>3</sup> and Alejandra Nieto-Garibay 2,\***


**Abstract:** The objective of this study was to compare the morphology of *M. spicata* and *R. officinalis* plants, and the relative abundance quantification, colony-forming units, ribotypes, and biofilm former bacteria under an inorganic fertilizer and the use of vermicompost leachate in the rhizosphere under a closed hydroponic system. In mint (*Mentha spicata*) plants treated with the vermicompost leachate, growth increase was determined mainly in root length from an average of 38 cm in plants under inorganic fertilizer to 74 cm under vermicompost leachate. In rosemary (*Rosmarinus officinalis*), no changes were determined between the two treatments. There were differences in the compositions of microbial communities: For *R. officinalis*, eight ribotypes were identified, seven for inorganic fertilizer and four for vermicompost leachate. For *M. spicata*, eight ribotypes were identified, three of them exclusive to vermicompost leachate. However, no changes were observed in microbial communities between the two treatments. Otherwise, some changes were observed in the compositions of these communities over time. In both cases, the main found phylum was Firmicutes, with 60% for *R. officinalis* and 80% for *M. spicata* represented by the *Bacillus* genus. In conclusion, the use of vermicompost leachate under the hydroponic system is a viable alternative to achieve an increase in the production of *M. spicata*, and for both plants (mint and rosemary), the quality of the product and the microbial communities that inhabited them remained unaltered.

**Keywords:** organic fertilizer; hydroponic; ribotypes; vermicompost leachate

#### **1. Introduction**

At present, the growing global population has put pressure on agriculture in different ways: the increase in demand for food and the need to meet this demand in an environmentally friendly manner. Although the use of chemical fertilizers has led to an enhancement in crop production, several major health- and environment-related concerns are associated with their use [1,2]. Pollution and the increase in global temperature are predicted to have negative consequences for agriculture in the coming decades [3]. Likewise, future climatechange scenarios predict a more frequent occurrence of extreme conditions [4]. In this sense, hydroponic systems have emerged as an alternative to improve yield, product quality, water management, land saving, nutrient recycling, and environmental and pathogen control. Hydroponic systems are cultivation technologies that use nutrient solutions rather

Troyo-Diéguez, E.; Murillo-Amador, B.; Barraza, A.; Caamal-Chan, G.; Lucero-Vega, G.; Nieto-Garibay, A. Effects of Vermicompost Leachate versus Inorganic Fertilizer on Morphology and Microbial Traits in the Early Development Growth Stage in Mint (*Mentha spicata* L.) And Rosemary (*Rosmarinus officinalis* L.) Plants under Closed Hydroponic System. *Horticulturae* **2021**, *7*, 100. https://doi.org/10.3390/ horticulturae7050100

**Citation:** Loera-Muro, A.;

Academic Editors: Nazim Gruda and Juan A. Fernández

Received: 16 April 2021 Accepted: 28 April 2021 Published: 6 May 2021

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

**Copyright:** © 2021 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/).

than soil substrates, and can use natural or artificial media to provide physical support to plants [5–7]. However, there is an intense debate about which hydroponic practices align or do not align with the Organic Foods Production Act (OFPA) and USDA organic regulations [8]. Furthermore, hydroponic systems are a form of soilless food production, and one of the points of conflictive in points in organic agriculture is the use of inorganic nutrition in water solutions, which many people strongly believe should not be allowed [8]. Hydroponic production has increased in recent years due to its multiple benefits. Thus, it is convenient to understand the role of microorganisms and natural sources of nutrients to improve hydroponic systems for the production of healthy food beyond reaching certification in organic agriculture. At present, the use of vermicompost leachate coupled with hydroponic systems seems to be a viable alternative. Vermicompost is the resulting product from the processing of organic waste in the digestive tract of earthworms [9,10]. This process involves the bio-oxidation and stabilization of organic compounds by the joint action of earthworms and microorganisms [11]. Consequently, the obtained vermicompost is a fertilizer with available nutrients for plants and a strong charge of beneficial bacteria [12,13]. Likewise, vermicompost is an effective technique to reduce the toxicity of waste material [14]. Vermicompost leachate is a subproduct of the vermicompost process with nutrients, microorganisms, and biologically active substances, such as fulvic acids and humic acids, and the released water during the decomposition of the organic material [15,16]. One of the positive effects of the use of vermicompost leachate is an increase in the population of plant-growth-promoting bacteria (PGPB) [17]. PGPB can promote plant growth by both direct and indirect mechanisms. Direct mechanisms include the production of auxin, ACC deaminase activity, cytokinin, gibberellin, the nitrogen fixation process, phosphorus solubilization, and the sequestration of iron by bacterial siderophores. Indirect mechanisms refer to the bacterial capability to inhibit the proliferation of plant pathogenic organisms, such as fungi and bacteria [2,18]. Most studies on hydroponic systems reported the role of indigenous bacteria and the effects of bacterial addition, and indirect bacterial mechanisms for biological pathogen control, but scarce data are available about the existence of differences between the bacteria content and plant growth when applying vermicompost leachate to a hydroponic system [13,19]. The influence of agricultural management practices on plant microbial communities is not completely clear [20]. Opportune microorganism identification in hydroponic systems which uses vermicompost leachate as a low-cost organic fertilizer is essential to select the most adequate microorganisms for an efficient pathogen biocontrol program, also to define a fertilization protocol for this system environmentally friendly and accessible to any producer [21]. Mint (*Mentha spicata* L.) and rosemary (*Rosmarinus officinalis* L.) are two plants of agronomic importance belonging to the *Lamiaceae* family [19,22], a family with many wild and cultivated officinal species, rich in essential oils and antioxidant compounds that are useful to humans [23,24]. The leaves of *M. spicata* are dried and used for tea infusions, and cultivated for the production of essential oils that are widely used in the pharmaceutical and cosmetic industries [19]. *R. officinalis*, besides its culinary uses due to its characteristic aroma, is also widely employed by indigenous populations in areas where it spontaneously grows. Rosemary extracts are used as a natural antioxidant, improving the shelf life of perishable foods [22,25]. This study assessed the effect of two types of fertilizer (inorganic versus organic fertilizer) on the growth of mint (*M. spicata*) and rosemary (*R. officinalis*) plants under a hydroponic production system, as an alternative agronomic method contributing to a reduction in pollution, water use, and fertilizer consumption, and low-cost production.

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

#### *2.1. Study Area*

The experiment was conducted in a shade-enclosure environment that served as a greenhouse facility in La Paz, located in a Bw (h') hw (e) climate, which is considered to be semiarid and sustains the xerophytic vegetation of Baja California Sur, northwest Mexico, at 7 m above sea level. Mean, maximal, and minimal temperature in the shade-enclosure

facility were 21.4, 31.8, and 8.9 ◦C, respectively, with a mean of 70% relative humidity. Meteorological records were obtained during the study from an automated weather station located inside the shade-enclosure facility.

#### *2.2. Plant Cultivation Conditions and Hydroponic System*

The experiment was carried out from September to November. *M. spicata* and *R. officinalis* cuttings were obtained from mother plants within their regional cultivars and were placed in pots with vermiculite until they developed enough roots to be able to absorb nutrients from fertilizers after applying the treatments. The pots were placed in 30 propylene containers of 20 L (24.5 × 16 × 10 cm (length × width × height)) filled with water. Oxygen supplementation in containers was provided with a Blogger Sweetwater pump (model SST20, 50 Hz). The water volume was maintained constant to build a closed hydroponic system; there was no recirculating water because the study was on the early vegetative stage (September to November).

#### *2.3. Treatments and Experimental Design*

The experimental design consisted of two treatments: one applying vermicompost leachate (L) and the other applying inorganic fertilizer (SS; control group) [26]. Vermicompost leachate (L) was produced at the CIBNOR experimental field according to recommendations by Gunadi et al. [27]. The vermicomposting process was carried out in 200 L containers cut in half, to which 5 holes were made in its base. Subsequently, a 5 cm thick layer of gravel and an antiaphid mesh were placed to separate the gravel from the bed where the earthworms developed. Kitchen waste and manure were used as food for the earthworms in a ratio of 1:1 volume:volume. Both kitchen waste and manure were precomposted for 21 days before being used as food for the earthworms. The feeding process was carried out using 5 cm thick layers of precomposted food every week for 12 weeks. The vermicomposting process was considered to have ended when a homogeneous material was observed without the presence of remnants of the original material. The vermicompost was separated to be laid and sheltered in a dry place and away from light for 90 days for its mineralization. Vermicompost leachate was obtained according to the methodology described by García-Galindo et al. [28], where 5 kg of vermicompost was placed in a container. Three liters of distilled water was poured into the container, and the leachate was collected. Information of the nutrient content of both inorganic fertilizer and vermicompost leachate is shown in Table 1. The experiment was established under a completely randomized design with 15 replicates for each treatment (vermicompost leachate and inorganic fertilizer). Each replicate consisted in a container before descripted with 12 pots, each pot with one plan. Treatments were applied once at five days after sowing (DAS), for inorganic fertilizer a commercial fertilizer of 17% NPK was used to prepared 10 mL that contained 0.0079, 0.000087, 0.070 (parts per million of N, P K, respectively) diluted in 40 L of top water (the capacity of pot container). For the vermicompost-leachate treatment, 140 mL that contained 0.00709, 0.000259, and 0.074 (parts per million of N, P K, respectively) was diluted in 40 L of tap water. The nutrient doses of N–P–K corresponded to the minimum established for these crops in the region to examine if any differences could be detected in microbial and morphological traits in the use of an organic versus inorganic fertilizer. Plants were analyzed in early-stage growth at 35 days after fertilizer application.

#### *2.4. Morphological Traits and Relative-Growth Analysis*

Stem length (SL, cm), fresh stem weight (FSW), dry stem weight (DSW), foliar area (FA), fresh foliar weight (FFW), dry foliar weight (DFW), root length (RL), fresh root weight (FRW), and dry root weight (DRW) were evaluated in five *M. spicata* plants and five *R. officinalis* rosemary plants before treatment application and at the end of the experiment (35 DAS). Stem and root weights (g) were obtained using an analytical scale (Mettler Toledo, AG204); for dry weights, an oven was used with forced air circulation at 70 ◦C (Shel-Lab®, FX-5, series 1000203) until constant weight. Data of initial and final dry weights were used

to calculate total relative growth rate (TGR), foliar growth rate (FGR), root growth rate (RGR), and stem growth rate (SGR) in grams per day, according to Hunt [29], following Formula (1):

$$\text{TGR} = \text{((lnDW2)} \times \text{(lnDW1)} / \text{(t2 - t1)},\tag{1}$$

where DW2 and DW1 are the total plant (TGR), foliar (FGR), root (RGR) and stem (SGR) dry weight (g), recorded at times t2 (time of sampling) and t1 (beginning of the experiment), respectively. The difference (t2 − t1) is expressed in days. TGR, FGR, RGR, and SGR are expressed in g−<sup>1</sup> day<sup>−</sup>1.

**Table 1.** Solution-component analyses of nutritional source for *M. spicata* and *R. officinalis* in hydroponic system.


SS: inorganic fertilizer, L: vermicompost leachate.

#### *2.5. Photosynthetic Pigments*

For *M. spicata* and *R. officinalis* plants under organic and inorganic treatments, we determined chlorophyll with seven plants (one leaf per plant) per treatment. *M. spicata* SPAD values [30,31] were recorded for 20 consecutive days after the beginning of both organic and inorganic treatments application. In *R. officinalis* plants, chlorophyll was evaluated two times: before any treatment application, and 20 days after both treatment applications. For *R. officinalis*, the chlorophyll was extracted following the acetone extraction methodology from leaf tissue, and the absorbance measure was carried out with a UV/visible spectrophotometer (model HELIOS OMEGA, Thermo Scientific, Vantaa, Finland). Chlorophyll a and b concentrations were estimated by applying the following functions [32]:

$$\text{Chlorophyll a (mg mL}^{-1}) = 11.64 \text{ (A663) } -2.16 \text{ (A645)}\tag{2}$$

$$\text{Chlorophyll b (mg mL}^{-1}) = 20.97 \text{ (A645)} - 3.94 \text{ (A663)}, \tag{3}$$

where A663 and A645 correspond for the absorbance values at wavelengths (λ) of 663 and 645 nm, respectively.

#### *2.6. Sampling for Bacterial-Community Characterization*

To determine the influence of organic and inorganic fertilizers on rhizobial microbial communities from the plant rhizosphere, samples of the root rhizosphere were taken in the hydroponic system as follows: a water sample of 50 mL with the roots (0–0.5 cm) from three different reservoirs at three times (1, 7, and 35 DAS). The collected samples were processed immediately for: (i) total DNA isolation from water (rhizosphere) samples, and (ii) bacterial isolation from *R. officinalis* and *M. spicata* root samples with the methodology that follows below. Vermicompost was free of pathogens.

#### *2.7. Colony-Forming Units (CFU) Quantification and Isolation of Bacteria from M. spicata and R. officinalis Cultivated by Hydroponic System*

The water and root samples were vorticed for 30 s. Then, 25 mL of the sample was transferred to a new tube for DNA isolation. The remaining 25 mL was used to determine the colony-forming units (CFU). One milliliter of the remaining sample was used to perform serial dilutions in saline solution 0.85% (*w*/*v*) (from 10-2 to 10-7). Lastly, 100 μL for each dilution (from 10-2 to 10-7) was plated on nutrient agar (NA) and incubated for 24 h at 30 ◦C. After 24 h, the CFU count was performed.

After the CFU count, bacterial colonies were isolated on the basis of their morphology. A representative colony of the five most abundant colonial morphologies was reseeded by streak dilution in a new plate of NA and incubated at 30 ◦C overnight. This step was repeated until a pure isolate in each case (a single bacterial morphology per isolate) was obtained. The obtained pure isolates were stored in glycerol 30% (*v*/*v*) at −80 ◦C until their use.

#### *2.8. DNA Isolation*

The total DNA isolation of the water samples and bacterial isolates was carried out according to the protocol with slight modifications [33]. For water samples, 25 mL was centrifuged at 5000× *g* for 10 min, and the supernatant was discarded. For bacterial isolates, 3 mL of liquid culture was placed in nutrient broth (NB) at 30 ◦C overnight and centrifuged at 5000× *g* for 5 min, and the supernatant was discarded. Both the pellet from water samples and the bacterial isolate pellets were processed in the same way. The resulting pellet was resuspended in 1 mL of a lysis buffer (15% sucrose, 0.3 mg/mL lysozyme, 0.05 M EDTA and 1 M Tris, pH 8) and incubated for 30 min at 37 ◦C. Then, 100 μL of 10% SDS (*w*/*v*), 100 μL of 5 M NaCl, and 5 μL of proteinase K (0.4 mg/mL) were added and incubated under agitation for 1 h at 50 ◦C. After incubation, 200 μL of phenol–chloroform–isoamyl alcohol (25:24:1) was added to 500 μL of the solution, briefly vorticed, and then centrifuged at 12,000× *g* for 5 min. The aqueous phase was recovered, and 200 μL of ammonium acetate (7.5 M) and 500 μL (1 volume) of absolute ethanol were added to be mixed by inversion and precipitate at 4 ◦C overnight to centrifuge at 4 ◦C at 12,000× *g* for 15 min. The supernatant was discarded, and the pellet was washed twice with 100 μL of ethanol 70% (*v*/*v*). The DNA was dried at room temperature, resuspended in molecular-biology-grade water, and stored at −20 ◦C until use.

#### *2.9. Relative-Abundance Quantification by qPCR*

The relative abundance of the bacterial population was assessed through qPCR to determine the effect of treatments. The qPCR was performed on a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) according to the instructions of the iTaq™ Universal SYBR® Green Supermix (Bio-Rad, Hercules, CA, USA). The relative abundance of the total bacteria in the rhizosphere samples for each treatment was assessed according to the methodology described by López-Gutiérrez et al. [33] with slight modifications.

#### *2.10. Characterization of Bacterial Communities by Ribotype Assay Analysis (16S rRNA Gene)*

Ribotype assay analysis was conducted according to the Bogino et al. [34] methodology. A total DNA of 36 water samples (3 samples × 3 times × 2 treatments × 2 species of plants = 36 samples in total) and 60 bacterial isolate strains (30 isolate strains for each plant for both organic and inorganic fertilization treatments) were characterized by amplified ribosomal DNA restriction analysis (ARDRA). Bacterial genomic DNA was extracted from each isolate as mentioned previously. For 16S rRNA gene amplification, we used primers fD1 (5 -AGAGTTTGATCCTGGCTCAG-3 ) and rD1 (5 - AAGGAGGTGATCCAGCC-3 ). PCR amplification products (~1500 bp) were processed by a restriction endonuclease assay with HaeIII (Thermo Fisher Scientific), and the resulting fragments were electrophoretically separated on a 2% (*w*/*v*) agarose gel, stained with ethidium bromide to visualize them with UV radiation, and the corresponding image was photographed. Ribotype identification is directly associated with a specific restriction fragment fingerprint. The community structure dendrogram was constructed on the basis of ribotypes of the bacterial isolates with GelCompar II software. Bacterial isolate strains belonging to either unique majority ribotypes or common ribotypes were selected for further identification through 16S rRNA

gene nucleotide sequence analysis with primers COM 1 (5 -CAGCAGCCGCGGTAATAC-3 ) and COM 2 (5 -CCGTCAATTCCTTTGAGTTT-3 ) with the methodology described by Stach et al. [35]. The 16S rRNA gene sequences were analyzed using the BLAST (blastn) search program (National Center for Biotechnology Information (NCBI)).

#### *2.11. Biofilm-Formation Assay*

Biofilms are microbial communities that adhere to surfaces and are enclosed in a protective matrix; this is also the primary structure from which bacteria interact with plants and other eukaryotes. Thus, to characterize the bacterial capability of the rhizosphere (water samples) isolate strains from *M. spicata* and *R. officinalis* to form biofilms, we carried out the crystal violet (CV) staining quantitative assay of Labrie et al. [36] with slight modifications. CV staining absorbance was measured at 590 nm using a spectrophotometer (Multiskan Spectrum, Thermo Scientific, Wilmington, DE, USA).

#### *2.12. Statistical Analysis*

Data were analyzed using univariate and multivariate analysis of variance (ANOVA and MANOVA) for one-way classification, and the nutrition source was the study factor. For chlorophyll content, multiple analysis of variance (MANOVA) and significant differences between means for each recorded date were determined by two-way analysis of variance (ANOVA). Least significant differences (LSD) in Tukey's HSD test (*p* = 0.05) were estimated for one-way ANOVA. For all cases, significant differences between means were considered to be significant at *p* < 0.05. All statistical analyses were performed with Statistica software program v10.0 and GraphPad Prism version 6.0 (GraphPad Software, San Diego, CA, USA).

#### **3. Results**

#### *3.1. Plant Morphology and Photosynthetic Pigments*

#### 3.1.1. *M. spicata*

Stem height (SL), dry foliar weight (DFW), fresh foliar weight (FFW), foliar area (FA), and root length (RL) showed a significant increase in the vermicompost leachate treatment compared with the inorganic treatment for *M. spicata* (Table 2). There was no difference between the vermicompost leachate treatment and the inorganic treatment for relative growth rates of leaves (FGR), stems (RGS), total growth rate (TGR), and roots (RGR), which was lower for vermicompost leachate than inorganic fertilizer was (Table 3). Chlorophyll a and b, and total content did not show any differences between plants with vermicompost leachate or inorganic treatment (Table 4 and Figure 1).

**Figure 1.** Chlorophyll SPAD readings in *M. spicata* plants under leachates of inorganic and vermicompost leachate fertilizers. Vertical bars represented mean ± standard error.


**Table 2.** Morphometric parameters in *M. spicata* and *R. officinalis* plants under fertilization treatments.

SS: inorganic fertilizer, L: vermicompost leachate, SL: stem length, FSW: fresh stem weight, DSW: dry stem weight, FA: foliar area, FFW: fresh foliar weight, DFW: dry foliar weight, RL: root length, FRW: fresh root weight, DRW: dry root weight. Data represent means ± standard error (*n* = 3). *M. spicata* and *R. officinalis* data were treated as independent ANOVA analyses. Different letters for each column denote statistical difference.

**Table 3.** Total growth rate (TGR), foliar growth rate (FGR), root growth rate (RGR), and stem growth rate (SGR) expressed in grams per day of *M. spicata* and *R. officinalis* plants.


SS: inorganic fertilizer, L: organic fertilizer (vermicompost leachate). Data represent means ± standard deviation (*n* = 5). Different letters denote statistical differences.

> **Table 4.** Chlorophyll (Chl) a and b, and total (mg·mL<sup>−</sup>1) content in *M. spicata* and *R. officinalis* plants under different nutrient sources in two times before (BT) and after (AT) application of vermicompost leachate and inorganic treatments.


SS: inorganic fertilizer, L: organic fertilizer (vermicompost leachate). Data represent means ± standard deviation (*n* = 5). Different letters denote statistical differences. \* Denote statistical differences between sampling dates.

#### 3.1.2. *R. officinalis*

For all morphological traits, there were no differences between the vermicompost leachate and inorganic treatments (Tables 2 and 3) except for rosemary under treatment with leachate in RGR, which showed lower growth (Table 3). Organic treatment did not affect chlorophyll a and b, and total content did not undergo alterations in either organic or inorganic treatment, and the only variable that exerted an effect was the time (date) of chlorophyll sampling (Table 4).

#### *3.2. CFU Quantification and Relative Abundance of Bacterial Communities*

The relative abundance of total bacterial communities due to the effect of treatments was assessed by CFU estimation and by a qPCR-based assay. For both *M. spicata* and *R. officinalis*, no differences were determined between the vermicompost leachate and inorganic treatments regarding the abundance of bacterial populations; however, an increase in relative abundance in time was more evident for the vermicompost leachate (Figure 2).

Bacterial community structure kinetics between both vermicompost leachate and inorganic treatments was analyzed. Thirty-six total DNA water samples were analyzed by amplified ribosomal DNA restriction analysis (ARDRA). As this test showed for *M. spicata* and *R. officinalis*, bacterial community structures underwent changes through time without a significant effect between treatments (Figure 3a,b). Thus, these results highlight the feasibility of replacing inorganic fertilizer with the vermicompost leachate without significant impact on the bacterial abundance or bacterial community structures of *M. spicata* and *R. officinalis* in hydroponic systems.

**Figure 2.** Colony-forming unit (CFU) quantification and relative abundance (qPCR) of bacterial communities in *M. spicata* and *R. officinalis*. CFU quantification in (**a**) *M. spicata* and (**b**) *R. officinalis*; relative abundance (qPCR) of bacterial communities in (**c**) *M. spicata* and (**d**) *R. officinalis* (M1SS: mint composed sample, time 1, inorganic fertilizer; M2SS: *M. spicata* composed sample, time 2, inorganic fertilizer; M3SS: *M. spicata* composed sample, time 3, inorganic fertilizer; M1L: *M. spicata* composed sample, time 1, vermicompost leachate; M2L: *M. spicata* composed sample, time 2, vermicompost leachate; M3L: *M. spicata* composed sample, time 3, vermicompost leachate; R1SS: *R. officinalis* composed sample, time 1, inorganic fertilizer; R2SS: *R. officinalis* composed sample, time 2, inorganic fertilizer; R3SS: *R. officinalis* composed sample, time 3, inorganic fertilizer; R1L: *R. officinalis* composed sample, time 1, vermicompost leachate; R2L: *R. officinalis* composed sample, time 2, vermicompost leachate; R3L: *R. officinalis* composed sample, time 3, vermicompost leachate.

**Figure 3.** Dendrogram of general distribution of bacterial composition of communities between treatments in (**a**) *M. spicata* and (**b**) *R. officinalis* (M1SS: *M. spicata* composed sample, time 1, inorganic fertilizer; M2SS: *M. spicata* composed sample, time 2, inorganic fertilizer; M3SS: *M. spicata* composed sample, time 3, inorganic fertilizer; M1L: *M. spicata* composed sample, time 1, vermicompost leachate; M2L: *M. spicata* composed sample, time 2, vermicompost leachate; M3L: *M. spicata* composed sample, time 3, vermicompost leachate; R1SS: *R. officinalis* composed sample, time 1, inorganic fertilizer; R2SS: *R. officinalis* composed sample, time 2, inorganic fertilizer; R3SS: *R. officinalis* composed sample, time 3, inorganic fertilizer; R1L: *R. officinalis* composed sample, time 1, vermicompost leachate; R2L: *R. officinalis* composed sample, time 2, vermicompost leachate; R3L: *R. officinalis* composed sample, time 3, vermicompost leachate).

#### *3.3. Composition and Diversity of Bacterial Communities*

A total of 60 bacterial isolate strains (30 isolate strains for each plant for both vermicompost leachate and inorganic fertilization treatments) were characterized by ARDRA. From ARDRA, 15 ribotypes were identified in *M. spicata* and *R. officinalis* according to the yielded fingerprint after the restriction assay with the HaeIII restriction enzyme (Table 5). In the case of *R. officinalis*, eight different ribotypes were identified (Figure 4). Of these eight ribotypes, seven were present in inorganic treatment, and four in the vermicompost leachate. Of the ribotypes present in the inorganic treatment, four were exclusively present in this treatment, while only one ribotype was exclusive of the vermicompost leachate. In the case of *M. spicata*, there were also eight different ribotypes for both the vermicompost leachate and the inorganic treatment. For the inorganic treatment, there were five ribotypes, and none was exclusive to this treatment. For the vermicompost leachate treatment, eight ribotypes were present, and three ribotypes were exclusive of this treatment. However, it was not possible to characterize the ribotype to which three bacterial isolates from *M. spicata* belonged (two from inorganic treatment and one from organic treatment).

Representative bacterial strains were identified by 16S rRNA gene sequencing. Bacterial isolate strains were selected according to ribotype ARDRA profiles (Table 6). Most bacterial isolate strains belonged to the Firmicutes phylum, which was mainly composed of the Bacilli class, the Bacillaceae family, and the *Bacillus* genus. Bacterial isolate strains belonging to Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria classes from the Proteobacteria phylum were found (Table 6).

**Figure 4.** Ribotypes present in *M. spicata* and *R. officinalis* obtained by amplified ribosomal DNA restriction analysis (ARDRA; R: ribotype, number: number of ribotypes, and number in parentheses: number of isolates corresponding to each ribotype).


**Table 5.** Ribotypes of bacteria isolated from hydroponic system in *M. spicata* and *R. officinalis* plants.

SS: inorganic fertilizer, L: vermicompost leachate, MSS-number or MSSR-number: isolates from *M. spicata* inorganic fertilizer, ML-number or MLR-number: isolates from *M. spicata* vermicompost leachate, RSS-number or RSSR-number: isolates from *R. officinalis* inorganic fertilizer, and RL-number or RLR-number: isolates from *R. officinalis* vermicompost leachate. Note: MSSR2, MSSR3, and ML3 are missing from the table because they were unclassified.


Rt: ribotype.

Ribotypes found in rosemary bacterial isolate strains belonged to Firmicutes (60%), mainly composed of the *Bacillus* genus. Comparing the vermicompost leachate and inorganic treatments, we determined that the Firmicutes phylum was the most abundant between treatments, and the Alphaproteobacteria and Betaproteobacteria classes, and Gammaproteobacteria showed greater abundance in inorganic treatment than in the vermicompost leachate treatment (Figure 4, Table 6). The ribotypes found in *M. spicata* bacterial isolate strains belonged to Firmicutes (80% and were mainly composed of the *Bacillus* genus. Interestingly, 10% of the bacterial isolate strains were unclassified. Comparing the vermicompost leachate and inorganic treatments, the most abundant phylum was Firmicutes, followed by the Gammaproteobacteria class (Tables 5 and 6). For the vermicompost leachate, the Betaproteobacteria class showed greater abundance in the vermicompost

leachate treatment than in inorganic treatment (Tables 5 and 6). Therefore, the Firmicutes phylum was the most abundant in both *R. officinalis* and *M. spicata* plants, and in both the vermicompost leachate and the inorganic treatment.

#### *3.4. Biofilm-Forming Ability of Bacterial Communities*

All bacterial isolate strains from *R. officinalis* (30 isolates) and *M. spicata* (30 isolates) were assessed for adhesion and biofilm-establishment capability with a CV assay. The CV assay showed that all bacterial isolates were able to adhere to the surface and establish biofilms (Figure 5). Differences were found in biofilm formation that were categorized according to the capability to retain CV measured by the OD at 595 nm (CV-OD595) [28], for all bacterial isolate strains as follows: weak (<0.6), moderate (0.6–1.2), and strong (>1.2). *R. officinalis* bacterial isolate strains with the vermicompost leachate treatment showed that 3 bacterial isolates formed a moderate biofilm, 2 a strong biofilm, and the remaining 10 a weak biofilm. For the bacterial isolate strains from the inorganic treatment, 4 bacterial isolates formed a moderate biofilm, 1 a strong biofilm, and the remaining 10 a weak biofilm. The *M. spicata* bacterial isolate strains with the vermicompost leachate treatment showed that 1 bacterial isolate formed a strong biofilm, 2 a moderate biofilm, and the remaining 12 formed a weak biofilm. For the inorganic treatment, 2 bacterial isolates were able to form a strong biofilm, 1 a moderate biofilm, and the remaining 12 a weak biofilm. Altogether, for the *R. officinalis* and *M. spicata* plants and both the vermicompost leachate and the inorganic treatment, most bacterial isolates were able to form weak biofilms in the conditions assessed in this study.

**Figure 5.** Biofilm formation quantified by staining with crystal violet of isolates from (**a**) *R. officinalis* and (**b**) *M. spicata* (RSS-number or RSSR-number: isolates from *R. officinalis* inorganic fertilizer, RL-number or RLR-number: isolates from *R. officinalis* vermicompost leachate, MSS-number or MSSR-number: isolates from *M. spicata* inorganic fertilizer, and ML-number or MLR-number: isolates from *M. spicata* vermicompost leachate.

#### **4. Discussion**

The vermicompost leachate treatment for both *M. spicata* (mint) and *R. officinalis* (rosemary) plants did not affect their growth; even for *M. spicata* plants, we were able

to determine a growth increase for several morphometric parameters. Moreover, for *R. officinalis* plant growth, for all morphometric parameters, there were only differences for root growth, which was lower for vermicompost than for inorganic leachate; similar results were found by Peng et al. [37]. This is important since the aim of healthy food production is avoiding the application of inorganic fertilizer [25,38–41]. Furthermore, vermicompost leachate contains a high amount of plant hormones, such as auxins, gibberellins, and cytokinins from microbial origin, giving rise to plant-growth enhancement, and acting as a liquid fertilizer [15,42–45]. Emperor and Kumar [45] determined that organic matter processed in the earthworm gut and then excreted as vermicast undergoes an increased level of microbial population, microbial respiration, microbial enzyme activity, and N, P, and K enrichment, bacterial exopolysaccharide production, lignocellulolytic activity establishment, nitrifying, and nitrogen-fixing microorganism proliferation. The above allow for us to conclude that the use of vermicompost to replace inorganic fertilizers is a viable option under the use of hydroponic systems [43,46–49].

The bacterial communities' relative abundance showed no differences between the vermicompost leachate and inorganic treatments for both *R. officinalis* and *M. spicata* plants, showing time-related differences, as expected, in accordance with previous works, where the analyzed bacterial communities underwent the same behavior [50,51]. The bacterialcommunity structure for the *R. officinalis* and *M. spicata* plants and for both treatment types were mainly composed by the Firmicutes phylum, followed by the Proteobacteria phylum, which was represented by the Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria classes; we were also able to determine the presence of beneficial bacteria from the *Bacillus* (Firmicutes phylum) and *Pseudomonas* (Proteobacteria phylum) genera. Those bacteria are designated as beneficial or plant-growth-promoting (PGPB), and the characterization of the bacterial-community structures of the rhizosphere for other plant members (*Thymus vulgaris*, *T. citriodorus*, *T. zygis*, *Santolina chamaecyparissus*, *Lavandula dentata*, and *Salvia miltiorrhiza*) of the Lamiaceae family showed that Proteobacteria, Firmicutes, Bacteroidetes, Actinobacteria, Acidobacteria, and Gemmatimonadetes were among the most abundant bacterial phyla [5,52–56].

Lastly, the capability to establish biofilms was assessed for all 60 bacterial isolate strains from the *M. spicata* and *R. officinalis* plants and both treatments, with no differences highlighting the essential role of biofilm development in bacterial survival and physiology [36]. We determined that most of the isolates (66.67% in *R. officinalis* and 80% in *M. spicata*) had weak capacity (CV-OD595) to form a biofilm; a smaller proportion were able to produce a strong biofilm for both *M* plants and both treatments. In an aqueous environment, such as a hydroponic system, biofilm establishment follows other mechanisms that are not yet characterized. Authors should discuss the results and how they can be interpreted from the perspective of previous studies and working hypotheses. The findings and their implications should be discussed in the broadest context possible. Future research directions may also be highlighted.

#### **5. Conclusions**

In this study, we showed that the substitution of inorganic fertilizer by vermicompost leachate in a hydroponic system allows for us to maintain or increase the production of two crop plants with agricultural importance (*M. spicata* (mint), and *R. officinalis* (rosemary)). Furthermore, we determined that this fertilizer substitution modifies neither the bacterial communities for both plants nor their ability to form biofilms. Through time, the vermicompost leachate tendency showed an increase in relative abundance, which is important to consider for future studies. Therefore, we propose the use of vermicompost leachate fertilizer as a feasible replacement for inorganic fertilizer in hydroponic systems to achieve sustainable and ecofriendly agricultural production, in agreement with our results and recent research conducted on open-field cultures, to face the challenge of a growing population and pollution derived from the use of inorganic fertilizers.

**Author Contributions:** A.L.-M.: conceptualization, methodology, formal analysis, investigation, writing—original draft. A.B.: writing—review and editing. G.C.-C.: writing—review and editing. E.T.-D.: supervision and writing-review. B.M.-A.: Funding acquisition. A.N.-G.: conceptualization, formal analysis, writing—review and editing, funding acquisition, project administration. G.L.-V.: methodology. All authors have read and agreed to the published version of the manuscript.

**Funding:** The current investigation was supported by CONACYT/Mexico through the Ciencia Básica SEP-CONACYT grant No. 236240, and funds provided to Centro de Investigaciones Biologicas del Noroeste S.C. (CIBNOR) and Consejo Sudcaliforniano de Ciencia y Tecnología grant 20403.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data are available for transparency at www.cibnor.gob.mx (anieto04@ cibnor.mx, aloera04@cibnor.mx) (accessed on 27 April 2021).

**Acknowledgments:** Current investigations from the group are supported by CONACYT/México. We thank the technical assistance to all the team of the experimental field of the Agriculture Program in Arid Zones, Saúl Edel Briseño Ruiz, Pedro Luna García, Adrián Jordán Castro, and Raymundo Ceseña Nuñez. We also thank the technical assistance of Lidia Hirales-Lucero of Fitotecnia Laboratory and Angel Edgardo Carrillo García of Microbial Molecular Ecology Laboratory.

**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.

#### **References**


### *Article* **Calcium Carbonate Can Be Used to Manage Soilless Substrate pH for Blueberry Production**

**Michael J. Schreiber and Gerardo H. Nunez \***

Horticultural Sciences Department, University of Florida, Gainesville, FL 32611, USA; schreiberm@ufl.edu

**\*** Correspondence: g.nunez@ufl.edu

**Abstract:** Blueberry (*Vaccinium corymbosum* interspecific hybrids) production in soilless substrates is becoming increasingly popular. Soilless substrates have low pH buffering capacity. Blueberry plants preferentially take up ammonium, which acidifies the rhizosphere. Consequently, soilless substrates where blueberry plants are grown exhibit a tendency to get acidified over time. Agricultural lime (CaCO3) is commonly used to raise soil and substrate pH in other crops, but it is rarely used in blueberry cultivation. We hypothesized that substrate amendment with low rates of agricultural lime increases substrate pH buffering capacity and provides nutritional cations that can benefit blueberry plants. We tested this hypothesis in a greenhouse experiment with 'Emerald' southern highbush blueberry plants grown in rhizoboxes filled with a 3:1 mix of coconut coir and perlite. We found that substrate amendment with CaCO3 did not cause high pH stress. This amendment maintained substrate pH between 5.5 and 6.5 and provided Ca and Mg for plant uptake. When blueberry plants were grown in CaCO3-amended substrate and fertigated with low pH nutrient solution (pH 4.5), they exhibited greater biomass accumulation than plants grown in unamended substrates. These results suggest that low rates of CaCO3 could be useful for blueberry cultivation in soilless substrates.

**Keywords:** *Vaccinium corymbosum*; container; ammonium uptake; southern highbush blueberry

#### **1. Introduction**

Cultivation in containers filled with soilless substrates is rapidly becoming a popular growing system for blueberry (*Vaccinium corymbosum* interspecific hybrids) production. Soilless substrates based on sphagnum peat moss or coconut coir are generally acidic [1] and have high water holding capacity [2]. These substrate characteristics promote blueberry nutrient uptake and support vigorous growth [3,4]. As this growing system becomes widespread [5], there is a need for research focused on fertilization and management practices for substrate-grown blueberry.

Sphagnum peat moss and coconut coir have low pH buffering capacity [6,7]. Consequently, pH changes of up to 1 unit per month are not uncommon [3,4,8,9]. While blueberry roots exhibit limited ability to change the rhizosphere pH through H+ extrusion [10], ammonium uptake can lead to rapid rhizosphere acidification [11,12]. Considering blueberry growth and N content are enhanced by ammonium-based fertilization [11,13], substrate acidification appears inevitable in this production system.

Calcitic (CaCO3) and dolomitic [CaMg(CO3)2] lime are commonly used to raise soilless substrate pH, but amendment rates and effects are crop-specific (reviewed in [14]). The carbonate moiety in lime acts as a buffer that maintains the rhizosphere approximately at pH 6.4 [14]. The cations in lime are nutritionally relevant Ca and Mg. Substrates used for cultivation of other acid-loving plants are routinely amended with lime to limit substrate pH change [15,16]. Nevertheless, the effects of lime amendments in substrate-grown blueberry remain understudied.

Lime is rarely used in soil-based blueberry cultivation because high liming rates can raise soil pH excessively and cause plant stress. When grown in high pH soils, blueberry

**Citation:** Schreiber, M.J.; Nunez, G.H. Calcium Carbonate Can Be Used to Manage Soilless Substrate pH for Blueberry Production. *Horticulturae* **2021**, *7*, 74. https://doi.org/ 10.3390/horticulturae7040074

Academic Editors: Nazim Gruda and Juan A. Fernández

Received: 2 March 2021 Accepted: 2 April 2021 Published: 7 April 2021

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

**Copyright:** © 2021 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/).

plants exhibit nutritional deficiencies, stunted growth, and lower yields [17–19]. Lime is only used in situations where soil pH is very low to deliver Ca and Mg [20]. Hence, lime amendments must be meticulously used to avoid stressing blueberry plants.

This research investigates the effect of substrate amendment with CaCO3 on the substrate pH, growth, and nutrition of southern highbush blueberry. We hypothesized that substrate amendment with low rates of agricultural lime increases substrate pH buffering capacity and provides nutritional cations that can benefit blueberry plants. We tested this hypothesis in a greenhouse experiment with plants grown in rhizoboxes.

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

Rooted cuttings of 'Emerald' southern highbush blueberry (SHB; rooting volume = 3 cm3, average dry weight = 1.15 g, average height = 12 cm) were acquired from a commercial micropropagation nursery (Agristarts LLC, Apopka, FL, USA) and transplanted to benchtop rhizoboxes as per [21]. Rhizoboxes were built using two 35.56 cm × 35.56 cm plexiglass panels spaced 1.9 cm apart using wood inserts. Each rhizobox contained approximately 1.7 L of substrate and was irrigated or fertigated by two 1.89 l·h−<sup>1</sup> pressure-regulating emitters, spaced approximately 15.25 cm apart. Custom-made rhizobox stands kept roots in the dark at 33◦ inclination. There was one plant per rhizobox. Rhizoboxes were used as a tool to study root growth patterns in response to substrate amendment and fertigation pH treatments.

Rhizoboxes were filled with a 3:1 mixture of coconut coir (SpongEase™, Enroot Products LLC, Cromwell, CT, USA) and horticultural grade perlite (American Garden Perlite, LLC, Lake Wales, FL, USA) pre-treated to deliver two substrate amendment treatments. In one treatment, substrate was amended with CaCO3 (Garden Lime, Austinville Limestone, Austinville, VA, USA) at a rate of 6.18 Kg·m−3. This rate corresponds to half of the rate used in [22] where lime amendments were used to stress azalea (*Rhododendron* spp.). In the other treatment, substrate was amended with Ca-containing fertilizer produced from neutralized CaCO3 (Calexin ®, Miller Chemical & Fertilizer Corporation, Hanover, PA, USA) at a rate of 100.3 L·m<sup>−</sup>3. Guaranteed analysis and product density information were used to calculate a Calexin application rate that delivered the same amount of Ca as the CaCO3 amendment. Both amendments were incorporated into moist substrate 7 days before transplant.

Fertigation solution pH was a second variable in the experiment. Plants were fertigated with a solution containing 0.5 mM (NH4)2SO4, 0.5 mM K2PO4, 1.0 mM MgSO4, 0.5 mM CaCl2, 0.045 mM H3BO3, 0.01 mM MnSO4, 0.01 mM ZnSO4 with 0.3 mM CuSO4, 0.2 mM Na2MoO4, and 45 mM Fe provided as Sequestrene 330 (10% iron(III)-diethylenetriamine pentaacetic acid) (Becker Underwood, Inc., Ames, IA, USA). Ammonium was the only form of N provided, in agreement with industry practices [20]. The low N rate was selected because blueberry microcuttings exhibited ammonium toxicity when fertigated with higher N rates in a preliminary experiment. Fertigation solution was buffered using 5.0 mM 2-(4-morpholino)-ethane sulfonic acid to pH 4.5 or pH 6.5 using HCl or KOH. These fertigation pH treatments are referred to as low pH and high pH respectively in relation to fertigation pH used in previous studies [3]. There were 21 fertigation/irrigation events per week. Each plant received 1.75 L of fertigation solution (delivered through 7 events) and 3.96 L of irrigation water per week (delivered through 14 events). Fertigation events preceded irrigation events. Fertigation and irrigation volumes were measured with graduated cylinders connected directly to emitters.

Substrate samples were collected at the start (day 0) and end (days 75–77) of the experiment and submitted for analysis at a commercial laboratory (Waters Agricultural Laboratory, Camila, GA, USA). Ca, Mg, and K concentrations in the substrate were determined using inductively coupled plasma mass spectrometry [23]. Cation exchange capacity (CEC) was calculated from K, Ca, Mg, and H concentrations as per [24]. Substrate pH was measured in a 1:1 substrate:deionized water slurry [25].

Substrate pH and electrical conductivity (EC) were monitored using the pour-through method [26]. Deionized water samples that eluted through the substrate (hereon, leachate) were collected on a weekly basis (*n* = 3 per treatment). Rhizoboxes were removed from the stand and placed vertically on top of plastic trays (one rhizobox per tray) approximately 2 h after the last fertigation event. Then, 500 mL of deionized water were slowly poured on top of the substrate. Leachate was collected in the plastic tray for approximately 20 min. Then, leachate volume was measured with a graduate cylinder and 50 mL aliquots were transported to the laboratory for immediate measurement of leachate pH (Accumet AP110 Portable pH Meter, Thermo Fisher Scientific, Hampton, NH, USA) and EC (Accumet Excel Conductivity Meter, XL30, Thermo Fisher Scientific, Hampton, NH, USA) using standardized electrodes. In this manuscript and elsewhere [26], it is assumed that leachate pH and EC represent rhizosphere conditions. Leaf greenness was measured on the youngest fully expanded leaf of each plant using a SPAD-502 meter (Konica Minolta, Inc., Ramsey, NJ, USA).

Rhizoboxes were scanned using a flatbed scanner (LX1100, Seiko Epson Corp., Tokyo, Japan) at a resolution of 1000 dots per inch (dpi). The scanner was held at an inclination of 30◦ during scanning to avoid substrate loss. Rhizoboxes were scanned on a weekly basis starting on week 2 of the experiment. Rhizobox images were used to measure root system convex hull area using ImageJ version 1.51 [27]. Convex hull area is the area of the polygon formed by lines connecting the most distal root tips in a plant. Root system spread was computed as the ratio of the convex hull area to root dry weight.

Rhizoboxes were disassembled 75 to 77 days after the start of the experiment. Roots were washed clean of substrate using tap water. A subset of the root systems (*n* = 4 per treatment except for CaCO3 + pH 6.5 where one root image was lost due to human error) were scanned floating in water using the transparency unit of the flatbed scanner at 1000 dpi. Images were divided into 5 tiles using ImageJ. Then, total root length was determined using WinRhizo Pro 2013b (Regent Instruments, Quebec, QC, Canada). Organ and whole plant fresh weight were measured. Leaves were laid flat and photographed at a distance of 48.25 cm from the lens using mobile phone cameras (iPhone 7 and iPhone X, Apple Inc., Cupertino, CA, USA) on a white background with a scale bar of known size. Total leaf area was measured using ImageJ. Plant tissues were weighted after drying at 72 ◦C for a week. Dry tissue was ground until it passed through a size-20 mesh (sieve opening = 0.841 mm). Then, tissue was submitted for elemental analysis at a commercial laboratory (Waters Agricultural Laboratory, Camila, GA, USA).

The experiment was conducted in a greenhouse where average temperature and relative humidity were 22.53 ◦C and 70.19%, respectively. The experiment followed a completely randomized design with treatments in a 2 × 2 factorial arrangement. There were 10 single-plant replications per amendment × pH combination. Unless otherwise stated, *n* = 10 per treatment. Treatment effects on biomass accumulation, leaf area, substrate characteristics, elemental content, and root traits were assessed using two-way analysis of variance (R package agricolae, [28]). Where significant effects were identified, pairwise comparisons were made using the least significant difference method. Leachate pH and EC data were analyzed through linear mixed-effect analysis (R package lme4, [29]). Fertigation solution pH, substrate amendment, and their interaction were considered fixed effects. Repeated measures per plant and week were considered random sources of error. Leachate pH and EC were response variables analyzed in separate models. Statistical significance was determined by likelihood ratio tests comparing the full model against a model without the effect being investigated. All statistical analyses were conducted in R version 3.6.2 [30]. Data were illustrated using ggplot 2 [31].

#### **3. Results**

At the start of the experiment, substrates amended with CaCO3 exhibited higher pH, percentage base saturation, Mg content, and K content than substrates amended with Calexin (Table 1). Substrate CEC and Ca content were not different between the amendment treatments. At the end of the experiment, substrate pH was not different among treatments (Table 2). Substrates amended with CaCO3 exhibited higher CEC, percentage base saturation, Ca concentration, and Mg concentration than substrates amended with Calexin. Substrates that were fertigated at pH 4.5 exhibited lower K concentration than substrates that were fertigated at pH 6.5. The interaction of substrate amendment and fertigation pH did not affect substrate characteristics (*p* ≥ 0.158).

**Table 1.** Substrate characteristics before transplant. A substrate composed of a 3:1 mixture of coconut coir and perlite was amended with CaCO3 or a Ca-containing fertilizer (Calexin) 7 days before transplanting 'Emerald' southern highbush blueberry.


<sup>z</sup> Treatments were compared using ANOVA.

**Table 2.** Substrate characteristics after 75–77 days of growing 'Emerald' southern highbush blueberry with contrasting substrate amendments and fertigation pH.


<sup>z</sup> Data were analyzed by two-way ANOVA. The interaction of fertigation pH and substate amendment did not affect substrate characteristics (*p* ≥ 0.158).

> Substrate amendments and fertigation pH created contrasting leachate pH and EC during most of the experiment (Figure 1). Leachate pH gradually decreased in all treatments (χ<sup>2</sup> = 11.74, df = 1, *<sup>p</sup>* ≤ 0.001, estimate = −0.42). Amendment with CaCO3 (χ<sup>2</sup> = 93.34, df = 1, *p* < 0.001) and high pH fertigation (χ<sup>2</sup> = 28.69, df = 1, *p* < 0.001) led to high leachate pH. The interaction of substrate amendment and fertigation pH did not affect leachate pH (χ<sup>2</sup> = 2.57, df = 3, *p* = 0.11). Amendment with CaCO3 led to higher leachate EC (χ<sup>2</sup> = 5.26, df = 1, *p* = 0.02). Fertigation pH (χ<sup>2</sup> = 1.13, df = 1, *p* = 0.29), time (χ<sup>2</sup> = 0.22, df = 1, *p* = 0.63), and the interaction of substrate amendment and fertigation pH (χ<sup>2</sup> = 7.33, df = 3, *p* = 0.06) did not affect leachate EC.

> Substrate amendments and fertigation pH affected plant biomass accumulation (Table 3). Plants grown with a combination of Calexin amendment and low pH fertigation solution exhibited lower cane, leaf, and total dry weight than plants grown with CaCO3 amendments. Within a substrate amendment, fertigation pH did not affect biomass accumulation. Plants grown in substrates amended with CaCO3 exhibited larger root systems than plants grown in substrates amended with Calexin. Leaf area followed the same trends as leaf dry weight (data not shown). Leaf greenness was not affected by the treatments (average = 24.68, *p* = 0.23).

**Figure 1.** Leachate pH (**A**) and electrical conductivity (EC) (**B**) collected from rhizoboxes where 'Emerald' southern highbush blueberry grew with contrasting substrate amendments (CaCO3 or Ca-containing fertilizer Calexin) and fertigation pH (pH 4.5 and pH 6.5).

**Table 3.** Biomass accumulation of 'Emerald' southern highbush blueberry plants grown in rhizoboxes with contrasting substrate amendments and fertigation pH.


<sup>z</sup> Data were analyzed by two-way ANOVA. Means followed by the same letter were not significantly different according to Tukey LSD at α = 0.05.

> Substrate amendment and fertigation pH also affected root system characteristics. Root systems of plants grown with low pH fertigation and CaCO3 amendments exhibited larger convex hull areas than all other treatment combinations between weeks 3 and 9 (Figure 2A). Root systems of plants grown with low pH fertigation and Calexin amendments had smaller convex hull area than all other treatments initially (weeks 3 and 4). Plants grown with low pH fertigation and CaCO3 amendment exhibited higher total root length than plants grown with high pH fertigation and CaCO3 amendments and plants grown with low pH fertigation and Calexin (Figure 2B). High pH fertigation so

lution (274.53 cm2·g−<sup>1</sup> vs. 191.66 cm2·g−1) and CaCO3 amendments (338.97 cm2·g−<sup>1</sup> vs. 127.22 cm2·g−1) reduced root system spread (*<sup>p</sup>* < 0.008 in all cases). Root system spread was not affected by the interaction of substrate amendment and fertigation solution pH (*p* = 0.29).

**Figure 2.** Root system characteristics of 'Emerald' southern highbush blueberry grown with contrasting substrate amendments and fertigation pH. (**A**) Convex hull area during the treatment period. (**B**) Total root length after 77 days of cultivation. Means followed by the same letter were not significantly different according to Tukey LSD at α = 0.05.

Substrate amendment and fertigation pH affected root and leaf nutrient concentrations (Table S1). High pH fertigation decreased N, Zn, and Cu concentrations and increased Ca concentration in roots. Substrate amendment with CaCO3 decreased K, S, B, and Cu concentrations and increased Fe concentrations in roots. Other elements were not affected. The interaction of fertigation pH and substate amendment did not affect root nutrient concentrations (*p* ≥ 0.078). Plants grown in substrates amended with Calexin and fertigated with low pH solution exhibited the highest leaf N, P, Ca, Mg, S, Fe, Zn, and Cu concentrations (Table S2). Plants grown in substrates amended with CaCO3 generally exhibited the lowest leaf concentrations of these elements. Plants grown with low pH fertigation exhibited higher leaf Mn concentrations than plants grown with high pH fertigation. Plants grown in substrates amended with CaCO3 exhibited lower leaf K, Mn, and B. With the exception of K, treatment effects on nutrient concentration did not exhibit the same trends in roots and leaves.

#### **4. Discussion**

Soilless substrates have limited pH buffering capacity [6], which allows large pH changes over the cultivation period [3,4,8,9]. In this experiment, 'Emerald' SHB plants were fertigated with a nutrient solution where ammonium was the only form of N. Ammonium uptake leads to rhizosphere acidification [11,12]. As expected, leachate pH gradually decreased in all treatments. Similar leachate acidification has been previously observed in experiments with substrate-grown blueberry [4,12] and azalea [32].

CaCO3 is routinely used to raise soil or substrate pH in other crops [14–16], but not in blueberry. When blueberry and other acid-loving plants are grown in soils or substrates amended with high CaCO3 rates, they exhibit high pH stress symptoms such as interveinal chlorosis and stunted growth [17–19,33]. In this experiment, CaCO3 in the substrate did not cause high pH stress in 'Emerald' SHB, probably due to the low rate used. Plants grown in substrates amended with CaCO3 did not exhibit Fe deficiency symptoms either, but leaf Fe concentrations were lower than published recommendations [20]. These results suggest that even though CaCO3 raised substrate pH, the effect was mild enough to avoid causing high pH stress in 'Emerald' SHB. Further research will be necessary to determine if the CaCO3 rate used here is appropriate for other blueberry varieties.

In this experiment, CaCO3 in the substrate acted as a pH buffer that partially neutralized H+ from ammonium uptake and the fertigation solution, maintaining leachate pH between pH 5.5 and pH 6.5 for most of the experiment. Additionally, CaCO3 amendment replaced cations from the substrate adhesion sites with Ca and Mg. The combination of acidic substrate and nutritional cation availability supported vigorous growth above- and below-ground in 'Emerald' SHB, especially when CaCO3 amendment was matched with low pH fertigation solution.

When the substrate did not contain carbonates, leachate pH ranged between pH 4.5 and pH 5.0 and almost half of the adhesion sites were occupied by non-nutritional cations. The lack of nutritional cations was likely caused by the abundance of H+ and/or Calexin leaching out of the rhizoboxes. These substrate conditions affected shoot and root growth, particularly when the fertigation solution pH was low. Low pH increases Al solubility [34], which can cause Al toxicity in blueberry [35]. Perlite contains 10–15% Al2O3 [36]. Thus, it is possible that Al toxicity might have affected 'Emerald' SHB growth when substrate pH was extremely low. Al concentrations in the rhizosphere were not measured in this experiment. Further research will be necessary to establish if Al toxicity impacted plant responses.

Substrate characteristics affected blueberry root abundance and distribution. CaCO3 amendments increased 'Emerald' SHB root dry weight and, in combination with low pH fertigation, they led to large root systems that reached most of the substrate in the rhizoboxes. Nevertheless, large root systems were not always better at taking up nutrients. Previous research has shown that CaCO3 can affect nutrient uptake through pH-dependent and pH-independent effects [37]. Thus, fertilization practices might need to be adapted to maintain optimum plant nutrition in substrates amended with CaCO3.

Irrigation water can contain carbonates and bicarbonates (collectively called alkalinity). Alkaline water sources are not uncommon in blueberry production areas [38], but water acidification through sulfuric acid injection or sulfur dioxide generators is routinely used to neutralize alkalinity [39]. Our results suggest that increasing substrate pH buffering capacity can be beneficial for blueberry. Thus, it is important to recognize the tradeoff between irrigation water pH and pH buffering capacity when irrigation water is acidified. Considering the substrate acidification tendency observed here and elsewhere [4,12], it is tempting to speculate on the utility of alkaline water for substrate pH management. Future research should explore this potential management strategy.

Altogether, our results indicate that substrate amendment with low rates of CaCO3 is a viable tool to increase pH buffering capacity in coconut coir-based substrates used for blueberry cultivation. CaCO3 neutralized H<sup>+</sup> and contributed Ca and Mg for plant uptake. Access to a weakly acidic substrate with abundant nutritional cations supported vigorous

growth in 'Emerald' SHB. Further research should evaluate other CaCO3 amendment rates and other blueberry varieties to facilitate decision making when using CaCO3 in substrate-based blueberry cultivation.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2311-752 4/7/4/74/s1, Table S1: Nutrient concentrations in roots of 'Emerald' southern highbush blueberry, Table S2: Nutrient concentrations in leaves of 'Emerald' southern highbush blueberry.

**Author Contributions:** Conceptualization, G.H.N.; methodology, data collection, analysis, and visualization, writing, M.J.S. and G.H.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data presented in this study are available on request from the corresponding author.

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

#### **References**


### *Article* **Organic Fertilization of Growing Media: Response of N Mineralization to Temperature and Moisture**

**Patrice Cannavo 1,\*, Sylvie Recous 2, Matthieu Valé 3, Sophie Bresch 4, Louise Paillat 1, Mohammed Benbrahim <sup>5</sup> and René Guénon <sup>1</sup>**


**Abstract:** Managing plant fertilization is a major concern of greenhouse growers when it comes to sustainable production on growing media. Organic fertilization is popular, but more difficult to control since organic compounds first need to be mineralized by microbes. The objective of this study was to characterize the time course of N mineralization by different fertilizer–growing media pairs, in the absence of plants. Several incubations were carried out at four temperatures (4, 20, 28, and 40 ◦C) and three suction potentials (−3.2, −10, and −31.6 kPa) on four growing media under two organic fertilization conditions to study the dynamics of NH4 <sup>+</sup> and NO3 − production. The results showed that the release of mineral N was strongly dependent on growing media, temperature, humidity, and fertilizer nature, varying from 10.7% to 71.3% of the N fertilizer applied. A temperature action law was established for the four growing media. The Q10 value of the growing media was 1.13, lower than the average Q10 value of arable soils. On the other hand, the specific behavior of the growing media did not yield a single humidity action law. Nevertheless, the nitrification process, evaluated by analyzing the ratio of NO3 − to total mineral N, showed a humidity-dependent relationship common to the four growing media and comparable to admitted observations on soils. Nitrification was optimal when growing media humidity was higher than 0.46 *v/v*.

**Keywords:** NH4 +; NO3 −; nitrification; Q10; modeling

57

#### **1. Introduction**

Consumers are concerned about food quality and the environmental impact of its production. The subject is thorny in horticulture, particularly in soilless production which consumes resources (water and other inputs). As a consequence, producers are moving toward agro-ecological practices such as organic fertilization and the development of growing media from renewable organic materials [1]. Indeed, organic fertilization introduces a recycling concept in agroecosystems, and the non-use of synthetic inorganic N fertilizers considerably reduces the CO2 emissions produced during the industrial N2 fixation.

In conventional soilless production (cultivation in pots and containers), the plant grows in a finite volume of a growing medium with limited buffering capacity for water, temperature, and pH in particular [2]. The physical, chemical, and, to a lesser extent, biological properties of growing media materials have been investigated over the last 40 years, but practical considerations have been relatively little investigated [3]. Professionals have good knowledge of the physicochemical properties of the growing media, allowing for the control of irrigation and mineral fertilization. Introducing organic fertilizers requires

**Citation:** Cannavo, P.; Recous, S.; Valé, M.; Bresch, S.; Paillat, L.; Benbrahim, M.; Guénon, R. Organic Fertilization of Growing Media: Response of N Mineralization to Temperature and Moisture. *Horticulturae* **2022**, *8*, 152. https://doi.org/10.3390/ horticulturae8020152

Academic Editors: Nazim Gruda and Juan A. Fernández

Received: 12 December 2021 Accepted: 7 February 2022 Published: 10 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/).

adapted practices because organic fertilizers first have to be mineralized by the microbiota of the growing media before being assimilated by plants.

Matching the rate of nutrient release by micro-organisms to the plant demands is essential [4,5]. Although microbial communities are widely used in growing media, few studies have characterized them. The authors of [6,7] studied microbial communities in peat, coir, and wood fiber growing media. They showed that organic growing media display specific activities and microbial structures depending on their origin and manufacturing process. Organic nutrient sources may be single or blended, and they may come from plant or animal byproducts or allowable mined sources [5]. Solid organic fertilizers are often unbalanced in their nutrient content, especially in their nitrogen, phosphorus, and potassium ratios, and a delay in shoot growth can result from their use [8,9]. The authors of [10] studied microbial activities involved in organic C, N, P, and S availability and the release of mineral forms in different growing media made of organic fertilizer combinations. Specific responses were observed, showing the complexity of the mineralization process. In particular, the mineralization rate varied greatly from one growing media to another.

Mineral nitrogen is the nutrient most used by plants, and it is often the most limiting element for plant growth [11]. The N preference of plants is variable and closely related to environmental conditions. For example, the N preference of plants changes from NO3 − to NH4 <sup>+</sup> from drier to wetter sites [12], while the preference shifts from NH4 <sup>+</sup> under an acidic environment to NO3 − at alkaline locations [13]. Moreover, NO3 − is more accessible to plants because it reaches the roots by mass flow, whereas ammonium reaches them by diffusion [14]. Nitrogen deficiency results in symptoms such as reduced growth, and yellowing of leaves occurs very fast after the onset of deficiency [15]. Conversely, nutrient excess due to too rapid mineralization of organic fertilizer or the presence of unwanted ions such as SO4 <sup>2</sup>−, sodium, or chlorine at high concentrations can result in salinization of the organic growing media [16]. It is difficult for growers to match the availability of dissolved growing media nutrients with plant demands at different stages of their developmental cycle [17]. Because a large proportion of organic nutrients are mineralized within the first few weeks and can leach out of containers, substrates containing organic fertilizers are typically used as the sole fertilizer source only for short-term crops. For long-term crops, substrates containing fertilizers are typically not enough to supply plant needs throughout the crop cycle and must be supplemented by top-dressing, side-dressing, or the use of liquid fertilizers in irrigation water [1].

The microbe-mediated N mineralization rate and the microbial community composition are highly variable and dependent on several factors such as growing media temperature, air porosity, and moisture content, as well as on the nature of the organic fertilizer source and the growing media composition (particle size and composition) [17–19]. The C:N ratio of growing media can also impact organic fertilizer mineralization. Substrates with C:N ratios exceeding 30:1 tend to immobilize N due to microbial decomposition of available C, which requires N [20]. Wood components such as composted barks, hammer milled, wood materials, and sawdust can have C:N ratios of 300:1 or more, and they have a high potential to immobilize N from applied fertilizers. Non-wood components with high C:N ratios, such as coir fiber, can also immobilize N [21].

N mineralization and nitrification have been thoroughly studied in soils, but knowledge gaps persist with regard to growing media; as they present low biodegradability, one can wonder whether indigenous microbial communities are suitable for organic N mineralization and nitrification, depending on temperature and moisture conditions. More generally, the transposition of mineralization and nitrification knowledge from soils to growing media is questioned. Thus, understanding the drivers of organic fertilizer mineralization and nitrification in horticultural growing media is necessary for a better prediction of mineral N availability for plants. The objectives of this study were to characterize the dynamics of NH4 <sup>+</sup> and NO3 − production, and to evaluate the impacts of temperature and humidity on the N dynamics in the growing media. The ambition was to set up temperature and moisture functions to be ultimately used for modeling the N dynamics

in fertilized growing media under fluctuating conditions of moisture and temperature, as met by producers. A laboratory incubation experiment was conducted to characterize the organic N fertilizer mineralization of four commercial growing media under different growing media moisture and temperature regimes. We hypothesized that the action laws for temperature and moisture established for soils would be applicable to the growing media, whatever the growing media.

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

#### *2.1. Materials*

Four marketed growing media (GM1 to GM4) were studied. They were selected to be representative of the mostly used growing media (frequency and commercial volume, growing media producers survey, results not shown). Their properties are presented in Table 1. GM1 was made of black peat and composted plant fibers (80–20 vol.%), GM2 contained blond and black peat, coconut fiber, and composted plant fibers (50–20–10–20 vol.%), GM3 contained blond peat, coconut fiber, and composted bark (70–20–10 vol.%), and GM4 contained blond and black peat, coconut fiber, and green waste compost (60–10–20–10 vol.%). The properties of GM3 were somewhat different from those of the other growing media, with a coarser particle size, a higher OM content, a higher C:N ratio, and a lower bulk density. Three fertilizer modalities were studied: no fertilization (F0), organic fertilizer of an animal-based origin (F1), and organic fertilizer of a plant-based origin (F2). The fertilizer compositions are presented in Table 2, but commercial names were kept confidential.

**Table 1.** Physicochemical properties of the growing media.


<sup>1</sup> Bulk density (g·cm−3) was determined following [22], and <sup>2</sup> total porosity (*v/v*) was determined following [22], <sup>3</sup> EAW: easy available water (%, *v/v*) and <sup>4</sup> AFP air-filled porosity (%, *v/v*) were calculated from water retention curves determined using a tension table draining at pressure potentials ranging from −1 to −10 kPa [22]; <sup>5</sup> pH was determined following [23]; <sup>6</sup> EC: electrical conductivity was determined following [24]; <sup>7</sup> OM (organic matter; % dry mass) was determined by loss of ignition (550 C, 7 h); <sup>8</sup> total N was determined by dry ignition according to [25]; <sup>9</sup> basal respiration was obtained by the MicrorespTM method with growing media maintained at 60% of the water holding capacity at 25 ◦C for 1 week.

#### *2.2. Experimental Design*

The four growing media were incubated at four temperatures (4, 20, 28, and 40 ◦C) and three matric suctions (humidity maintained at −3.2, −10.0, and −31.6 kPa corresponding to pF1.5, pF2.0, and pF2.5, respectively), with or without added fertilizer, in the dark, in the absence of plants. Destructive samples were used to measure NH4 +-N and NO3 −-N contents after 3, 7, 14, 28, and 49 days. They consisted of 90 mL vials filled with growing media depending on its bulk density (Table 1), i.e., 16, 18, 11, and 17 g dw per vial for GM1, GM2, GM3, and GM4, respectively. Vials were destroyed at each date for measurements. They were placed in trays with a lid on each vial, and the growing media water content was maintained by weighing control. The organic fertilizer was applied at a rate of 55 g N·kg−<sup>1</sup> growing media dw. Caps were placed on the vials, unsealed to permit air circulation but limit fast water evaporation. The amount of applied fertilizer was calculated on the basis of usual producers' practices (200 g fertilizer N·m<sup>−</sup>3). The growing media were analyzed for their NH4 +-N and NO3 −-N contents at the beginning and after the start of the experiment. Three replicates were prepared per modality.

**Table 2.** Fertilizer composition.


OM was determined following [26]; inorganic N was extracted with 1 M KCl (1:5 *v/v* ratio). Ammonium and nitrate concentrations were determined by colorimetry using a continuous flow analyzer (Skalar Analytical). Total P, K, Mg, and Mn were extracted following [27] and measured following [28]; pH was measured following [23].

#### *2.3. Analysis*

NH4 +-N and NO3 −-N were extracted with deionized water (1:1.5 vol.) for 1 h. Concentrations were determined by colorimetry using a continuous flow analyzer (Skalar Analytical). To obtain a growing media water suction curve, samples were saturated with distilled water for 48 h, with three replicates per growing media. Then, they were gradually dried using sand suction tables [29] with potentials equivalent to 0, −3.2, and −10 kPa. A ceramic pressure press was used for suction equivalent to −31.6 kPa [30,31]. When equilibrium was reached (2–3 days), the samples were dried in an oven at 105 ◦C for 48 h and weighed.

Nitrifying bacteria were quantified as follows: total nucleic acids were extracted from growing media samples using a Qiagen DNeasy PowerSoil kit (Cat No./ID: 12888-100). Then, quantitative polymerase chain reaction was performed using primers 968 R and 1401 R [32] for total bacteria and primers amoA-1F and amoA-2R [33] for nitrifying bacteria.

#### *2.4. Data Treatments*

#### 2.4.1. Ammonium and Nitrate Concentrations

N mineralization was estimated by measuring NH4 +-N and NO3 −-N concentrations, and their sum was used as the total mineral N concentration at each sampling time. Since growing media already contained mineral N (Table 1), NH4 +-N and NO3 −-N concentrations were expressed by subtracting each respective initial mineral N content. Nitrate was the final product of N mineralization we monitored; hence, we expressed NO3 − production as the relative proportion of total mineral N (i.e., NO3 − + NH4 +) released from fertilizer degradation at each timepoint.

#### 2.4.2. Abiotic Factors

The temperature action law was determined in two steps, as described below.

First, it corresponded to a ratio of the mineralization rate at a given temperature over the mineralization rate at a reference temperature. Second, this ratio was plotted against growing media temperature, and modeled using the STICS crop model temperature action law dedicated to the simulation of organic matter decomposition [34].

$$f(T) = \frac{\left( [N]\_{i,fin} - [N]\_{i,init} \right)}{\left( [N]\_{ref,fin} - [N]\_{ref,init} \right)} = \frac{B}{1 + \mathbb{C} \times \exp(-k \times T)},\tag{1}$$

where *T* is the temperature (◦C), [*N*]*i,fin* and [*N*]*i,init* are the total mineral N on days 49 and 0 at a given temperature *Ti*, respectively, and [*N*]*ref,fin* and [*N*]*ref,init* are the total mineral N on days 49 and 0 at the reference temperature *Tref* = 20 ◦C, respectively. This reference temperature was made to be close to that of the STICS model (i.e., 15 ◦C). B was a dimensionless adjusted parameter, and k was an adjusted parameter (◦C<sup>−</sup>1). C was a parameter and was recovered by solving the following equation:

$$\mathbb{C} = (B - 1) \times \exp\left(k \times T\_{ref}\right), \quad \mathbb{C} = (B - 1) \times \exp\left(k \times T\_{ref}\right). \tag{2}$$

The water content action law was also determined in two steps, as described below.

First, it corresponded to a ratio of the mineralization rate at a given growing media matric suction over the mineralization rate at a reference matric suction. Second, this ratio was plotted against growing media moisture, and modeled using the STICS crop model moisture action law dedicated to the simulation of organic matter decomposition [34].

$$f(H) = \frac{\left( [N]\_{i,fin} - [N]\_{i,init} \right)}{\left( [N]\_{ref,fin} - [N]\_{ref,init} \right)} = \frac{H - H\_{wp} \times H\_{fc}}{\left( H\_{fc} - H\_{wp} \right) \times H\_{fc}} \tag{3}$$

where *H* is the volumetric water content of the growing medium (*v/v*), *[N]i,fin* and *[N]i,init* are the total mineral N on days 49 and 0 at a given water content H*i*, respectively, and *[N]ref,fin* and *[N]ref,init* are the total mineral N on days 49 and 0 at the reference water content corresponding to water suction −10 kPa, respectively. *Hwp* is the volumetric water content at the wilting point (water suction −100 kPa), and *Hfc* is the volumetric water content at field capacity (water suction −1 kPa).

To further understand the temperature and moisture interactions, the ratio of NO3 − to total mineral N was calculated as a mean ratio for the whole incubation period. This allowed us to identify abiotic conditions that may have slowed down or favored the nitrification process.

#### *2.5. Statistical Analyses*

We used three-way repeated-measures ANOVA (rmANOVA) to test the interaction of the growing medium type, the fertilizer type (Fert), temperature (Temp), and matric water suction (ψ) on NO3 −, NH4 +, and NH4 <sup>+</sup> + NO3 − concentrations following fertilization. We analyzed these effects separately depending on the significant interactions. We present the results for each growing medium, comparing temperatures or matric water suctions, and only with or without addition of fertilizers 1 or 2 to simplify the viewing of these effects. Significant differences were tested by the least significance difference test (LSD, *p* < 0.05). Correlations were tested using Pearson correlations (*p* < 0.05). When data seemed to present segmented regressions, we tested piecewise regressions with SegReg free software.

#### **3. Results**

#### *3.1. Dynamics of Growing Media N Mineral Content*

We did not find a significant four-level interaction among growing medium, fertilizer type (Fert), temperature (Temp), and humidity (Hum) over time as hypothesized (Table S1, four-way repeated-measures ANOVA). Instead, we did find significant threeway interactions such as GM × Temp × Hum. This interaction was the most powerful one (F > 3.9, *p* < 0.001, within effect) and was also confirmed independently of time

(GM × Temp × Hum: F > 1.7, *p* < 0.001, between effect). As a result, we focused on these interaction factors to present our results (i.e., without considering the fertilizer type, even though we detected some minor modularity of the results per fertilizer type compared to GM, Temp and Hum).

Temperature significantly controlled the total mineral N content (Figure 1), as well as NH4 <sup>+</sup> and NO3 − contents (Figures S1 and S2, respectively), in the four growing media. The pattern of total mineral N differed depending on the growing media. We generally observed a plateau between 28 and 49 days modulated by temperature, except GM1 and GM2 that displayed a linear increase in mineral N content at 40 ◦C. GM1 and GM2 presented the best mineral N content at 40 ◦C after 49 days of incubation (877 and 807 mg N·kg−<sup>1</sup> dw growing media, respectively, Tables S2 and S3). The temperature increase from 4 to 20 ◦C was always significant, whereas 20 ◦C and 28 ◦C tended to have similar effects on GM1 and GM2. For GM3 and GM4, 28 ◦C gave the best mineral N content, whereas 40 ◦C gave a lower content. In the absence of organic fertilization, the rates were close to zero except for GM1 and GM2, where total mineral N significantly increased at 40 ◦C; moreover, a negative N content was found in GM2 and GM4, especially at 28 ◦C, corresponding to organization of the initial mineral N content. At 28 ◦C, GM3 contained the highest mineral N content reached in these incubations (1053 mg N·kg−<sup>1</sup> dw growing media, Table S4).

**Figure 1.** Influence of temperature on net N mineralization at −10 kPa water matric suction, for (**A**) GM1, (**B**) GM2, (**C**) GM3, and (**D**) GM4, with fertilizer F2 or without fertilizer (F0).

Humidity significantly controlled the mineral N content (Table S1, Figure 2), as well as NH4 <sup>+</sup> and NO3 − contents (Figures S3 and S4, respectively), in the four growing media, but less markedly so than temperature. We observed similar trends, with a decreasing mineral N content between 28 and 49 days of incubation. GM1 showed the best mineral N content at −3.2 kPa (the highest humidity rate) and the lowest one at −31.6 kPa (the lowest humidity rate), whereas GM2 presented its best mineral N content at −31.6 kPa. GM3 showed very contrasted mineral N dynamics from 0 to 28 days, and finally reached the same level of mineral N content after 49 days of incubation whatever the humidity level (Figure 2). GM4 showed the highest contrasts between humidity levels, with −31.6 kPa giving the highest mineral N content and −3.2 kPa giving the lowest one (712 mg N·kg−<sup>1</sup> dw GM

and 378 mg N·kg−<sup>1</sup> dw growing media, respectively, Figure 2; Table S5). In the absence of fertilization, GM1 provided mineral N at −3.2 kPa (151 mg N·kg−<sup>1</sup> dw growing media), while GM2 provided a similar content at −31.6 kPa, indicating that humidity did not drive the mineral N content in the same way as in GM1. In GM4, we observed a strong organization of the mineral N content, with no significant effect of humidity (Figure 2, Table S5).

**Figure 2.** Influence of water matric suction on net N mineralization at 20 ◦C, for (**A**) GM1, (**B**) GM2, (**C**) GM3, and (**D**) GM4, with fertilizer F2 or without fertilizer (F0).

#### *3.2. Fertilizer Mineralization*

An increase in temperature from 4 ◦C induced an increase in the percentage of fertilizer N mineralized (Table 3), with slightly contrasting results depending on the growing medium, the fertilizer type, and humidity leading to 20, 28, or 40 ◦C with the highest percentage mineralization of the applied fertilizer N. Whatever the factor, F2 mineralized faster than F1 (10.7–69.2% vs. 14.7–71.3%). GM1 and GM4 reached the highest percentage of fertilizer mineralization at −31.6 kPa, but at different temperatures (40 ◦C and 20 ◦C, respectively). GM2 and GM3 reached the highest percentage of fertilizer mineralization at −10 kPa and 28 ◦C.


**Table 3.** Total mineralized fertilizer F1 and F2 as a percentage of applied N on day 49 (i.e., at the end of the experiment) (*n* = 3, SD = standard deviation).

#### *3.3. Relative Proportion of NO3* − *to Total Mineral N*

We observed four patterns for the relative proportion of nitrate to total mineral N depending on the GM type, temperature, and humidity (Figure 3). GM1 was affected by a decrease in humidity (i.e., a suction decrease), with a weak influence of temperature. GM2 was affected mostly at 40 ◦C and in the driest and wettest conditions (−3.2 and −31.6 kPa, respectively). GM3 was affected by temperature but not by humidity; the ratio was almost the same for all temperatures. The very low values of the ratio revealed that NH4 <sup>+</sup> accumulated substantially in this growing medium type, especially at 4 ◦C. GM4 was the least affected growing medium, with a slow but linear decrease in the values of the ratio as humidity decreased. These decreases were constant, but more pronounced at 20 and 28 ◦C than at 40 and 4 ◦C.

#### *3.4. Temperature and Humidity Action Laws*

A temperature action law was established for all growing media, all humidity levels, and by combining fertilizers F1 and F2 (Figure 4A). The model (Equation (1)) fitted the observed data well. It was calibrated for all humidity levels taken together. Table 4 presents the calibrated parameters and statistical performances (RMSE, *R*2). The lowest RMSE and the best *<sup>R</sup>*<sup>2</sup> corresponded to the model adjustment with humidity at −10 kPa (Figure 4B).

The humidity action law f(H) is presented in Figure 5, combining fertilizers F1 and F2. Different patterns were observed depending on the growing medium, and they also changed according to temperature. At 28 ◦C, f(H) presented less variation for all growing media, with values around 1 in most cases (Figure 5B). This was almost the same at 40 ◦C, except in GM4 (Figure 5C). However, strong variations of f(H) were observed at 4 ◦C, except in GM1 where it was around 1 whatever the H-to-Hcc ratio (Figure 5A).

No correlation was found between the amount of mineralized N and the humidity level whatever the growing medium, but a relationship was established between the ratio of NO3 —N to total mineral N and the humidity content H (Figure 6). The ratio first increased with increasing H, whatever the temperature and considering all growing media, with a breakpoint of the slope when H reached 0.46 *v/v*, and a plateau thereafter. The segmented regression gave a very good correlation coefficient (*R*<sup>2</sup> = 0.83, *p* < 0.001).

**Figure 3.** Relative proportion of nitrate to total mineral N (ratio: NO3 −/total min N) depending on humidity and temperature in GM1 (**A**), GM2 (**B**), GM3 (**C**), and GM4 (**D**) fertilized with F2. Bars represent standard deviations (*n* = 3).

**Figure 4.** (**A**) Temperature response during N mineralization for all growing media incubated at −3.2, −10, and −31.6 kPa; (**B**) temperature response of all growing media incubated at −10 kPa. Data are the means of six replicates (i.e., three each with fertilizers F1 and F2). The temperature action law f(T) was calculated according to Equation (1).


**Table 4.** Adjusted parameters B and k and statistical performance of the temperature action law model.

**Figure 5.** Water content action law at (**A**) 4 ◦C, (**B**) 28 ◦C, and (**C**) 40 ◦C for the four growing media. Data are the means of six replicates.

**Figure 6.** Effect of growing media volumetric water content on the average ratio of NO3 − to total mineral N. The dataset compiles all temperatures.

#### **4. Discussion**

In soils, it has been widely demonstrated that N mineralization is particularly dependent on temperature [35], humidity [36], and texture [37,38]. Growing media are made of organically stable compounds that strongly limit biological activity in the absence of fertilization [10]. Compared to soils, growing media do not provide available nutrients, especially nitrogen, and they usually require mineral or organic fertilization for biological activation of microbes and plant growth. Studies on growing media are mostly focused on physical properties such as hydrodynamic parameters and aeration, or they lay the emphasis on chemical properties such as water pH, electrical conductivity, or cationic exchange capacity [39–42]. These properties are very important and easily monitored, but they weakly reflect the biological aspects and nutrient availability needed for organically fertilized crops. Growers attempt to maintain these properties steadily throughout cropping, but alteration of growing media compounds and root growth can modify them [43,44]. This study was focused on the monitoring of the N mineralization process—which depends on many bio-physicochemical interactions—in four commercial organic growing media with similar physicochemical properties (Table 1). We tested different temperature and moisture levels and compared the responses of the four growing media types when added with two different organic fertilizers and when no fertilizer was added.

The four growing media were relatively similar in terms of physicochemical properties even if they were made from different materials and for different cropping purposes. We rather focused the discussion on the biochemical and microbial aspects addressed in the body of the manuscript.

#### *4.1. Growing Media Type*

The organic growing media were biochemically stable and characterized by high C-to-nutrient ratios, much higher than the C:N:P stoichiometry of microbial biomass [45], which ranges between 42:6:1 and 60:7:1 according to [46] and [47], respectively. Due to their homeostatic stoichiometry, microbes are extremely constrained by the low resource availability in the growing medium; the growing media are considered as biologically inactive, such that microbes strongly respond to organic fertilization (Figure 1) [10]. Consequently, N immobilization can occur in growing media with a C:N ratio exceeding 30:1 due to microbial decomposition of available C—a process requiring N [48]. In the present study, the C:N ratios were similar: around 30 for three growing media types (GM1, GM2, and GM4), and twice higher in GM3 (Table 1). Thus, high N immobilization (i.e., a greater microbial demand) was expected in the four growing media, with the strongest effects in GM3. However, the time course of N mineralization did not confirm these expectations whatever the growing media or the temperature and humidity conditions when fertilizer was added. Even so, GM3 presented the best performances in terms of mineral N release (Figure 1C). Net N immobilization (i.e., negative net N mineralization rates due to gross N organization higher than gross N mineralization) occurred and was only observed in the absence of fertilization and in all growing media; it was the highest in GM2 and GM4, not in GM3 as expected (Figure 1). Net N immobilization was only observed in growing media that already contained mineral N and plant-based compost (i.e., GM2 and GM4). Consequently, its intensity appeared to be limited by the very low initial mineral N content of GM3. Significant N mineralization occurred at 40 ◦C in the absence of fertilization. However, we failed to distinguish whether it resulted from the activation of microbes mineralizing growing media compounds or from dead cells, since this temperature can be critical for some microbial populations such as nitrifiers [49] (Supplementary Materials Table S6). These results could be confirmed by decreased N mineralization rates at this temperature in the fertilized treatments, but further investigations are required. N immobilization occurred in all four commercial growing media, especially because they received a compost fraction during formulation that produced mineral nitrogen before they were used. However, this immobilization effect was easily outperformed by organic fertilization. On the contrary, the raw materials showed no such immobilization effect and even no biological activity or very weak activity in the absence of organic fertilization [10].

#### *4.2. Fertilizer Type*

F1 and F2 were commercial animal-based and plant-based fertilizers, respectively. Due to confidentiality rules, the fertilizer compositions were unavailable, but we analyzed them for pH, elemental composition, and C:N:P stoichiometry (Table 2). We detected significant differences between the two organic fertilizers (Table S1). However, the patterns obtained with either fertilizer were very close, and we only showed curves of unfertilized versus F2-fertilized growing media to clarify illustrations (Figures 1 and 2). We previously showed that the huge C:N and C:P ratios of different growing media were the most important drivers of organic fertilizer mineralization and microbial activity, and they constrained the autochthonous microbial communities through C and nutrient availability [10]. The results of the current study confirm these effects. The C:N ratios of the growing media indeed largely overcame (Table 1) the different C:N ratios of the two fertilizers, and this explained the small difference in their N mineralization response. The C:N ratio of an organic fertilizer is usually considered as a good predictor of the N mineralization or immobilization balance following fertilizer incorporation in soil [50,51]. However, in growing media where nutrient availability is low, fertilizers have a smaller impact than growing media on N mineralization, and this raises the question of the importance of growing media formulation. Granular fertilizers present similar patterns of nutrient release, but other organic fertilizers such as raw or thermally treated horn can have a huge impact on the control of nitrogen release, while remaining driven by the unbalanced stoichiometry of the growing media types [10]. Since C cannot be decoupled from the N and P cycles [10], we hypothesized that the lower P content in F2 (Table 2) may also have increased the microbial mining effect [52] to access P compared to F1, leading to overall faster fertilizer mineralization. Further investigations on coupling microbial C, N, and P functions [10] would be necessary to confirm this assumption.

We also expressed total mineral N at the end of incubation as a percentage of N fertilizer addition and detailed the patterns according to temperature and humidity, as a function of the two fertilizers (F1 and F2, Table 3). However, these values were not cumulative because we only measured the mineral N content at different timepoints of incubation. Thus, at the end of the experiment, the mineral N content did not express the total mineral N produced from the organic fertilizers, but the total mineral N content as a fraction of fertilizer-added N. Only a minimum value of what was really mineralized from the organic fertilizers was expressed. We cannot rule out that some of the mineral N came from (i) microbial turnover [53], especially at 40 ◦C, which can be critical for some microbial populations, and (ii) growing media biodegradation. However, growing media biodegradation is believed to be very low in growing media and would necessitate specific C inputs to trigger a priming effect [54].

#### *4.3. Temperature Effects and Action Law*

Temperature influences transformation rates through the responses of microorganisms and enzymatic activities. We tested four temperatures frequently met during plant growth in horticulture. Specifically, 20 and 28 ◦C are classical temperatures in the greenhouse and supposed to be optimal, whereas 4 and 40 ◦C are extreme temperatures affecting nutrient availability by slowing down microbial activity; 40 ◦C potentially affects the microbes themselves. We observed maximum nitrification at 28 ◦C, close to the soil optimum of 30 ◦C [55]. However, 40 ◦C sometimes gave the best mineral N content depending on the growing media, suggesting that this temperature provided for the highest mineralization rates. Delving deeper into the ammonification and nitrification processes (Figure S2), this mineralization was not sustainable since the NH4 <sup>+</sup> content was higher (Figure S1) than at the other temperatures while the NO3 − content decreased, indicating that nitrifiers were probably affected [55,56]. This unbalance between ammonification and nitrification was also analyzed by studying the relative proportion of NO3 − content over total mineral N (Figure 3): 40 ◦C and mostly 4 ◦C consistently resulted in bad conditions, and even critical ones for GM3.

N mineralization increases exponentially within the range of temperatures met in farmed soils (0–40 ◦C) and can be successfully modeled with numerous functions [57]. This study showed that the formalism of the temperature action law proposed by the STICS model for soils is also adapted to organic growing media mineralization. This is a first modeling of the effect of temperature on N mineralization rates applied to growing media. Modeling performance was best when using incubations were run at −10 kPa, as this modality was most adequate to reveal fertilizer N mineralization. A common way to express temperature sensitivity is to use the Q10 function. A Q10 of 2, for example, means that the rate of a particular process doubles when the temperature increases by 10 ◦C [58]. Using 20 ◦C as the reference temperature, the Q10 value of the growing media

was 1.13, lower than the average Q10 value of arable soils, but within the large range of values reported in the literature (from 0.55 to 11.9 [59]).

#### *4.4. Humidity Effect and Action Law*

Matric suction (ψ) was used to study the effect of humidity (H). However, due to their composition, the growing media had a specific H at a given ψ value, and this made it more difficult to analyze the results (Table 5). H might have been a better parameter choice in the experimental design, even though ψ affected organic nitrogen mineralization statistically. The choice of ψ made agronomic sense in terms of water-filled porosity; a ψ of −1 kPa is equivalent to the retention capacity of a growing media (H*fc*), a ψ of −10 kPa corresponds to the temporary wilting point, and a ψ of −100 kPa corresponds to the permanent wilting point [60].

**Table 5.** Volumetric water content (*v/v*) and water-filled pore space (WFPS, %) values according to the growing media and ψ modalities.


Water is necessary for microbial activity, and its content has to be balanced with the oxygen required for root and microbial respiration [61]. Aerobic microbial activity is optimal at a humidity volumetric content ranging between 50% and 70% of the water holding capacity (WHC) [62,63] corresponding to water and oxygen availability in good equilibrium. Other studies estimate the maximum microbial activity (respiration and nitrification) in soils around 60% of the total porosity occupied by water (WFPS, "water-filled pore space") [64,65]. A major influence of the water content has been shown on microorganism activity in different organic growing media, with higher microbial respiration at 63% WFPS compared to 73% and 83% WFPS [66]. According to the theory mentioned above, fertilizer mineralization should be optimal around −3.2 kPa, as in GM1 and GM3. Yet, the mineralization rates of GM2 and GM4 were highest at −31.6 kPa, i.e., the driest conditions of this study. As a result, oxygen could be more important than water availability. Suction of 31.6 kPa is supposed to be too low for plant survival and growth. The ratio of NO3 − to (NH4 <sup>+</sup> + NO3 −) as a function of growing media humidity showed a similar trend to that observed for soils, regardless of temperature [67], i.e., a progressive increase with humidity up to 0.46 *v/v* (corresponding to a WFPS of 50–52% depending on the growing medium), followed by a plateau up to 0.67 *v/v* (corresponding to a WFPS of 73–76% depending on the growing media) (Figure 6). Thus, the optimal humidity for nitrification in the growing media was similar to that of soils. However, we failed to establish a humidity action law on the basis of experimental data. Indeed, at a given temperature each growing media had a specific response curve to H/H*fc* (Figure 5); when H/H*fc* increased, f(H) decreased for some growing media and then increased with increasing H/H*fc* values. For other growing media, f(H) progressively increased, or even reached a plateau. At another temperature, the growing media behaved again in a different way, making it difficult to establish a generic humidity action law.

Thus, an interaction between growing media temperature and moisture did occur, and this process is commonly encountered in soils [68,69]. Several humidity action laws have been established for arable soils and expressed as a function of the soil water content, the WFPS, or the matric potential. The need to improve the representation of this relationship in models has been highlighted. The authors of [70] presented a data-driven analysis of soil humidity–respiration relations based on 90 soils. They showed how the relationship between soil heterotrophic respiration and different soil humidity levels is consistently affected by soil properties. On the basis of the proportional response of soil respiration (PRSR) related to a 0.01 increase in soil humidity as the central unit for analysis, they found little or no effect of soil properties on the PRSR in organic soils (i.e., a soil organic content higher than 300 g·kg−1). Thus, due to their nature and their specific behavior, the formalisms known today are not adapted to growing media; research work has to be developed in this area.

#### *4.5. Management Implications*

In terms of professional applications, growing media and fertilizer type need to be considered at the same time to determine the rate of N release adequate for plant growth as precisely as possible. Depending on plant requirements, professionals could select a growing media according to its use (but physical characteristics would need to be checked) and use an organic fertilizer to provide nutrients. Our results tend to show that GM4 supplied slow nutrient release that quickly reached a plateau at the lowest level in this study, whereas GM3 supplied slowly the highest nutrient level. These results also indicate that detecting the plateau would have required a longer incubation time than 49 days for GM3 (Figure 1). Moreover, all four growing media presented a linear increase in total mineral N at 40 ◦C indicating that extreme temperature can cause fast N release with potential N loss if not synchronized with plant needs, whereas the process would be best controlled at a temperature maintained between 20 and 28 ◦C. Peat is a reference material in soilless production. Nevertheless, its use is questioned because exploiting peatland implies depleting a recognized carbon (C) sink. Thus, efforts and the immediate need for peat reduction in horticulture are a strong challenge for the future. Growing media tested in this study were peat-based, but combined with other alternative materials. We were able to show here that growing media with only 50% or 60% peat (GM2 and GM4, respectively) mineralized as much fertilizer as GM1 (80% of peat). The results are, therefore, encouraging and demonstrate that it is possible to progressively free ourselves from peat.

Mineralized F2 induced a high N release and appeared to be interesting for short- or medium-term crop cultivation, while the fertilizer dose could be defined accordingly. The temperature and ψ conditions that promoted the highest N fertilizer mineralization were 28 ◦C and −10 kPa (for GM2 and GM3) or −31.6 kPa (for GM1 and GM4). To go further, growing media–fertilizer combinations should be tested in actual growth experiments since the plant uptake could reveal inadequate growing media–fertilizer combinations or too limiting ones in terms of nutrient supply, as suspected with GM4. Indeed, synchronizing nutrient supply with the nutrient requirements of plants is a major issue for increasing nutrient use efficiency. While nitrogen is essential and often the most limiting element for plant growth, plants can be subjected to multiple nutrient limitations, especially colimitation of N and P [71]. A depressive effect of organic versus mineral fertilization is frequently observed [72,73]. For example, higher ammonification over nitrification rates is the main explanation for the lower performances of organically grown basil plant because roots are exposed to high levels of NH4 <sup>+</sup> without supplying enough NO3 − [74,75].

#### **5. Conclusions**

The N mineralization dynamics of two organic fertilizers in four growing media types at different temperature and humidity conditions showed a strong impact of the different treatments on NH4 <sup>+</sup> and NO3 − release. Under optimal conditions of temperature (20 ◦C) and humidity (−10 kPa), 32% to 57% of the applied fertilizer was mineralized after 49 days depending on the growing media. These results constitute major food for thought on fertilizer application strategies during crop itineraries. The introduction of plants in the system will have an impact on the mineralization process, which we plan to study in the future. We attempted to adapt temperature and humidity action laws, whose formalisms are derived from work on soils, to growing media. We succeeded in describing the effect of temperature with an action law common to the four growing media, but the response of

the growing media to humidity greatly varied among growing media and in a temperaturedependent manner. Therefore, we failed to establish an action law for humidity, although a satisfactory relationship between nitrification and humidity was demonstrated. Research is needed to further investigate the effect of humidity and temperature-humidity interactions on the mineralization of organic N from fertilizers. In addition, the present results need to be refined using other growing media–fertilizer pairs. This work will allow for the shortterm development of a prediction model of mineralization of organic N from fertilizers in soilless growing media production because such a model is lacking at present.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10.339 0/horticulturae8020152/s1: Figure S1. Influence of temperature on NH4 +-N at <sup>−</sup>10 kPa water matric suction, for (A) GM1, (B) GM2, (C) GM3, and (D) GM4; Figure S2. Influence of temperature on NO3 −- N content at −10 kPa water matric suction, for (A) GM1, (B) GM2, (C) GM3, and (D) GM4; Figure S3. Influence of water matric suction on NH4 +-N content at 20 ◦C, for (A) GM1, (B) GM2, (C) GM3, and (D) GM4; Figure S4. Influence of water matric suction on NO3 −-N content at 20 ◦C, for (A) GM1, (B) GM2, (C) GM3, and (D) GM4; Figure S5. NO3 − to total N min ratio, depending on matric suction and temperature and with fertilizer F1, in (A) GM1, (B) GM2, (C) GM3, and (D) GM4. Bars represent standard deviation (*n* = 3); Table S1. Results of three-way repeated-measures ANOVA with growing media, fertilizer (Fert), temperature (Temp), and matric water suction (ψ) as between subject and time (t) after fertilizer addition as within subject, on total mineralized N, NH4 +, and NO3 − contents; Table S2. NO3 −, NH4 +, and total mineral N in GM1, depending on ψ, temperature, and fertilizers modalities (*n* = 3, F0 = without fertilizer); Table S3. NO3 −, NH4 +, and total mineral N in GM2, depending on ψ, temperature, and fertilizer modalities (*n* = 3, F0 = without fertilizer); Table S4. NO3 −, NH4 +, and total mineral N in GM3, depending on ψ, temperature, and fertilizer modalities (*n* = 3, F0 = without fertilizer); Table S5. NO3 −, NH4 +, and total mineral N in GM4, depending on ψ, temperature, and fertilizer modalities (*n* = 3, F0 = without fertilizer); Table S6. AmoA content (log nb\_seq·g−<sup>1</sup> dw growing media) at <sup>−</sup>3.2 and −31.6 kPa, and temperatures of 20 and 40 ◦C, during the 49 day incubation (*n* = 3).

**Author Contributions:** Conceptualization, S.B., M.V., P.C., R.G. and M.B.; methodology, S.B., M.V., P.C., R.G. and M.B.; validation, S.B., M.V., P.C., R.G., M.B. and S.R.; formal analysis, P.C., R.G., S.R., M.V. and M.B., investigation, P.C., R.G., S.R., M.V., S.B. and M.B.; writing—original draft preparation, P.C., R.G. and S.R.; writing—review and editing, P.C., R.G., S.R., M.V., L.P., S.B. and M.B.; visualization, P.C., R.G., S.R., M.V., L.P., S.B. and M.B.; supervision, P.C., R.G. and S.R.; project administration, S.B.; funding acquisition, S.B., M.V., P.C., R.G., M.B. and S.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by CASDAR, OptiFaz project, grant number 5746.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data sharing not applicable.

**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.

#### **References**

