Abstract
Understanding the spatial variability of soil organic carbon (SOC) content and its functional pools under current farming practices is crucial for developing targeted C management. This study quantified and predicted the farm-scale variability of SOC pools across conventional tillage (CT) and no-tillage (NT; 8–14 years) practices at two sites (Rostov and Krasnodar) in Russia. The soil types at Rostov and Krasnodar farms were Calcic Chernozem (sunflower–wheat rotation) and Stagnic Chernozem (maize–wheat rotation), respectively. The average SOC content at the Rostov site was higher than the Krasnodar site by 41% and 28% in 0–10 and 10–30 cm, respectively. For both sites, there was no clear trend in SOC variability between NT and CT practices. However, topsoil microbial-available C pool (mineralized for 180 days) was most sensitive to tillage systems, unlike unchanged particle-size C pools. Specifically, it increased from CT to NT at the Rostov site (by 7–16%), but it showed a decreased trend at the Krasnodar site (by 11–29%). Gradient boosting machines statistical models with remote sensing data based explanatory variables (spectral, topography) accurately predicted the spatial distributions of topsoil C content (R2 = 0.99) and its microbial-available pool (R2 = 0.78) across the farmland areas. The main explanatory variables included topography, vegetation distribution, moisture and thermal regimes. For both sites, DNDC modeling showed that applying NT versus CT for 30 years could potentially increase SOC in the topsoil by 27–28% and decrease it in the subsoil by 6–9% (sunflower–wheat–maize–wheat rotation; annual N and P rates of 41–80 and 0–52 kg ha−1). This study provides insights into current agricultural challenges and the developing site-specific strategies for managing soil C accrual in the Chernozem region.
1. Introduction
Chernozems are recognized as the most fertile and agriculturally valuable soils, which globally extend over ~230 million hectares in the mid-latitude grasslands [1]. According to the international soil classification system, Chernozems are grouped with other highly fertile soils (Phaeozems, Luvisols, Ultisol, Kastanozems) as Mollisoils, which have historically been extensively cultivated for crop production, playing an important role in ensuring food security [2]. Since the 1960s, the majority of Chernozems have been affected by accelerated erosion and physical or chemical degradation due to the use of increasingly powerful machinery and potent chemicals by farmers [3]. As a result, tilled Chernozems—like other Mollisoils—have lost about half of their initial organic carbon (C) stocks to date [4]. This has significantly reduced the sustainability of crop production, highlighting the urgent need to adapt modern farming practices to long-term C accrual in the Chernozem region.
The efficiency of agronomic practices to long-term C sequestration and storage in soils can be characterized by the C distribution in fast-cycling particulate organic matter (POM) and slow-cycling mineral-associated organic matter (MAOM) pools [5,6]. The different cycling rates of these functional pools are caused by their distinct interaction strength with the mineral surfaces and Fe-/Al-(hydr)oxides, which physically and chemically protects C against microbial decomposition [7]. Specifically, molecular organic fragments of MAOM (<53 μm) with varying degrees of amphiphilicity can strongly associate with the silt and clay minerals of variable chemical reactivity, unlike larger organic fragments of POM (53–2000 μm) [8,9]. The formation of MAOM requires favorable soil conditions for the microbial transformation of organic matter into simpler forms, especially microbial necromass residues and plant-derived water-soluble monomers [6], which constitute a microbial-available C pool. At the same time, the relationship between the functional C pools and their contribution to the total C variability across different farming practices in the Chernozem region remains poorly investigated [10,11].
Among conservation agriculture practices, no-tillage has been extensively studied for its potential to enhance soil C stocks by minimizing soil disturbance and promoting residue retention [12,13,14]. Despite ongoing debates about its effectiveness, no-tillage is widely recognized as a reliable tool for reducing soil degradation from water/wind erosion and for optimizing water usage [15]. Furthermore, this conservation agriculture practice fosters favorable topsoil microclimatic and nutritional conditions for microorganisms, leading to a steady increase in their population and diversity [16,17], and microbial-available C pool size [18]. These, in turn, create the potential to increase topsoil MAOM and thus sustainably reduce the existing C saturation deficit in agricultural soils. However, the effectiveness of no-tillage in long-term C sequestration is site-specific and influenced by factors such as climate, soil texture, initial MAOM level, cropping system, fertilizer strategies, and duration of the practice [19,20,21,22]. In real farming settings, assessing no-till efficiency is barely impossible because baseline soil C content is typically unavailable. A useful approach to addressing this problem is using biochemical process-based models for agroecosystems. Currently, many validated simulation models are available for evaluating potential C dynamics in agricultural soils under different management scenarios, including tillage practices [23].
Adapting agricultural practices to long-term C sequestration at the farm scale is challenging due to the pronounced spatial heterogeneity in soil conditions. This heterogeneity results from complex interactions between natural factors (e.g., topography, thermal and water regimes, and soil texture) and diversification of cropping management from field to field (e.g., crop rotation, fertilization, and tillage systems). In this regard, studies aimed at predicting and mapping the local-scale variability of soil C content and its functional pools are particularly relevant and valuable. While soil organic C content has often been mapped by digital soil mapping (DSM) approach at a local scale with topography, agricultural management, and soil properties to be often the most important covariates [24], mapping of different C pools with respect to contrasting agronomic practices remain overlooked. Furthermore, the relative importance of SCOPRAN-oriented controls, often used in DSM [25], is case-specific. Therefore, predictive DSM requires developing a case-specified workflow that optimizes the sampling design, remote sensing data, and statistical modeling algorithm.
Given the above background, this study had the following three objectives: (1) assessing farm-scale variability in soil C content and its functional pools across tilled and no-tilled Chernozems in two different environmental settings, (2) evaluating the potential of no-tillage to increase C content in Chernozems using simulation modeling, and (3) developing robust statistical models for predictive mapping of soil C content and its pools across the studied farmland areas.
2. Materials and Methods
2.1. Study Area
The study was conducted at the following two sites in southern European Russia: Rostov (47.5° N, 38.9° E) and Krasnodar (45.0° N, 37.9° E). Both sites are located in the steppe zone, which has a temperate continental climate. The mean annual temperatures in Rostov and Krasnodar are 10.9 °C and 13.1 °C, respectively, and the mean annual precipitation is 579 mm and 561 mm, respectively (data from the closest WMO weather stations, “34720 Taganrog” and “37001 Anapa,” from 1991 to 2020). The intra-annual distributions of temperatures and precipitation are shown in Figure 1.
Figure 1.
Mean monthly climate for (a) Rostov and (b) Krasnodar sites (1991–2020). Lines show monthly mean maximum (TMAX) and minimum (TMIN) temperatures in °C (±1 standard deviation). Bars represent total monthly precipitation (PRCP) in mm.
The sites have distinct landscape positions and Chernozem subtypes. Specifically, the Rostov site is in an automorphic position at an elevation of 68 ± 5 m above sea level (m a.s.l.) and has Calcic Chernozem. The Krasnodar site is in a lowland area at an elevation of 11 ± 3 m a.s.l. and has Stagnic Chernozem. Both soils formed from loess parent materials and have a silt loam texture. However, Calcic Chernozem versus Stagnic Chernozem has a higher carbonate content (5.7–15.7% vs. 0.1–1.0%) and pH value (8.3–8.7 vs. 7.3–8.1) (Table A1).
Crop cultivation at the Rostov site dates back about a century, as evidenced by an online archive from the 1920s. In contrast, crop cultivation at the Krasnodar site began in the late 1960s, when a drainage canal network was constructed, according to interviews with local farmers. Currently, Rostov’s crop rotation included winter wheat and sunflower, and Krasnodar’s one included winter wheat and maize. Historically, both sites have practiced conventional tillage (CT), which involves moldboard plowing to a depth of 22–25 cm. Since the 2010s, some farms changed soil treatment from CT to no-tillage (NT).
In 2023, we selected three adjacent fields at each site: one with CT, and two with NT for 8–14 years (Figure 2). All of the chosen fields were owned by commercial farms. Only mineral fertilizers were used, with application rates varying from year to year depending on the crop grown and the farm’s budget. Generally, the annual rates for nitrogen vary between 0 and 100 kg N ha−1, for phosphorus between 0 and 70 kg P ha−1, and for sulfur between 0 and 14 kg S ha−1, as determined through communication with farmers.
Figure 2.
Location of sampling points across the studied fields at Rostov (left) and Krasnodar (right) sites in southwest Russia.
The crops grown in 2022–2023, the agrochemical soil characteristics and bulk density in 2023 (our lab measurements) for the studied fields are presented in Table 1. Overall, all of the Chernozems were depleted of available phosphorus and were slightly to moderately alkaline. At the same time, Krasnodar’s soil was characterized by rather increased bulk density compared to Rostov’s soil.
Table 1.
Characteristics of studied fields with conventional tillage (CT) and no-tillage (NT) over 8–14 years. Notes: CH, Chernozem; FA, field area; W, winter crop; BD, bulk density; N, total nitrogen; P, labile phosphorus; K, labile potassium; NA, not available. Values for soil properties are means for n = 30.
2.2. Soil Sampling Design
Both sites employed the conditioned Latin hypercube sampling (cLHS) design [26], which accounted for variations in both spectral characteristics (surface reflectance in blue (490 nm), red (665 nm), red-edge (740 nm), near-infrared (NIR) (842 nm), and short-wave-infrared (SWIR) (1610 nm) as well as crop productivity via the normalized difference vegetation index—NDVI [27] and terrain parameters (elevation in m above EGM96 geoid, slope in degrees, aspect in degrees, flow direction, flow accumulation, and topographic wetness index, TWI [28]). The last three terrain parameters are all unitless. We have used a median composite (median value across dates) of cloud-free Sentinel-2 (Level-2A) images acquired within June and September of 2023 for retrieving surface reflectance in blue, red, red-edge, NIR, SWIR ranges of electromagnetic spectrum as well as for NDVI. We have further used the Shuttle Radar Topography Mission (SRTM) digital elevation model with 30 m resolution to calculate terrain parameters [29]. We prepared remote sensing data using GEE (Google Earth Engine) cloud-computing platform. Terrain derivatives were calculated in QGIS Desktop software (version 3.38) and the R environment (version 4.4.1) [30]. Spatial data were prepared as multiband raster file at 30 m spatial resolution by resampling (nearest neighbor) and aligning Sentinel-2 10 and 20 m resolution bands to SRTM extent and shape.
Within each field of every site, 30 locations for sampling have been selected using twelve raster layers in the R package clhs [31]. The location of the sampling points across the studied fields is shown in Figure 2. At each sampling point, a soil sample was taken from the topsoil (0–10 cm) and subsoil (10–30 cm) layers in November 2023. The soil samples were placed in individual plastic bags and delivered to the laboratory.
2.3. Analysis of Soil Organic Carbon Content and Modeling Its Dynamics in Response to Different Tillage Systems
Soil samples were air-dried at the room temperature and sieved through a 2 mm mesh to remove any plant debris and roots. Total carbon (TC) content in a finely ground soil sample (~20 mg) was analyzed using a dry combustion technique with an ECS 8024 NC Soil Special analyzer (Italy). Inorganic carbon (IC) content was measured using the manometric method, which is based on the change in air pressure within a closed vial after adding 10% HCl to the soil sample (2 g). The soil organic C content (SOC) was determined by the difference between the corresponding TC and IC.
The DNDC (DeNitrification-DeComposition) model [32] was employed to project the 30-year changes in SOC under NT and CT management for the study fields. We simulated the established tillage practices for each field at both sites over an extended period under fixed crop rotation and mineral fertilizer rates. The latter two parameters were chosen based on the actual management practices used the farm operations at two sites, which are detailed in Table 2. The sole variable was the tillage method as follows: the CT scenario involved ploughing with a disk or chisel to 10 cm depth followed by moldboard ploughing to 20 cm, while the NT scenario involved surface mulching only with no soil disturbance. The defined crop rotation was then cycled continuously throughout the entire 30-year simulation period.
Table 2.
Cropping system and management parameters used to initialize the DNDC model for simulating 30-year SOC dynamics under conventional tillage (CT) and no-till (NT) scenarios.
Soil parameters for the 0–10 cm topsoil were initialized using a combination of laboratory measurements and established protocols (see Table A2). Soil texture was determined using the USDA textural triangle based on particle size distribution, while core properties including bulk density, pH, clay fraction, and surface SOC content were assigned directly from laboratory analyses. To initialize the model’s C pools, the resistant litter and humads + humus (stable) fractions were derived from the measured proportion of POM to MAOM. The depth to the water retention level was defined for each site as the elevation difference between the soil surface and the nearest open water surface. Finally, the model parameter governing the SOC decrease rate below the topsoil was specifically calibrated for each field using measured SOC data from the 10–30 cm layer, with the depth of topsoil exhibiting uniform SOC content set to 0.1 m.
Meteorological data were initialized using a 30-year daily dataset (1991–2020) of maximum and minimum air temperature and precipitation from the nearest WMO stations: “34720 Taganrog” and “37001 Anapa”.
2.4. Particle-Size Fractionation of Soil Organic Carbon
The SOC of topsoil layer (0–10 cm) was separated into POM and MAOM using a particle size-based technique that combined sonication, wet sieving, and centrifugation. Air-dried and 2 mm sieved soil subsamples (5 g) were dispersed in 50 mL of distilled water with a sonication energy input of 60 J mL−1 (calorimetric calibration). Next, the soil suspension was passed through a 53 µm sieve and washed thoroughly with distilled water. Material retained on the sieve was designated as the POM fraction and dried to constant weight at 60 °C. The <53 μm suspension was centrifuged at 4000 g for 30 min (ULAB UC-1536E centrifuge, “Xieli International Trading Co., Ltd.”, Hong Kong, China) to sediment MAOM-containing particles. The supernatant was discarded, and the pellet was dried to constant weight at 60 °C. The SOC content in the POM and MAOM fractions was determined as for the total soil sample, as described in the previous section.
2.5. Analysis of Soil Microbial-Available Carbon Pool
The potentially microbial-available C pool in the topsoil (0–10 cm) was defined as the amount of SOC that mineralized to CO2 during six months of soil incubation under optimal hydrothermal conditions (22 °C and 65% water-holding capacity). Prior to analysis, 10 g soil subsamples were placed in 100 mL glass vials and standardized by oven-drying at 65 °C for 24 h. This procedure rehydrates microbial cells and incorporates their C into the potentially microbial-available C pool [33,34]. Then, the soil was rewetted with distilled water to 65% of its water-holding capacity. Immediately following, the vials with soil hermetically sealed with silicone stoppers and incubated at 22 °C. The first measurement of the CO2 release rate was performed 3–4 h after closing the vials, followed by daily measurements during the 1st week of incubation, three times a week starting from the 2nd week, twice a week starting from the 2nd month, once a week starting from the 3rd month and up to six months taking into account CO2 flux dynamics. After each measurement, the soil samples in vials were air-exchanged for 15–20 min. Constant soil moisture was maintained by periodically weighing the vials with soil and adding distilled water in case of weight loss. The CO2 concentration in the gas phase of the vial was determined by gas chromatography (Crystallux-4000M chromatograph, LLC Scientific Production Company “Meta-Chrom”, Yoshkar-Ola, Russia). Finally, the total CO2 fluxes released during the six-month incubation period (microbial-available C pool) for each soil sample were calculated. Additionally, the proportion (%) of the microbial-available C pool from the total SOC was calculated.
2.6. Statistical Data Analysis and Field-Scale Mapping
A one-way ANOVA was used to evaluate differences in soil properties among the three fields within each site. If there were significant differences, we used Tukey’s test for multiple comparisons of field means. The spatial variability of SOC pools was quantified by the coefficient of variation (CV, %), which is the ratio of the standard deviation to the mean. Relationships between chemical and physical soil properties were examined with Pearson’s correlation analysis. The analyses were performed in R (version 4.4.1) using the following packages: car for ANOVA [35], DescTools for Dunnett’s test [36], and ggplot2 for data visualization [37].
We have mapped within-field spatial distribution of topsoil (0–10 cm) SOC parameters with predictive digital soil mapping (pDSM). We have used a non-linear ensemble tree-based model GBM (gradient boosting machines) [38] to statistically describe dependencies between target variable and a set of covariates, including spectral and terrain properties. In pDSM, we used a set of thirteen covariates. Among terrain parameters, we have used elevation, slope, aspect, and TWI. Spectral properties at this time have been extracted from Landsat-8/9 (May–July median composite for 2023). The reason of using Landsat data form mapping is the resolution matching SRTM. For each field, all Landsat-based covariates undergo a statistical t-test (sampled pixels versus statistical population). Among spectral properties, we have used the following four spectral bands and three spectral transformation bands: blue, red, NIR, SWIR, NDVI, and MNDWI (Modified Normalized Difference Water Index) [39] as well land surface temperature (LST) calculated as in [40]. Fields (management) and regions were also rasterized and added as covariates to a set of predictors. Uncertainties of GBM models were assessed through a nine-fold cross-validation RMSE. Optimal tuning parameters were established through an iterative looping procedure with different combinations (n = 882). Figure 3 shows the general scheme of the mapping workflow, from developing the sampling design to creating the final maps of the target soil parameters. Processing of remote sensed data has been done in GEE and in R environment (version 4.4.1) [30]. Modeling has been performed using a caret package [40]. We have further assessed feature relative influence of the main covariates using gbm package in R.
Figure 3.
General scheme of sampling design and mapping workflow.
3. Results
3.1. Soil Organic Carbon Content Across Different Chernozems and Field Managements
Across the studied farming fields, the SOC content was significantly higher in Calcic Chernozem than in Stagnic Chernozem (t-test, p < 0.001), averaging 24.1 vs. 17.1 g kg−1 for topsoil (0–10 cm) and 20.9 vs. 16.3 g kg−1 for subsoil (10–30 cm). The coefficients of spatial variability (CV) were similar for both soils: 15.3–14.5% for topsoil and 12.5–16.7% for subsoil.
Regarding contrasting field management, topsoil SOC variability did not show a clear trend between NT and CT practices for both Chernozems (Figure 4). Specifically, there were NT fields with higher (two NT11), lower (NT8), or the same SOC content (NT14) as the reference CT field. At the same time, a common pattern appeared in the subsoil: NT fields had either the same (two NT11) or lower SOC content (NT8 and NT14) compared to the reference CT field.
Figure 4.
Soil organic carbon (SOC) content in 0–10 and 10–30 cm layers of Calcic and Stagnic Chernozems under conventional tillage (CT) and no-tillage (NT). The dotted lines show the mean values; *** p-value ≤ 0.001 for Tukey’s test, n = 30.
DNDC modeling showed that 30-year applying NT practices to both Chernozems could potentially lead to higher topsoil SOC gains (27–28%) than CT practices with the same crop rotation and mineral fertilization (Figure 5). However, the trend is inverse for subsoil, i.e., NT reduces SOC by 6–9% compared to CT.
Figure 5.
DNDC modeling results showing potential dynamics of soil organic carbon (SOC) content in 0–10 and 10–30 cm layers under conventional tillage (CT) and no-tillage (NT) during 30 years for Calcic and Stagnic Chernozems. The solid line shows the average trend, and the dotted line shows the variability during a four-year crop rotation (sunflower–spring wheat–maize–spring wheat sequence).
Thus, for both Chernozems, NT had the potential to enhance soil C sequestration relative to CT, which concentrates in the topsoil layer (0–10 cm). Consequently, we further considered only this soil layer when assessing the qualitative characteristics of SOC under different field management.
3.2. Soil Organic Carbon Quality: POM:MAOM Ratio and Microbial-Available Pool
The POM:MAOM ratio was similar for both soils, ranging from 0.07 to 0.29 for Calcic Chernozem and from 0.05 to 0.25 for Stagnic Chernozem (Figure 6). This indicates that POM contributes only 6–23% to the total SOC, while MAOM contributes 77–93%. Notably, the average POM:MAOM ratio remained mainly unchanged across contrasting field management practices, with the exception of the NT8 field. This suggests that long-term and intensive agricultural use of Chernozems makes this indicator of SOC quality uniform.
Figure 6.
Particulate organic matter (POM) to mineral-associated organic matter (MAOM) ratio in Chernozems’ topsoil (0–10 cm) under conventional tillage (CT) and no-tillage (NT). The dotted lines show the mean values; *** p-value ≤ 0.001 for Tukey’s test, n = 30.
The microbial-available C pool (mineralized SOC for 180 days) in Calcic Chernozem was significantly lower than in Stagnic Chernozem (t-test, p < 0.001), averaging 0.88 mg C g−1 (or 3.7% of SOC) versus 1.45 mg C g−1 (or 7.6% of SOC). Interestingly, this parameter exhibited an inverse trend with the average SOC content in these two soils (Section 3.1).
In terms of contrasting field management, the microbial-available C pool increased from CT to NT for Calcic Chernozem, but it showed an inverse trend for Stagnic Chernozem (Figure 7A). However, the C pool’s contribution to the total SOC exhibited no clear tendency for Calcic Chernozem, in contrast to Stagnic Chernozem (Figure 7B).
Figure 7.
Cumulative soil organic carbon (SOC) mineralization over a six-month incubation period (A) and its percentage from initial soil organic carbon content (B) for Chernozems’ topsoil (0–10 cm) under conventional tillage (CT) and no-tillage (NT). The dotted lines show the mean values; * p-value ≤ 0.05, ** 0.01 and *** 0.001 for Tukey’s test, n = 30.
Thus, in contrast to the uniform POM:MAOM ratio, the microbial-available C pool exhibited a site-specific pattern of change in response to different field management practices.
3.3. Correlations Between Soil Organic Carbon, Agrochemical Properties and Topography
The relationships between soil properties and topography across the studied farming fields was mostly site-specific. In the case of Calcic Chernozem, the SOC content increased with the SOC:N ratio and the labile potassium (K) (Figure 8). At the same time, no correlation was observed between the different C pools, nor between SOC and elevation. For Stagnic Chernozem, the SOC content increased with K content but decreased with both bulk density and elevation. In addition, the microbial-available C pool was positively correlated with the POM:MAOM ratio; in other words, it increased with the POM pool.
Figure 8.
Pearson’ correlation coefficient among soil organic carbon (SOC), its quality (particulate organic matter (POM) to mineral-associated organic matter (MAOM) ratio; Min_SOC, mineralized pool for 180 days), labile potassium (K), labile phosphorus (P), bulk density (BD) in the topsoil (0–10 cm) and elevation (Elev) variability across farming fields with contrasting management practices. Significant coefficients are shown in bold italics at p-value ≤ 0.001 (n = 90 and ‡ n = 60).
3.4. Predictive Modeling and Mapping of Soil Organic Carbon Content and Quality
We selected the SOC content in the topsoil (0–10 cm) and its microbial-available C pool (mineralized SOC over six months) as the target variables for pDSM because they were the most sensitive to contrasting field management. Non-linear statistical models with a set of predictors (Section 2.6) were able to explain 99 and 78% of the variance for above mentioned target variables across six farming fields studied at the two sites (Figure A1). Cross-validation RMSE (RMSEcv) for these models were 2.29 g C kg−1 and 0.29 mg C g−1 of soil, respectively. For SOC content, there was a perfect model fitting, whereas for the microbial-available C pool was a slight underestimation especially for highest values (1.8–2.5 mg C g−1 of soil).
Spatial variability of SOC content has mainly been driven by elevation (relative influence (RI) 28.3%), surface reflectance in near-infrared range (RI 15.1%) and land surface temperature (LST) (RI 7.3%). Impact of different tillage practices has only been visible for no-till field over 11 yrs (RI 6.3%) (Figure 9). Within-field variation (but mainly for the Rostov fields) has further been explained by the vegetation impact as follows: more crop biomass in the pixel—more SOC (statistically determined by increasing near-infrared surface reflectance). Similar to the SOC, the spatial variability of the microbial-available C pool has been driven by elevation and vegetation response (RI: 18.6 and 8.7%, respectively). Additionally, this variable was rather driven by the moisture regime—through MNDWI spectral index (RI of 25.9%)—positive relationship (wetter—higher).
Figure 9.
Spatial distribution of predicted soil organic carbon content (SOC) (on top) and its mineralized pool for 180 days (bottom) for Chernozems’ topsoil (0–10 cm) under conventional tillage and no-tillage across studied farm areas.
Considering the entire studied area, the inter-field distributions of the predicted SOC and its microbial-available pool generally aligned with their measured values (Figure 10). The predicted SOC showed higher spatial variability for Calcic Chernozem than Stagnic Chernozem (ranges from minimum to maximum values: 21.9 versus 14.0 g kg−1 of soil). Notably, this difference was due to the high SOC heterogeneity within the CT field alone.
Figure 10.
Spatial variability of predicted soil organic carbon content (SOC) (on top) and its mineralized pool for 180 days (Min_SOC) (bottom) for Chernozems’ topsoil (0–10 cm) under conventional tillage and no-tillage across studied farm areas.
Unlike the SOC, the predicted microbial-available C pool exhibited higher spatial variability in Stagnic Chernozem than in Calcic Chernozem (value ranges: 1.19 versus 0.71 mg g−1 of soil). Variability within individual fields did not markedly differ among contrasting field management.
4. Discussion
4.1. SOC Content Under Contrasting Tillage Practices in Farmland
Across the studied commercial fields, there was no clear trend in SOC variability between NT and CT practices (Figure 3). This is expected for several reasons, such as the following: (1) the baseline level of SOC content is unknown, (2) crop rotation and fertilizer management are not uniform across fields, and (3) the no-tillage application is short-term. After fixing all of these factors in the DNDC modeling, we found that NT practice has the potential to increase topsoil SOC content relative to CT practice across the studied farmland areas. However, the opposite trend can be observed for the subsoil. These findings collaborate with a literature review evidencing that NT promotes SOC gain in the near-surface soil layers (up to 10–20 cm deep) and SOC depletion in subsoil layers compared to CT [12,13]. Such profile stratification of SOC can be explained by the elimination of the mechanical incorporation of post-harvest plant residues into the soil depth. Additionally, it can be attributed to the specific pattern of root distribution within the soil profile under different tillage practices. Specifically, NT farming increases root biomass at the 0–20 cm depth but reduces it at the 20–30 cm depth compared to plowing, which can be referred to the effect of soil compaction [41]. The limitation of increasing SOC in the entire soil profile under NT system raises some skepticism about its potential to mitigate climate change [12,14]. Consequently, enhancing the climate-regulating potential of this conservation agriculture practice requires combining it with other management approaches that redistribute SOC throughout the soil profile. For example, incorporating perennial cover grasses into the crop rotations increases the C supply to subsoil horizons through rhizodeposits and mortmass itself [42]. Another option is to select crop varieties with deeper and thicker root systems that contain high levels of chemical recalcitrant compounds, such as suberin [43].
4.2. Spatial Variability of POM to MAOM Ratio and Microbial-Available C Pool
The observed POM:MAOM ratios in the studied Chernozems are generally close to the median values for agricultural soils worldwide [44]. This is mainly due to reduced organic matter inputs relative to native ecosystems and the enhanced POM decomposition through tillage and mineral fertilization [44]. At the same time, an important functional role of POM in forming organo-mineral associations has been demonstrated by Witzgall et al. [45] by tracking the decomposition of isotopically labeled litter in soils. The authors showed that POM surfaces are not only constituting hotspots for microbial activity, but also the nucleus of aggregate formation and MAOM buildup. This highlights the necessity of developing sustainable approaches to enhance this parameter in agricultural soils. According to a recent global meta-analysis, the effective practices for increasing POM and MAOM pools are combinations of mineral N and organic matter inputs, such as straw, manure, and biochar [46]. Under such management, POM increased more rapidly (after six years) than MAOM (after two decades). Other useful approaches for managing the POM:MAOM ratio, as well as total SOC content, include diversifying crop rotations [47] and using cover crops [48].
We found that microbial-available C pool was sensitive to contrasting tillage practices, unlike the POM:MAOM ratio (Figure 6 and Figure 7). Previous studies have evidenced that NT practice significantly increases the abundance of simple carbohydrates, amino sugars, lignin phenols, long-chain aliphatic (≥C20) and cyclic lipids in the topsoil compared to CT practice [18,49,50]. In our case, however, changes in the microbial-available C pool relative to different tillage practices showed site-specific patterns (Figure 7). Specifically, this parameter increased under NT compared to CT at the Rostov site, but it showed an opposite trend at the Krasnodar site. This can be associated with distinct soil aeration conditions due to differences in soil bulk density (Table 1) and water table depth mediated by topography. At the Krasnodar site located in the lowland, the SOC content decreased significantly with soil bulk density, in contrast to the Rostov site (Figure 8). Apparently, applying NT relative to CT in lowlands worsens soil physical conditions, which controls the flow and bioavailability of gases (e.g., O2, CO2 and N2O), water, and nutrients [51]. These factors, in turn, are essential for soil microbial functioning and crop productivity.
4.3. Study Limitations and Further Prospects
The main challenge in predicting SOC dynamics in farmlands is the high spatial and temporal variability of management practices, including crop rotations, tillage systems, fertilizer application rates and methods, seeding frequency and depth, crop protection systems, machinery used, etc. Furthermore, due to the occasional change in farmland ownership, it is impossible to completely restore this information over time. Therefore, the observed patterns of SOC variability across cultivated Chernozems cannot be attributed to a fixed-factor effect, such as tillage systems. This gap underscores the necessity of systematically and uniformly monitoring all applied agricultural treatments in relation to each spatial (e.g., field) together with continuous tracking of SOC dynamics, as discussed previously by Olson et al. [52].
Nowadays, the mandatory monitoring system for Russian farmers is being actively developed and includes an annual report on crop-specific harvests, the used agrochemicals (fertilizers and pesticides), machinery, and tillage system. Starting in 2025, these data began to be integrated with the cadastral inventory map system, which provides their spatial linking at the field scale. In this context, our study can serve as a baseline for the further monitoring of SOC stocks and stability across farmlands. Moreover, accurate farm-scale maps of SOC with its constituent pools (i.e., POM and microbial-available C) are highly helpful for farmers’ decision-making in developing effective management strategies for long-term SOC accrual and enhanced total soil quality.
5. Conclusions
This study examined the farm-scale variability of SOC content and its functional pools across contrasting tillage practices in two different environmental settings of the Chernozem region. Two sites differed markedly in SOC content and microbial-available C pool. However, the POM:MAOM ratio revealed equal dominance of MAOM in the SOC of both sites. This suggests that long-term and intensive agricultural use of Chernozems makes this indicator of SOC quality uniform across difference environmental conditions.
Regarding contrasting tillage practices, only the microbial-available C pool showed clear and site-specific patterns, unlike SOC content and the POM:MAOM ratio. At the same time, the DNDC modeling showed that applying NT versus CT for 30 years could potentially increase SOC in the topsoil (0–10 cm) and decrease it in the subsoil (10–30 cm) across the studied Chernozems.
GBM models trained using remote sensing data (spectral and topographic) as explanatory spatial covariates achieved high predictive accuracy for topsoil SOC content and its microbial-available pool. This enabled us to create high-resolution maps of their distributions across the studied farmland areas. These results are valuable for designing targeted soil C management strategies at the farm scale, taking into account current agronomic practices.
Author Contributions
Conceptualization, S.S. and K.I.; methodology, S.S., K.I., Y.D., A.D. and V.S.; formal analysis, S.S., K.I., Y.D. and A.G.; investigation, A.P., A.G., G.F., A.R., E.M., D.S. and D.K.; data curation, S.S., K.I. and V.S.; writing—original draft preparation, S.S., Y.D. and A.D.; writing—review and editing K.I., I.A. and V.S.; visualization, S.S., Y.D. and A.D.; supervision, S.S., K.I. and V.S.; project administration, S.S. and K.I.; funding acquisition, S.S. and K.I. All authors have read and agreed to the published version of the manuscript.
Funding
Analysis of soil organic carbon pools was funded by the Russian Science Foundation under grant No. 24-76-10079 (https://rscf.ru/en/project/24-76-10079/). Soil sampling, physical and chemical analyses, and soil prediction modeling and mapping were supported by the state assignment of the Ministry of Science and Higher Education of the Russian Federation No. 125061107005-4.
Data Availability Statement
The raw data supporting the conclusions of this article will be made available by the authors on request.
Acknowledgments
We are grateful to the farmers in the Rostov and Krasnodar regions who gave us the opportunity to conduct our study in their fields. We would like to express our gratitude for the soil agrochemical analysis carried out by the Pushchino Scientific Center for Biological Research’s Center of Common Facilities.
Conflicts of Interest
The authors declare no conflicts of interest.
Appendix A
Table A1.
Soil chemical and physical properties along a 1 m profile of Calcic and Stagnic Chernozems under no-tillage over 11 yrs.
Table A2.
Soil properties used for initializing the DNDC model in the 0–10 cm layer. Txt: soil texture class; BD: bulk density (g cm−3); SOC: soil organic carbon content in the surface soil (kg C kg−1 soil); SOC dec.: SOC decrease with depth below the topsoil; Res. lit.: initial resistant litter fraction; HH: initial combined humads and humus fraction; DWRL: depth to the water retention level (m).
Figure A1.
Fitted versus observed values of soil organic carbon content (SOC) (left) and its mineralized pool for 180 days (right) for Chernozems’ topsoil (0–10 cm) under conventional tillage and no-tillage across studied farm areas.
References
- Anjos, L.; Gaistardo, C.C.; Deckers, J.; Dondeyne, S.; Eberhardt, E.; Gerasimova, M.; Harms, B.; Jones, A.; Krasilnikov, P.; Reinsch, T.; et al. World Reference Base for Soil Resources 2014 International Soil Classification System for Naming Soils and Creating Legends for Soil Maps. Available online: https://publications.jrc.ec.europa.eu/repository/handle/JRC91947 (accessed on 10 December 2025).
- Liu, X.; Lee Burras, C.; Kravchenko, Y.S.; Duran, A.; Huffman, T.; Morras, H.; Studdert, G.; Zhang, X.; Cruse, R.M.; Yuan, X. Overview of Mollisols in the World: Distribution, Land Use and Management. Can. J. Soil Sci. 2012, 92, 383–402. [Google Scholar] [CrossRef] [Scilit]
- Boincean, B.; Dent, D. Farming the Black Earth: Sustainable and Climate-Smart Management of Chernozem Soils; Springer International Publishing: Cham, Switzerland, 2019. [Google Scholar]
- Xu, X.; Pei, J.; Xu, Y.; Wang, J. Soil Organic Carbon Depletion in Global Mollisols Regions and Restoration by Management Practices: A Review. J. Soils Sediments 2020, 20, 1173–1181. [Google Scholar] [CrossRef] [Scilit]
- Just, C.; Armbruster, M.; Barkusky, D.; Baumecker, M.; Diepolder, M.; Döring, T.F.; Heigl, L.; Honermeier, B.; Jate, M.; Merbach, I.; et al. Soil Organic Carbon Sequestration in Agricultural Long-Term Field Experiments as Derived from Particulate and Mineral-Associated Organic Matter. Geoderma 2023, 434, 116472. [Google Scholar] [CrossRef] [Scilit]
- Angst, G.; Mueller, K.E.; Castellano, M.J.; Vogel, C.; Wiesmeier, M.; Mueller, C.W. Unlocking Complex Soil Systems as Carbon Sinks: Multi-Pool Management as the Key. Nat. Commun. 2023, 14, 2967. [Google Scholar] [CrossRef] [Scilit]
- Lützow, M.V.; Kögel-Knabner, I.; Ekschmitt, K.; Matzner, E.; Guggenberger, G.; Marschner, B.; Flessa, H. Stabilization of Organic Matter in Temperate Soils: Mechanisms and Their Relevance under Different Soil Conditions—A Review. Eur. J. Soil Sci. 2006, 57, 426–445. [Google Scholar] [CrossRef] [Scilit]
- Sollins, P.; Swanston, C.; Kramer, M. Stabilization and Destabilization of Soil Organic Matter—A New Focus. Biogeochemistry 2007, 85, 1–7. [Google Scholar] [CrossRef] [Scilit]
- Lavallee, J.M.; Soong, J.L.; Cotrufo, M.F. Conceptualizing Soil Organic Matter into Particulate and Mineral-associated Forms to Address Global Change in the 21st Century. Glob. Change Biol. 2020, 26, 261–273. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Semenov, V.M.; Lebedeva, T.N.; Sokolov, D.A.; Zinyakova, N.B.; Lopes De Gerenu, V.O.; Semenov, M.V. Measurement of the Soil Organic Carbon Pools Isolated Using Bio-Physical-Chemical Fractionation Methods. Eurasian Soil Sci. 2023, 56, 1327–1342. [Google Scholar] [CrossRef] [Scilit]
- Sushko, S.; Dvornikov, Y.; Neprimerova, S.; Ivashchenko, K.; Yudina, A.; Grigorova, A.; Kovaleva, N.; Orlova, L.; Hu, C. Soil Mineral-Associated Organic Matter under Conventional and No-till Farming: Spatial Drivers and Local-Scale Mapping. Agric. Ecosyst. Environ. 2026, 399, 110207. [Google Scholar] [CrossRef] [Scilit]
- Powlson, D.S.; Stirling, C.M.; Jat, M.L.; Gerard, B.G.; Palm, C.A.; Sanchez, P.A.; Cassman, K.G. Limited Potential of No-till Agriculture for Climate Change Mitigation. Nat. Clim. Change 2014, 4, 678–683. [Google Scholar] [CrossRef] [Scilit]
- Du, Z.; Angers, D.A.; Ren, T.; Zhang, Q.; Li, G. The Effect of No-till on Organic C Storage in Chinese Soils Should Not Be Overemphasized: A Meta-Analysis. Agric. Ecosyst. Environ. 2017, 236, 1–11. [Google Scholar] [CrossRef] [Scilit]
- Ogle, S.M.; Alsaker, C.; Baldock, J.; Bernoux, M.; Breidt, F.J.; McConkey, B.; Regina, K.; Vazquez-Amabile, G.G. Climate and Soil Characteristics Determine Where No-Till Management Can Store Carbon in Soils and Mitigate Greenhouse Gas Emissions. Sci. Rep. 2019, 9, 11665. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lal, R.; Reicosky, D.C.; Hanson, J.D. Evolution of the Plow over 10,000 Years and the Rationale for No-till Farming. Soil Tillage Res. 2007, 93, 1–12. [Google Scholar] [CrossRef] [Scilit]
- Li, Y.; Chang, S.X.; Tian, L.; Zhang, Q. Conservation Agriculture Practices Increase Soil Microbial Biomass Carbon and Nitrogen in Agricultural Soils: A Global Meta-Analysis. Soil Biol. Biochem. 2018, 121, 50–58. [Google Scholar] [CrossRef] [Scilit]
- Li, Y.; Song, D.; Liang, S.; Dang, P.; Qin, X.; Liao, Y.; Siddique, K.H.M. Effect of No-Tillage on Soil Bacterial and Fungal Community Diversity: A Meta-Analysis. Soil Tillage Res. 2020, 204, 104721. [Google Scholar] [CrossRef] [Scilit]
- Ma, L.; Wang, X.; Fang, Y.; Vancov, T.; Jin, X.; Gao, Q.; Dong, W.; Du, Z. No-Tillage Farming for Two Decades Increases Plant- and Microbial-Biomolecules in the Topsoil Rather than Soil Profile in Temperate Agroecosystem. Soil Tillage Res. 2024, 241, 106108. [Google Scholar] [CrossRef] [Scilit]
- Briedis, C.; De Moraes Sá, J.C.; Lal, R.; Tivet, F.; Franchini, J.C.; De Oliveira Ferreira, A.; Da Cruz Hartman, D.; Schimiguel, R.; Bressan, P.T.; Inagaki, T.M.; et al. How Does No-till Deliver Carbon Stabilization and Saturation in Highly Weathered Soils? Catena 2018, 163, 13–23. [Google Scholar] [CrossRef] [Scilit]
- Blanco-Moure, N.; Gracia, R.; Bielsa, A.C.; López, M.V. Long-term No-tillage Effects on Particulate and Mineral-associated Soil Organic Matter under Rainfed M Editerranean Conditions. Soil Use Manag. 2013, 29, 250–259. [Google Scholar] [CrossRef] [Scilit]
- Conceição, P.C.; Dieckow, J.; Bayer, C. Combined Role of No-Tillage and Cropping Systems in Soil Carbon Stocks and Stabilization. Soil Tillage Res. 2013, 129, 40–47. [Google Scholar] [CrossRef] [Scilit]
- Fabrizzi, K.P.; Morón, A.; García, F.O. Soil Carbon and Nitrogen Organic Fractions in Degraded vs. Non-Degraded Mollisols in Argentina. Soil Sci. Soc. Am. J. 2003, 67, 1831–1841. [Google Scholar] [CrossRef] [Scilit]
- Garsia, A.; Moinet, A.; Vazquez, C.; Creamer, R.E.; Moinet, G.Y.K. The Challenge of Selecting an Appropriate Soil Organic Carbon Simulation Model: A Comprehensive Global Review and Validation Assessment. Glob. Change Biol. 2023, 29, 5760–5774. [Google Scholar] [CrossRef] [Scilit]
- Huang, H.; Yang, L.; Zhang, L.; Pu, Y.; Yang, C.; Wu, Q.; Cai, Y.; Shen, F.; Zhou, C. A Review on Digital Mapping of Soil Carbon in Cropland: Progress, Challenge, and Prospect. Environ. Res. Lett. 2022, 17, 123004. [Google Scholar] [CrossRef] [Scilit]
- McBratney, A.B.; Mendonça Santos, M.L.; Minasny, B. On Digital Soil Mapping. Geoderma 2003, 117, 3–52. [Google Scholar] [CrossRef] [Scilit]
- Minasny, B.; McBratney, A.B. A Conditioned Latin Hypercube Method for Sampling in the Presence of Ancillary Information. Comput. Geosci. 2006, 32, 1378–1388. [Google Scholar] [CrossRef] [Scilit]
- Tucker, C.J. Red and Photographic Infrared linear Combinations for Monitoring Vegetation. Remote Sens. Environ. 1979, 8, 127–150. [Google Scholar] [CrossRef] [Scilit]
- Beven, K.J.; Kirkby, M.J. A Physically Based, Variable Contributing Area Model of Basin Hydrology/Un Modèle à Base Physique de Zone d’appel Variable de l’hydrologie Du Bassin Versant. Hydrol. Sci. Bull. 1979, 24, 43–69. [Google Scholar] [CrossRef] [Scilit]
- Farr, T.G.; Rosen, P.A.; Caro, E.; Crippen, R.; Duren, R.; Hensley, S.; Kobrick, M.; Paller, M.; Rodriguez, E.; Roth, L.; et al. The Shuttle Radar Topography Mission. Rev. Geophys. 2007, 45, 2005RG000183. [Google Scholar] [CrossRef] [Scilit]
- R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2024. [Google Scholar]
- Roudier, P.; Beaudette, D.E.; Hewitt, A.E. A Conditioned Latin Hypercube Sampling Algorithm Incorporating Operational Constraints. In Digital Soil Assessments and Beyond; CRC Press: Boca Raton, FL, USA, 2012; p. 6. [Google Scholar]
- Li, C.; Frolking, S.; Frolking, T.A. A Model of Nitrous Oxide Evolution from Soil Driven by Rainfall Events: 1. Model Structure and Sensitivity. J. Geophys. Res. 1992, 97, 9759–9776. [Google Scholar] [CrossRef] [Scilit]
- Blagodatskiy, S.A.; Blagodatskaya, Y.V.; Gorbenko, A.Y.; Panikov, N.S. A Rehydration Method of Determining the Biomass of Microorganisms in Soil. Sov. Soil Sci. 1987, 19, 119–126. [Google Scholar]
- Semenov, V.M.; Kravchenko, I.K.; Ivannikova, L.A.; Kuznetsova, T.V.; Semenova, N.A.; Gispert, M.; Pardini, J. Experimental Determination of the Active Organic Matter Content in Some Soils of Natural and Agricultural Ecosystems. Eurasian Soil Sci. 2006, 39, 251–260. [Google Scholar] [CrossRef] [Scilit]
- Fox, J.; Weisberg, S. An R Companion to Applied Regression, 3rd ed.; AGE Publications Inc.: Thousand Oaks, CA, USA, 2019. [Google Scholar]
- Signorell, A. DescTools: Tools for Descriptive Statistics. R Package Version 0.99.60.24. 2025. Available online: https://github.com/andrisignorell/desctools (accessed on 15 September 2025).
- Wickham, H. Ggplot2: Elegant Graphics for Data Analysis, 2nd ed.; Springer International Publishing: Cham, Switzerland, 2016. [Google Scholar]
- Friedman, J.H. Greedy Function Approximation: A Gradient Boosting Machine. Ann. Stat. 2001, 29, 1189–1232. [Google Scholar] [CrossRef] [Scilit]
- Xu, H. Modification of Normalised Difference Water Index (NDWI) to Enhance Open Water Features in Remotely Sensed Imagery. Int. J. Remote Sens. 2006, 27, 3025–3033. [Google Scholar] [CrossRef] [Scilit]
- Ermida, S.L.; Soares, P.; Mantas, V.; Göttsche, F.-M.; Trigo, I.F. Google Earth Engine Open-Source Code for Land Surface Temperature Estimation from the Landsat Series. Remote Sens. 2020, 12, 1471. [Google Scholar] [CrossRef] [Scilit]
- Ruis, S.J.; Blanco-Canqui, H. How Does No-till Affect Soil-Profile Distribution of Roots? Can. J. Soil Sci. 2024, 104, 350–361. [Google Scholar]
- Pausch, J.; Kuzyakov, Y. Carbon Input by Roots into the Soil: Quantification of Rhizodeposition from Root to Ecosystem Scale. Glob. Change Biol. 2018, 24, 1–12. [Google Scholar] [CrossRef] [Scilit]
- Lorenz, K.; Lal, R. The Depth Distribution of Soil Organic Carbon in Relation to Land Use and Management and the Potential of Carbon Sequestration in Subsoil Horizons. In Advances in Agronomy; Elsevier: Amsterdam, The Netherlands, 2005; Volume 88, pp. 35–66. [Google Scholar]
- Sokol, N.W.; Whalen, E.D.; Jilling, A.; Kallenbach, C.; Pett-Ridge, J.; Georgiou, K. Global Distribution, Formation and Fate of Mineral-associated Soil Organic Matter under a Changing Climate: A Trait-based Perspective. Funct. Ecol. 2022, 36, 1411–1429. [Google Scholar] [CrossRef] [Scilit]
- Witzgall, K.; Vidal, A.; Schubert, D.I.; Höschen, C.; Schweizer, S.A.; Buegger, F.; Pouteau, V.; Chenu, C.; Mueller, C.W. Particulate Organic Matter as a Functional Soil Component for Persistent Soil Organic Carbon. Nat. Commun. 2021, 12, 4115. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zhou, R.; Kumar, A.; Gunina, A.; Ling, J.; Cui, Z.; Kuzyakov, Y.; Tian, J.; Xue, Y. Particulate and Mineral-Associated Organic Matter in Cropland Soils: Meta-Analysis of Management Effects. Agric. Ecosyst. Environ. 2026, 400, 110218. [Google Scholar] [CrossRef] [Scilit]
- Adingo, S.; Jiang, R.; Guo, Z.; Gao, L.; Liu, J.; Liu, S.; Peng, X.; Jiang, F. Crop Rotation Impacts on SOC Fraction Contents and Stability: Insights from a Global Meta-Analysis. Agric. Ecosyst. Environ. 2026, 399, 110128. [Google Scholar] [CrossRef] [Scilit]
- Zhang, Z.; Kaye, J.P.; Bradley, B.A.; Amsili, J.P.; Suseela, V. Cover Crop Functional Types Differentially Alter the Content and Composition of Soil Organic Carbon in Particulate and Mineral-associated Fractions. Glob. Change Biol. 2022, 28, 5831–5848. [Google Scholar] [CrossRef] [Scilit]
- Veloso, M.G.; Dick, D.P.; Costa, J.B.D.; Bayer, C. Cropping Systems Including Legume Cover Crops Favour Mineral–Organic Associations Enriched with Microbial Metabolites in No-till Soil. Soil Res. 2019, 57, 851–858. [Google Scholar] [CrossRef] [Scilit]
- Veloso, M.G.; Angers, D.A.; Chantigny, M.H.; Bayer, C. Mineral–Organic Associations Are Enriched in Both Microbial Metabolites and Plant Residues in a Subtropical Soil Profile under No-Tillage and Legume Cover Cropping. Soil Res. 2022, 60, 590–600. [Google Scholar] [CrossRef] [Scilit]
- Hartmann, M. Soil Structure and Microbiome Functions in Agroecosystems. Nat. Rev. Earth Environ. 2023, 4, 4–18. [Google Scholar] [CrossRef] [Scilit]
- Olson, K.R.; Al-Kaisi, M.M.; Lal, R.; Lowery, B. Experimental Consideration, Treatments, and Methods in Determining Soil Organic Carbon Sequestration Rates. Soil Sci. Soc. Am. J. 2014, 78, 348–360. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 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.












