1 Introduction

Cucumber is a high‑value vegetable widely cultivated in protected facilities, where continuous monoculture is common to maximize land use and profit (Gao et al., 2021; Huang et al., 2023). However, long‑term single‑crop systems often trigger continuous cropping obstacles, including soil degradation, yield decline, and increased disease, threatening the sustainability of greenhouse cucumber production. Understanding the processes and mechanisms of soil degradation under continuous cucumber cropping is therefore crucial for guiding rational management and maintaining stable production.

 

Facility cucumber monoculture can cause salinization, acidification, nutrient imbalance, and shifts in soil microbial communities, which together lead to soil quality deterioration and crop obstacles (Huang et al., 2023; Gao et al., 2021; Zhao et al., 2020). Studies in different cucumber production bases show increasing soil total salts, decreasing pH, and accumulation of nitrate nitrogen and available nutrients with increasing years of continuous cropping, often accompanied by deterioration of soil microbial community structure. These changes can promote soilborne diseases, disrupt plant-soil interactions, and reduce fruit yield and quality. More broadly, continuous cropping of various crops is now recognized as a major driver of land degradation and nutrient loss, with clear impacts on soil health and agroecosystem stability. Elucidating soil degradation characteristics specific to continuous cucumber cropping can provide a theoretical basis and technical reference for optimizing fertilization, improving cultivation systems, and designing remediation measures in protected vegetable production. 

 

Existing research on continuous cucumber cropping focuses mainly on changes in soil physicochemical properties, enzyme activities, and microbial communities under different monoculture durations or cultivation patterns. Long‑term cucumber monoculture in greenhouses has been shown to increase soil organic carbon and nitrogen but also to raise salinity and acidity and reduce bacterial and fungal diversity, with fungal communities particularly sensitive (Gao et al., 2021; Zhao et al., 2020). High‑throughput sequencing studies reveal that continuous cucumber cropping decreases overall microbial diversity, alters dominant bacterial and fungal taxa, and reshapes microbial networks, often favoring potential pathogens and reducing beneficial groups. At the same time, rotation or alternative patterns such as paddy‑upland and garlic rotation can improve soil microbial structure, increase beneficial microbiota, alleviate acidification, and enhance cucumber growth. More broadly, reviews on continuous cropping emphasize that persistent monoculture alters multiple abiotic indicators (salts, acids, toxic metabolites, structure) and biotic indicators (microbial diversity, networks, enzyme activities, soil food webs), accelerating soil degradation. However, most cucumber studies emphasize microbials or yield response, while systematic characterization of soil degradation characteristics-including coupled evolution of physical, chemical, and biological properties across continuous‑cropping gradients-remains relatively limited for typical cucumber facility systems. 

 

In view of the above gaps, the present study takes continuous cucumber cropping soils as the object and aims to clarify the soil degradation characteristics under different continuous‑cropping years. The specific objectives are: (1) to analyze the evolution of key soil physicochemical indicators (pH, salinity, organic matter, major nutrients) with increasing years of continuous cucumber cropping and identify dominant degradation trends; (2) to reveal changes in soil biological properties, focusing on microbial diversity, community composition, and functional groups related to soil health and crop growth (Sun et al., 2021); and (3) to comprehensively evaluate the relationships between soil environmental changes and continuous cropping intensity, thereby summarizing the main degradation patterns and potential driving mechanisms (Pervaiz et al., 2020). Technically, the study will select soils under different continuous‑cropping durations and non‑continuous controls in typical cucumber greenhouse systems, and combine measurements of soil physicochemical properties and enzyme activities with high‑throughput sequencing of bacterial and fungal communities, as used in recent cucumber and other continuous‑cropping studies. Multivariate statistical analyses such as principal coordinates analysis, redundancy analysis, and correlation analysis will be employed to link soil properties with microbial community changes and cropping duration. By constructing an integrated framework of “continuous cropping years-soil environment-microbial community-soil degradation characteristics”, the study seeks to provide scientific support for diagnosing soil degradation under continuous cucumber cropping and for developing targeted management and remediation strategies in facility vegetable production (Zhao et al., 2017).

 

2 Vineyard Microclimate Factors and Their Mechanisms

2.1 Effects of temperature on grape quality

Temperature controls berry development rate and the balance between sugars, acids, and phenolics, thereby shaping ripening trajectory and typicity of wines such as Bordeaux reds. Moderate warmth accelerates ripening, but sustained elevated temperatures can decouple sugar accumulation from color development, reducing anthocyanin concentration at a given soluble solids level (Arrizabalaga et al., 2018; De Rosas et al., 2022). 

 

High or extreme high temperatures directly affect flavonoid biosynthesis, often decreasing total anthocyanins and altering the profile of delphinidin-, cyanidin-, and malvidin-derived pigments (Lecourieux et al., 2017; Gouot et al., 2018). These thermal effects are cultivar- and clone-dependent, with some varieties or clones showing greater plasticity or resilience in color attributes under a +2 °C regime than others. 

 

2.2 Regulatory role of light and radiation conditions

Light quantity and spectral quality act jointly with temperature to regulate berry metabolism, particularly flavonols and anthocyanins (Figure 1) (Blancquaert et al., 2019). Increased solar irradiance within the cluster zone modifies spatial patterns of sugars, organic acids, amino acids, and phenylpropanoids, often lowering malate but enhancing stress-related metabolites such as proline and GABA (Reshef et al., 2017). 

 

Figure 1 Conceptual diagram of the joint effects of light quantity, spectral quality, and temperature on grape berry metabolism. Increased solar irradiance and UV-B exposure can stimulate flavonol and anthocyanin biosynthesis, whereas excessive heat may reduce anthocyanin accumulation and alter ripening dynamics (Adopted from Blancquaert et al., 2019)

 

Ultraviolet-B (UV-B) radiation specifically stimulates flavonol and anthocyanin biosynthesis, partly counteracting the anthocyanin losses caused by elevated CO₂ and temperature (Martínez-Lüscher et al., 2016). Canopy and net shading that filter sunlight or UV-B can reduce berry surface temperature and wind speed, but also diminish photosynthetically active radiation and UV fractions, leading to complex changes in flavonoid composition and ripening dynamics (Crouchett-Rojas et al., 2025). 

 

2.3 Combined effects of humidity and wind speed

Humidity and wind modify the vineyard microclimate by influencing leaf and berry boundary layers, pathogen development, and heat exchange. High relative humidity around 80-85% strongly favors powdery mildew development and conidial germination on grape foliage, while very high humidity can reduce germination frequency (Carroll and Wilcox, 2003). In contrast, lower humidity combined with adequate air movement tends to suppress epidemic intensity by reducing leaf wetness duration and spore viability. 

 

Wind speed interacts with radiation and temperature to alter berry surface temperature and susceptibility to sunburn necrosis. Higher wind speeds enhance convective cooling of clusters, lowering fruit temperature and reducing sunburn severity even under strong solar exposure (Müller et al., 2023). Conversely, reduced wind, together with low temperature variability and high humidity, supports the rapid build-up of diseases such as anthracnose, demonstrating how humidity-wind interactions can explain most of the variation in disease severity under field conditions (Chavarria et al., 2009). 

 

Temperature, light-radiation, humidity, and wind act together at berry scale to determine grape metabolic profiles, color, acidity, and disease pressure. High temperatures and intense radiation generally threaten color and phenolic quality, but targeted use of shading or UV-B exposure can mitigate some losses. Humidity and wind primarily modulate disease risk and sunburn damage, meaning that successful vineyard management must consider the full microclimate ensemble rather than any single factor in isolation.

 

3 Construction of Grape Quality Evaluation System

3.1 Physicochemical indicators

The basic physicochemical profile of grape berries is largely defined by sugars and organic acids, typically measured as total soluble solids (TSS), titratable acidity (TA), pH, and specific sugar-acid compositions (Kunter et al., 2024). Glucose and fructose are the predominant sugars in table grapes and together account for nearly all soluble sugars, while sucrose is usually absent or negligible. Their absolute contents and ratio determine sweetness intensity and are strongly affected by cultivar genetics and seasonal conditions, leading to marked year‑to‑year variability in TSS and sugar composition (Lu et al., 2024). 

 

Organic acids, mainly tartaric and malic acids, provide the acidic counterpart to sugars and are critical for taste balance and technological properties of grape juice and wine (Zhang et al., 2021). Variation in TA and in the tartaric/malic ratio among cultivars is substantial, and cultivars with high TA show much lower SSC/TA (sweetness-acidity) ratios than those with low acidity, even at similar TSS. The SSC/TA or TSS/acid ratio is widely used as a maturity or harvest index because it better reflects the perceived balance of sweetness and acidity than TSS or acidity alone (Kumar et al., 2025). 

 

3.2 Sensory quality and flavor compounds

Grape sensory quality is jointly determined by appearance, texture, taste, and aroma, with flavor (taste + aroma) playing a dominant role in consumer preference (Wang et al., 2023). Panel studies on diverse table grape lines show that overall flavor impression strongly correlates with liking scores, and within flavor categories, fruity notes are particularly associated with higher preference (Maoz et al., 2020). Sensory evaluations that integrate sweetness, acidity, firmness, and aroma have been used to construct multivariate quality indices, revealing that cultivars with balanced sweet-sour taste and crisp texture often receive the highest comprehensive scores. 

 

Volatile organic compounds (VOCs) underpin cultivar‑specific aromas such as Muscat, foxy, or herbaceous types (Maoz et al., 2020). In many table grapes, a small set of C6 aldehydes, especially 1‑hexanal and (E)‑2‑hexenal, represents the “core” volatiles and can account for most of the total volatile concentration, contributing fresh and green notes. Broader profiling across cultivars identifies terpenes, esters, higher alcohols, and other volatiles as key drivers of floral, fruity, and varietal aromas, and higher overall volatile concentrations or the presence of unique compounds are often associated with stronger or more preferred flavors (Moriyama et al., 2024). 

 

3.3 Secondary metabolites and functional components

Grapes are rich in secondary metabolites such as polyphenols (anthocyanins, flavonols, flavanols, stilbenes) and vitamins, which strongly influence color, astringency, and nutritional value (Flamini et al., 2013; Zhou et al., 2022). Anthocyanins determine the red to black coloration of many cultivars and contribute to perceived quality, while also providing antioxidant and potential cardioprotective, anticancer, and anti‑inflammatory benefits . Flavonols and hydroxycinnamic acids participate in copigmentation and photo‑protection and add to the overall antioxidant capacity of fresh grapes and derived products. 

 

Stilbenes, particularly resveratrol and its derivatives, act as phytoalexins in the berry and are implicated in many of the reported health effects of grape and wine, including cardioprotective and neuroprotective activities (Figure 2) (Hegedüs et al., 2022; Buljeta et al., 2023). Comparative studies among table grape cultivars show large differences in total phenolics, resveratrol, vitamin C, carotenoids, and tocopherols, with some black or Muscat‑type cultivars especially rich in these bioactives. These compositional differences mean that cultivar choice and production systems can be tailored not only for sensory quality but also for enhanced functional properties and potential health benefits.

 

Figure 2 Visual comparison of grape cultivars with different skin colors associated with anthocyanin accumulation. Higher anthocyanin content contributes to deeper red-to-black coloration and enhanced perceived quality in grapes (Adopted from Hegedüs et al., 2022)

 

4 Microclimate Data Collection and Processing Methods

4.1 Sensor deployment and data acquisition

High‑quality monitoring of soil degradation under continuous cucumber cropping depends first on rational deployment of in‑situ sensors and robust data acquisition architecture. Recent agricultural IoT systems typically couple soil temperature, moisture, EC and pH probes with microcontrollers or embedded computers, and then relay data via Wi‑Fi, LoRaWAN or cellular links to a cloud platform for storage and visualization (Wu et al., 2023). In multi‑sensor soil monitoring, nodes are often buried at several depths to capture vertical gradients in water and salinity, while spatially distributed nodes describe heterogeneity across the field or greenhouse (Selvanarayanan et al., 2024). 

 

Power management and sampling design are also central to stable long‑term soil monitoring. Low‑power hardware, duty‑cycled operation, and modular power supplies are used to extend node lifetime while maintaining sufficiently high sampling frequency for time‑series diagnosis. Wireless sensor networks developed for soil moisture or fertigation monitoring illustrate how customized communication protocols and multi‑hop topologies can ensure data delivery even in vegetated or obstructed environments. Such architectures provide continuous, high‑resolution data streams required to track subtle degradation processes in intensively cropped soils (Chamara et al., 2022). 

 

4.2 Data preprocessing and quality control

Raw time‑series collected from soil sensor networks inevitably contain gaps, spikes, drifts and other anomalies due to sensor aging, power interruptions or communication failures. Reviews of real‑time continuous soil monitoring emphasize that careful preprocessing-unit harmonization, temporal alignment, noise filtering and gap filling-is essential before analysis or modeling (Fan et al., 2022). Global soil moisture networks similarly rely on automated routines to flag values outside geophysical ranges and to check consistency with auxiliary variables such as soil temperature and rainfall (Dorigo et al., 2013). 

 

Beyond threshold checks, advanced quality‑control schemes apply time‑series and machine‑learning methods to detect complex anomalies. For example, bidirectional LSTM models have been used to identify nonphysical patterns in soil moisture sensor records more accurately than traditional spectral or rule‑based approaches (Bandaru et al., 2024). Other work embeds sensor data into process‑based feature spaces and trains classifiers to distinguish compromised sensors from normal behavior, reducing manual QA/QC burden in large environmental networks (Schmidt and Kerkez, 2023). These approaches can be transferred to soil‑degradation monitoring to maintain reliable long‑term datasets.

 

4.3 Multi‑source data integration techniques

Understanding soil degradation in continuous cucumber systems benefits from combining in‑situ soil sensing with other data sources such as microclimate records, remote sensing imagery and farm management logs. Ag‑IoT frameworks describe complete data pipelines where sensor, drone, and satellite measurements are integrated through multi‑layer communication architectures and cloud platforms to support crop modeling and decision‑making (Odedeyi et al., 2025). Geo‑intelligent agriculture concepts further highlight the value of GIS and geostatistical modeling for fusing spatial sensor data with remotely sensed indices to delineate management zones and degradation hotspots (Wassay et al., 2026). 

 

At the analytical level, multi‑source integration often relies on data fusion and feature‑engineering techniques. Reviews of smart sensing in precision agriculture discuss edge and cloud computing strategies that aggregate heterogeneous sensor streams, apply calibration and validation, and then feed harmonized datasets to machine‑learning models for soil assessment or pollutant monitoring (Akhtar et al., 2021). Case studies in yield optimization have already demonstrated that combining soil moisture, EC, temperature and atmospheric variables improves prediction of crop performance compared with single‑source inputs (Odedeyi et al., 2025). Applying similar fusion frameworks to continuous cucumber cropping can reveal linkages between soil property dynamics, microclimate fluctuations and management practices, thereby clarifying degradation mechanisms.

 

5 Modeling Methods for Vineyard Microclimate

5.1 Statistical models

Statistical models are widely used to quantify relationships between soil properties, management factors and degradation indicators, and to build prediction equations for unsampled locations or future scenarios. Multiple linear regression and related techniques have been applied to link soil physicochemical variables and topographic factors to yield or erosion parameters, showing that simple linear structures can explain a substantial portion of variability in agricultural systems when relationships are approximately linear (Burdett and Wellen, 2022). For example, non‑linear regression has also been used to derive adjustment functions for soil erodibility and critical shear stress based on crop and management information, providing compact formulas that can be embedded into larger erosion models (Lee et al., 2021). 

 

In degradation studies, such regression models are useful for exploring which soil variables (e.g., pH, organic matter, salinity) most strongly influence crop performance under continuous cropping. However, comparisons with more flexible approaches show that linear models often underperform when responses involve threshold effects or strong interactions (Liu et al., 2023). In mapping soil chemical properties from remote sensing, multiple linear regression could not adequately capture non‑linear relationships between spectral indices and soil parameters, resulting in lower prediction accuracy than more advanced methods. This limitation is important for continuous cucumber systems, where soil processes such as salinization, acidification and nutrient imbalance may respond non‑linearly to cropping duration and management intensity. 

 

5.2 Machine learning models

Machine learning models have been increasingly introduced to agricultural and soil studies to address non‑linear and high‑dimensional relationships. Ensemble tree methods such as Random Forest and related decision‑tree regressors have repeatedly outperformed classical regression for predicting crop yield from soil, topographic and management attributes, achieving R² values above 0.8 at fine spatial scales (Burdett and Wellen, 2022). Similar superiority has been demonstrated when estimating multiple soil properties (texture, EC, pH, nutrients, organic matter) from long‑term satellite NDVI products and ancillary data, where Random Forest yielded the highest mapping accuracy across most indicators (Liu et al., 2023). 

 

Support Vector Machines (SVM/SVR) and other kernel‑based methods also provide strong performance for both classification and regression tasks based on soil and climate features. In integrated crop and soil decision systems, SVM and Random Forest have been used together to classify soil quality, recommend crops, and predict yields from key soil fertility indicators and environmental variables (Sadasivan et al., 2025). Other work focusing on soil health under climate stress found that Random Forest provided the most robust predictions of pH, organic matter and moisture content among several machine learning techniques, due to its ability to represent complex non‑linear interactions (Alsalami et al., 2025). For continuous cucumber cropping, such algorithms are well suited to learn degradation patterns from multi‑source datasets combining soil tests, remote sensing, and management history. 

 

5.3 Mechanistic models and coupled models

Mechanistic models describe soil processes through explicit representations of carbon, nitrogen, water and energy flows, and can be used to infer degradation trajectories under different cropping systems. New generations of soil organic matter models such as MEMS 2.0 and ORCHIMIC explicitly simulate particulate and mineral‑associated organic matter, microbial biomass, and nitrogen cycling through vertically resolved soil profiles, allowing evaluation of how management and climate alter carbon and nutrient stocks over time (Figure 3) (Zhang et al., 2021; Huang et al., 2021). Microbially explicit frameworks like MIMICS‑CN couple carbon and nitrogen dynamics across litter, microbial and soil organic matter pools, and have been calibrated across wide climatic gradients to produce realistic equilibrium C and N stocks and process rates. 

 

Figure 3 Conceptual representation of soil carbon and nitrogen cycling in vertically structured soil profiles. Mechanistic models simulate interactions among particulate organic matter, mineral-associated organic matter, microbial biomass, and nutrient fluxes to evaluate soil processes under different environmental conditions (Adopted from Zhang et al., 2021)

 

Beyond stand‑alone soil modules, coupled ecosystem models link plant growth, soil biogeochemistry and hydrology, making it possible to analyze feedbacks between continuous cropping, root inputs, soil water, and microclimate. For example, process‑based models have been used to simulate long‑term C and N mineralization of exogenous organic matter amendments, partitioning inputs among multiple soil pools and quantifying mineralization kinetics under controlled conditions (Pérémé et al., 2023). Such mechanistic or coupled models provide a process‑oriented complement to data‑driven approaches, and, when combined with statistical or machine‑learning components for parameter estimation or spatial extrapolation, can form hybrid frameworks to diagnose and forecast soil degradation under continuous cucumber cropping and evaluate remediation scenarios.

 

6 Quantitative Analysis of Microclimate Effects on Grape Quality

6.1 Single-factor impact analysis

Temperature alone can alter berry growth, sugar loading and phenolic balance in a quantifiable way. In controlled in vitro systems and greenhouse regimes, moderately higher temperatures increased sugar accumulation but accelerated softening and reduced titratable acidity, indicating a shift toward faster, but often less balanced, ripening (Arrizabalaga et al., 2018; Zha et al., 2024). Under elevated day/night temperatures, anthocyanin concentration and color intensity declined at a given sugar level, confirming a thermal decoupling between sugar and pigment accumulation. High temperature effects on flavonoids are largely negative and often linear in magnitude, with total anthocyanins generally decreasing as thermal load rises across varieties and experimental setups. Quantitative reviews and cultivar-specific trials report that elevated temperatures reduce anthocyanins more consistently than other flavonoids, making color traits particularly vulnerable in warm sites. 

 

Light exposure acts as a second key single factor, especially for skin phenolics and flavonols. Cluster and berry-scale studies show that sun-exposed berries can be statistically separated from shaded ones based on flavonol profiles, with higher quercetin-3-glucoside, kaempferol-3-glucoside and related compounds under stronger light. At the bunch and berry-half level, total flavonol concentration increases approximately linearly with estimated incident radiation, demonstrating a direct dose-response of flavonols to local light conditions (Pieri et al., 2016). Microclimate manipulations that modify light or UV, such as plastic covers or varying leaf removal intensity, also shift berry metabolites. Excluding most UV but slightly warming the fruit zone changed amino acid and organic acid composition, while still allowing discrimination of shaded versus lit berries through metabolic fingerprints. In several cultivars, reduced UV combined with warmer bunch temperatures decreased total anthocyanins and specific di-hydroxylated forms, indicating that both light quantity and quality (UV fraction) quantitatively shape color composition (Wilson et al., 2024; Wang et al., 2025). 

 

6.2 Multi-factor interaction analysis

Interactions among temperature, light, and soil/air conditions often explain more variation in grape quality than any single driver. In high-altitude vineyards analyzed with principal component analysis and structural equation modeling, microclimatic temperature variables-especially soil temperature, day-night temperature range and daytime air temperature-had stronger direct effects on composite quality indices than light alone, yet their influence was expressed through traits such as malic acid, total phenols, tannins, anthocyanins and skin thickness (Zhang et al., 2024). Multi-year mesoclimate-microclimate comparisons further show that altitude primarily modulates temperature, while row orientation shapes light intensity and quality, together determining soluble sugar and anthocyanin patterns across canopy sides. 

 

Row orientation-driven microclimate gradients create complex temporal and cumulative effects on berry temperature and composition. Canopy faces that received morning radiation but cooled later in the day displayed slightly advanced sugar ripening and generally higher skin total anthocyanins and phenols without major changes in pH or titratable acidity, whereas bunches accumulating peak heat in late afternoon showed composition less compatible with high quality (Hunter et al., 2021). Detailed analyses at berry and bunch levels reveal nested interactions: bunch azimuth, berry position and berry side jointly control local radiation and temperature, leading to large within-bunch heterogeneity in flavonols and anthocyanins that is partly masked when data are aggregated to whole clusters. 

 

Microclimate also interacts with management practices such as mulching and canopy manipulation. In a semi-arid table-grape system, black geotextile mulch raised air and soil temperatures and increased soluble solids, reducing sugars and sugar-acid ratios, while grass mulch cooled the soil, decreased soluble solids but still slightly increased total anthocyanins relative to clean tillage (Hu et al., 2022). Canopy microclimate modified by shoot length and lateral shoot retention altered total soluble solids, titratable acidity, phenolics and anthocyanins; optimal lateral retention depended on whether the season was hot and dry or cool and humid, underscoring that the same structural practice can enhance or impair quality depending on prevailing macro- and mesoclimate (Candar, 2019). 

 

6.3 Sensitivity and contribution analysis

Quantitative modeling has been used to rank the sensitivity of grape quality traits to different microclimate components and management-related drivers. Multiblock partial least squares models partitioned variance in sugar, acids, pH, nitrogen and bunch rot among climate, soil, perennial structures and annual practices; for many traits, soil and annual soil practices contributed as much or more explanatory power than the climate block, highlighting that microclimate effects are filtered through soil-management interactions (Beauchet et al., 2020). In machine-learning models of yield components, accumulated growing degree days, irrigation amounts and reference evapotranspiration emerged as dominant predictors, with non-linear responses indicating optimal ranges beyond which additional heat or water reduced cluster weight or yield (Ohana-Levi et al., 2024). 

 

At berry composition level, path models and structural equation approaches quantify how strongly specific microclimate variables propagate through metabolic networks to sugars, acids and phenolics. In high-altitude sites, soil temperature showed the highest standardized contribution to composite quality indices, followed by day-night temperature range and daytime air temperature, while light played a smaller but still significant role through anthocyanins and skin traits (Zhang et al., 2024). Network-based path modeling of Cabernet Sauvignon under various climate scenarios found that changes in temperature and radiation jointly increased sugars and a suite of aromatic compounds (tannins, phenols, flavanols, anthocyanins), predicting future wines with higher alcohol potential and flavor intensity under moderate warming and earlier ripening onset (Wood et al., 2024). 

 

Microclimate around grape clusters has quantifiable, often strong effects on berry sugars, acids and phenolic composition. Temperature and radiation are primary drivers, but their impacts are modulated by altitude, row orientation, canopy structure, soil and management. Multivariate and path models show that quality traits are highly sensitive to specific thermal and light regimes, with non-linear responses and substantial within-bunch heterogeneity. These insights provide a basis for targeted microclimate management to optimize grape quality under variable and warming climates.

 

7 Model Validation and Optimization

7.1 Model accuracy evaluation metrics

Model performance in predicting soil or degradation‑related properties is generally evaluated using error‑based and correlation‑based indices. Common error metrics include root mean square error (RMSE) and mean absolute error (MAE), which quantify average deviation between predicted and measured values for compaction parameters, soil organic carbon, or contaminants (Matinfar et al., 2021). Lower RMSE or MAE values are interpreted as higher predictive accuracy, and are often complemented by relative indicators such as mean absolute percentage error (MAPE) or weighted MAPE to account for scale differences among samples (Khatti and Grover, 2023). 

 

Correlation metrics such as the coefficient of determination (R²) and Pearson correlation coefficient (r) are used to assess how well the model captures overall variation patterns rather than absolute error. R² close to 1.0 indicates strong agreement, while additional indices like variance accounted for (VAF), concordance coefficient, and index of agreement (IOA) provide more nuanced views of model consistency across ranges of soil properties (Khatti and Grover, 2023; Safaee et al., 2024). In spatial applications, these metrics are typically computed on independent validation sets or through K‑fold cross‑validation to avoid optimistic bias. 

 

7.2 Comparative analysis of different models

Comparative studies show that advanced machine‑learning models often outperform simpler statistical approaches for predicting soil parameters relevant to degradation. For compaction behavior in fine‑grained or expansive soils, Gaussian process regression, support vector machines, and XGBoost achieved higher R² and lower RMSE than alternative standalone models and k‑nearest neighbors, indicating better capture of complex, nonlinear relationships (Khatti and Grover, 2023; Almuaythir et al., 2025). Similar patterns appear in digital soil mapping, where random forest and hybrid RF-ordinary kriging models exceed partial least squares regression and simple kriging in predicting soil organic carbon. 

 

In larger‑scale digital soil mapping and property prediction, model ranking is site‑ and variable‑dependent, but tree‑based ensembles generally perform better than ordinary kriging alone. For some properties such as cation exchange capacity, random forest yielded substantially higher R² than kriging, whereas all models performed poorly for highly variable soil organic matter, underscoring the role of property characteristics in model selection (Safaee et al., 2024). Comparative erosion‑risk mapping also found support vector machines and random forests both effective, with only small differences in overall accuracy and F1‑scores across classes (Fernández et al., 2023). 

 

7.3 Parameter optimization and generalization improvement

Hyperparameter tuning is critical to improve accuracy and generalization, especially for nonlinear models used in soil and degradation studies. For support vector regression applied to mid‑infrared soil spectroscopy, systematic optimization of epsilon and cost based on validation RMSE maintained or improved prediction accuracy across diverse soil datasets, and optimal settings differed by property and region (Deiss et al., 2020). Grid search and K‑fold cross‑validation are also used to tune random forest parameters before integrating them with kriging, leading to minimized RMSE and maximized R² in hybrid RF-OK contamination maps (Han and Suh, 2024). 

 

More automated strategies further enhance generalization. Bayesian optimization within automated machine‑learning frameworks efficiently explores hyperparameter spaces for ensembles, SVMs, and neural networks, selecting models with minimal cross‑validation error while checking against overfitting using separate RMSE, MAE, and error‑distribution analyses (Shah et al., 2024). In addition, hybrid ML-geostatistical approaches and the inclusion of geomorphometric and remote‑sensing covariates increase model robustness across heterogeneous croplands, allowing field‑scale prediction of soil organic carbon with high R² and low RMSE (Matinfar et al., 2021). 

 

Across soil and degradation‑related modeling, accuracy is evaluated mainly by RMSE/MAE and R²‑type metrics, often supplemented by agreement indices. Comparative analyses consistently show that well‑tuned machine‑learning and hybrid models outperform simpler baselines, though optimal algorithms depend on property and site. Careful hyperparameter optimization, cross‑validation, and enriched covariate sets are central for improving model generalization in soils under intensive or degrading land use.

 

8 Case Study: Analysis of the Relationship Between Vineyard Microclimate and Grape Quality

8.1 Study area and experimental design

The study area can be represented by intensive greenhouse cucumber systems typical of northern China, where long-term monoculture has already caused pronounced soil degradation. In one representative case, experiments were conducted for multiple years in solar greenhouses of the Loess Plateau region, with winter-spring cucumber grown repeatedly on typical sandy soils under a semi-arid climate and loess-derived topography (Liang et al., 2013). In another key site along the Yellow River irrigation area, cucumber had been cultivated continuously for 0, 4, 8, and 12 years, providing a clear gradient of degradation intensity under facility horticulture (Huang et al., 2023). 

 

Experimental designs combined re‑cropping duration and cropping system structure to capture both temporal and management-driven changes in soil quality. In Yan’an, treatments with 0, 1, 4, and 8 years of continuous winter-spring cucumber were established in replicated plots, and a second experiment compared seven rotations of cucumber with greengrocery, legumes, tomato, and fallow during the summer period (Figure 4) (Liang et al., 2013). In the Yellow River irrigation area, replicated plots across the 0-12 year gradient were sampled for physicochemical properties, enzyme activities and high‑throughput sequencing of rhizosphere microbial communities, enabling the model to link degradation indices to microbial shifts. 

 

Figure 4 Representative view of intensive greenhouse cucumber production systems in northern China. Long-term monoculture under protected cultivation conditions has been associated with progressive soil degradation (Adopted from Liang et al., 2013)

 

8.2 Model application and results analysis

The degradation model was applied to these datasets to quantify how continuous cucumber cropping alters soil salinity, acidity, nutrient status and microbial structure. Along the Yellow River gradient, model outputs reproduced the observed increase in total salt content and significant pH decline with increasing years of continuous cropping, while capturing transient rises in organic matter and available N, P, and K after 4-8 years. In the greenhouse re‑cropping experiment, simulated yield trajectories matched the empirical pattern of initial yield increase followed by a marked decline after about five consecutive cucumber crops, indicating that modeled degradation thresholds correspond to real productivity losses (Liang et al., 2013). 

 

Model diagnostics also reflected biological degradation, as predicted indices of microbial imbalance increased with cropping duration and were consistent with observed declines in bacterial and fungal diversity at 4 and 12 years (Huang et al., 2023). Sensitivity analyses highlighted pH and salt accumulation as dominant drivers of shifts in microbial community structure, aligning with redundancy analysis that identified pH as the main factor shaping microbial dominance in the rhizosphere. In the long‑term greenhouse system, modeled declines in soil biological function were further supported by independent evidence of reduced beneficial bacterial taxa and weakened co‑occurrence networks under prolonged monoculture (Liu et al., 2020). 

 

8.3 Optimization strategies for vineyard management

Although the core focus is continuous cucumber, insights from vineyard soil management provide useful optimization strategies where similar degradation processes occur. In European vineyards, paired comparisons of inter-row cover crops versus bare or intensively tilled soil show that permanent or low‑intensity cover cropping generally increases soil organic carbon and improves bulk density, percolation stability and hydraulic conductivity, except in severely water‑limited sites (Liebhard et al., 2024). These results indicate that maintaining vegetation cover and reducing tillage intensity can mitigate compaction, erosion and organic matter loss, which are also critical issues under protected horticulture and continuous vegetable systems. 

 

Additional vineyard experiments demonstrate that grass strips without NPK fertilization sustain favorable sorption capacity and an optimal nutrient regime, while high mineral fertilization rates reduce pH and sorption capacity and increase the risk of hazardous element mobility, especially Cu from fungicides (Šimanský et al., 2023). Physical modeling of runoff and erosion in steep-slope vineyards further shows that continuous tillage of inter‑rows exacerbates soil erosion, whereas single annual tillage or nectariferous covers markedly reduce sediment losses and highlight the negative role of wheel tracks as runoff pathways (Straffelini et al., 2022). Together, these findings support optimization strategies that prioritize cover crops, reduced tillage and balanced fertilization when adapting soil management to prevent degradation, whether in vineyards or intensively cropped cucumber greenhouses. 

 

The model was grounded in long‑term greenhouse cucumber experiments spanning gradients of re‑cropping years and diversified rotations, and its outputs aligned with observed changes in salinity, pH, nutrients, yield and microbial communities. Evidence from vineyard soil management underscores that cover cropping, minimized tillage and moderated fertilization are key levers to reduce erosion, maintain organic matter and stabilize nutrient and contaminant dynamics, offering transferable principles for optimizing soil degradation control under continuous cucumber cropping.

 

9 Conclusions and Future Perspectives

Long‑term continuous cucumber cropping consistently altered soil physicochemical properties and nutrient patterns. Across greenhouse systems, extended monoculture lowered soil pH and increased electrical conductivity and total salt content, while organic matter and available N, P and K initially accumulated before declining or becoming less available to plants. Phosphorus fractions shifted toward larger pools of total, labile and non‑labile P, with a marked rise in total P and a decline in the phosphorus activation coefficient as cropping rounds progressed. Soil biological communities showed clear degradation signals, including reduced microbial diversity and restructuring of bacterial and fungal assemblages. With increasing years of continuous cucumber cultivation, beneficial bacterial genera involved in nutrient cycling and pathogen suppression (e.g. Bacillus, Sphingomonas) and beneficial fungi (e.g. Chaetomium, Mortierella, Penicillium) declined, while dominant phyla such as Ascomycota and several potentially harmful groups became more abundant. Co‑occurrence network analyses in related cucumber and greenhouse vegetable systems indicated weakened microbial interactions and simpler networks, suggesting reduced soil ecological stability under prolonged monoculture. 

 

These findings support a systems view of continuous cropping obstacles in which soil degradation arises from coupled shifts in chemistry, microbial communities and rhizosphere metabolism. Reviews of continuous cropping across crops emphasize that changes in soil ecological environment, rather than nutrients alone, are central drivers of yield decline, disease increase and quality loss. Phase‑change patterns described for strawberry and other crops-physicochemical imbalance, followed by biotic community disruption, and finally allelochemical accumulation-fit well with the progressive changes observed in long‑term cucumber systems. Practically, the evidence highlights that continuous cucumber cropping is unsustainable without active soil regulation and diversification measures. Rotation and cover‑crop systems, including paddy‑upland and garlic rotations or leafy‑vegetable cover crops in cucumber fallow periods, improved nutrient cycling, raised pH, enriched beneficial microbiota and enhanced cucumber yield, demonstrating effective pathways to reverse degradation. Amendments such as biochar‑based fertilizers and vegetable‑residue incorporation also alleviated continuous‑cropping obstacles by improving soil fertility, buffering capacity and microbial structure, offering scalable management options for intensive greenhouse cucumber production. 

 

Future work should further resolve the mechanisms linking nutrient accumulation, especially of N and P, with microbial community restructuring and functional loss in continuous cucumber soils. Studies on tomato and other greenhouse monocultures show that long‑term enrichment of organic matter and nutrients can drive a shift from bacterial to fungal dominance and down‑regulate C‑ and N‑cycling genes, suggesting similar functional tipping points may exist in cucumber systems that remain poorly quantified. Integrating high‑throughput sequencing with metabolomics and functional gene profiling will help clarify how specific microbial guilds and metabolites mediate soil degradation and plant health responses over time. In addition, more comparative and long‑term trials are needed to evaluate mitigation strategies under diverse environmental and management contexts. Meta‑analyses indicate that not all continuous‑cropping systems develop severe obstacles and that disease‑suppressive, microbially “self‑healing” soils can emerge, but the conditions favoring these outcomes in cucumber remain unclear. Systematic testing of rotations, intercropping, biochar and residue‑return regimes, combined with soil‑quality indices and network analyses, could identify design principles for resilient cucumber production systems and guide region‑specific recommendations for sustainable soil management.

 

Acknowledgments

I extend our sincere gratitude to the anonymous reviewers for their valuable and insightful comments, which have greatly strengthened this paper.

 

Conflict of Interest Disclosure

The author affirms that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

 

References

Akhtar M.N., Shaikh A.J., Khan A., Awais H., Bakar E.A., and Othman A.R., 2021, Smart sensing with edge computing in precision agriculture for soil assessment and heavy metal monitoring: a review, Agriculture, 11(6): 475.

https://doi.org/10.3390/agriculture11060475

 

Almuaythir S., Zaini M.S.I., and Lodhi R.H., 2025, Predicting soil compaction parameters in expansive soils using advanced machine learning models: a comparative study, Scientific Reports, 15(1): 24018.

https://doi.org/10.1038/s41598-025-09279-2

 

Alsalami Z., Mandapati A.H.A.H., and Sundari V.R., 2025, Modeling the impact of climate change on soil health using predictive analytics, SHS Web of Conferences, 216: 01053.

https://doi.org/10.1051/shsconf/202521601053

 

Arrizabalaga M., Morales F., Oyarzun M., Delrot S., Gomès E., Irigoyen J.J., Hilbert G., and Pascual I., 2018, Tempranillo clones differ in the response of berry sugar and anthocyanin accumulation to elevated temperature, Plant Science, 267: 74-83.

https://doi.org/10.1016/j.plantsci.2017.11.009

 

Bandaru L., Irigireddy B.C., Koutilya P.V.N.R., and Davis B., 2024, DeepQC: A deep learning system for automatic quality control of in-situ soil moisture sensor time series data, Smart Agricultural Technology, 9: 100514.

https://doi.org/10.1016/j.atech.2024.100514

 

Beauchet S., Cariou V., Renaud-Gentié C., Meunier M., Siret R., Thiollet-Scholtus M., and Jourjon F., 2020, Modeling grape quality by multivariate analysis of viticulture practices, soil and climate, OENO One, 54(3): 601-622.

https://doi.org/10.20870/oeno-one....1067

 

Blancquaert E.H., Oberholster A., Ricardo-da-Silva J.M., and Deloire A.J., 2019, Grape flavonoid evolution and composition under altered light and temperature conditions in Cabernet Sauvignon (Vitis vinifera L.), Frontiers in Plant Science, 10: 1062.

https://doi.org/10.3389/fpls.2019.01062

 

Buljeta I., Pichler A., Šimunović J., and Kopjar M., 2023, Beneficial effects of red wine polyphenols on human health: Comprehensive review, Current Issues in Molecular Biology, 45(2): 782-798.

https://doi.org/10.3390/cimb45020052

 

Burdett H., and Wellen C., 2022, Statistical and machine learning methods for crop yield prediction in the context of precision agriculture, Precision Agriculture, 23: 1553-1574.

https://doi.org/10.1007/s11119-022-09897-0

 

Candar S., Korkutal İ., and Bahar E., 2019, Effect of canopy microclimate on Merlot (Vitis vinifera L.) grape composition, Applied Ecology and Environmental Research, 17(6): 15431-15446.

https://doi.org/10.15666/aeer/1706_1543115446

 

Carroll J.E., and Wilcox W.F., 2003, Effects of humidity on the development of grapevine powdery mildew, Phytopathology, 93(9): 1137-1144.

https://doi.org/10.1094/phyto.2003.93.9.1137

 

Chamara N., Islam M.D., Bai G., Shi Y., and Ge Y., 2022, Ag-IoT for crop and environment monitoring: Past, present, and future, Agricultural Systems, 203: 103497.

https://doi.org/10.1016/j.agsy.2022.103497

 

Chavarria G., Cardoso L.S., Bergamaschi H., Santos H.P., Mandelli F., and Marodin G.A.B., 2009, Microclima de vinhedos sob cultivo protegido, Ciência Rural, 39(7): 2029-2034.

https://doi.org/10.1590/s0103-84782009005000147

 

Crouchett-Rojas R., Araya-Alman M., Verdugo-Vásquez N., Fourment M., and Gutiérrez-Gamboa G., 2025, Shading nets: A current viticultural strategy to mitigate the negative impacts of global warming on grape and wine quality, Australian Journal of Grape and Wine Research, 2025: 9729885.

https://doi.org/10.1155/ajgw/9729885

 

De Rosas I., Deis L., Baldo Y., Cavagnaro J.B., and Cavagnaro P.F., 2022, High temperature alters anthocyanin concentration and composition in grape berries of Malbec, Merlot, and Pinot Noir in a cultivar-dependent manner, Plants, 11(7): 926.

https://doi.org/10.3390/plants11070926

 

Deiss L., Margenot A.J., Culman S.W., and Demyan M.S., 2020, Tuning support vector machines regression models improves prediction accuracy of soil properties in MIR spectroscopy, Geoderma, 365: 114227.

https://doi.org/10.1016/j.geoderma.2020.114227

 

Dorigo W.A., Xaver A., Vreugdenhil M., Gruber A., Hegyiová A., Sanchis-Dufau A.D., Zamojski D., Cordes C., Wagner W., and Drusch M., 2013, Global automated quality control of in situ soil moisture data from the International Soil Moisture Network, Vadose Zone Journal, 12(3): vzj2012.0097.

https://doi.org/10.2136/vzj2012.0097

 

Fan Y., Wang X., Funk T., Rashid I., Herman B., Bompoti N., Mahmud M., Chrysochoou M., Yang M., Vadas T., Lei Y., and Li B., 2022, A critical review for real-time continuous soil monitoring: Advantages, challenges, and perspectives, Environmental Science and Technology, 56(19): 13546-13564.

https://doi.org/10.1021/acs.est.2c03562

 

Fernández D., Adermann E., Pizzolato M., Pechenkin R., Rodríguez C., and Taravat A., 2023, Comparative analysis of machine learning algorithms for soil erosion modelling based on remotely sensed data, Remote Sensing, 15(2): 482.

https://doi.org/10.3390/rs15020482

 

Flamini R., Mattivi F., De Rosso M., Arapitsas P., and Bavaresco L., 2013, Advanced knowledge of three important classes of grape phenolics: Anthocyanins, stilbenes and flavonols, International Journal of Molecular Sciences, 14(10): 19651-19669.

https://doi.org/10.3390/ijms141019651

 

Gao Y., Lu X., Guo R., Hao J., Miao Z., Yang L., and Li S., 2021, Responses of soil abiotic properties and microbial community structure to 25-year cucumber monoculture in commercial greenhouses, Agriculture, 11(4): 341.

https://doi.org/10.3390/agriculture11040341

 

Gouot J.C., Smith J.P., Holzapfel B.P., Walker A.R., and Barril C., 2018, Grape berry flavonoids: A review of their biochemical responses to high and extreme high temperatures, Journal of Experimental Botany, 70(2): 397-423.

https://doi.org/10.1093/jxb/ery392

 

Han H., and Suh J., 2024, Spatial prediction of soil contaminants using a hybrid random forest-ordinary kriging model, Applied Sciences, 14(4): 1666.

https://doi.org/10.3390/app14041666

 

Hegedüs I., Andreidesz K., Szentpéteri J., Kaleta Z., Szabo L., Szigeti K., Gulyás B., Padmanabhan P., Budán F., and Máthé D., 2022, The utilization of physiologically active molecular components of grape seeds and grape marc, International Journal of Molecular Sciences, 23(19): 11165.

https://doi.org/10.3390/ijms231911165

 

Hu J., Bai S., Zhao R., Chen G., and Cai J., 2022, Effects of black geotextile mulch and grass mulch on the microclimate, fruit quality and anthocyanin components of ‘Xinyu’ table grape, New Zealand Journal of Crop and Horticultural Science, 52: 141-158.

https://doi.org/10.1080/01140671.2022.2108066

 

Huang S., Yu J., Hou D., Yue H., Zhang D., Li Y., Lyu J., Jin L., and Jin N., 2023, Response of soil microbial community diversity to continuous cucumber cropping in facilities along the Yellow River irrigation area, PLOS ONE, 18(8): e0289772.

https://doi.org/10.1371/journal.pone.0289772

 

Huang Y., Guenet B., Wang Y., and Ciais P., 2021, Global simulation and evaluation of soil organic matter and microbial carbon and nitrogen stocks using the microbial decomposition model ORCHIMIC v2.0, Global Biogeochemical Cycles, 35(4): e2020GB006836.

https://doi.org/10.1029/2020gb006836

 

Hunter J.J., Volschenk C.G., Mania E., Vicente Castro A., Booyse M., Guidoni S., Pisciotta A., Di Lorenzo R., Novello V., and Zorer R., 2021, Grapevine row orientation mediated temporal and cumulative microclimatic effects on grape berry temperature and composition, Agricultural and Forest Meteorology, 310: 108660.

https://doi.org/10.1016/j.agrformet.2021.108660

 

Kumar S., Singh I., and Yadav A., 2025, Determination of maturity standards of some grape (Vitis vinifera L.) cultivars under Ayodhya conditions, International Journal of Agricultural Sciences, 21(1): 68-72.

https://doi.org/10.15740/has/ijas/21.1/68-72

 

Kunter B., Unal O., Keskin S., Hatterman-Valenti H., and Kaya O., 2024, Comparison of the sugar and organic acid components of seventeen table grape varieties produced in Ankara (Türkiye): A study over two consecutive seasons, Frontiers in Plant Science, 15: 1321210.

https://doi.org/10.3389/fpls.2024.1321210

 

Lecourieux F., Kappel C., Pieri P., Charon J., Pillet J., Hilbert G., Renaud C., Gomès E., Delrot S., and Lecourieux D., 2017, Dissecting the biochemical and transcriptomic effects of a locally applied heat treatment on developing Cabernet Sauvignon grape berries, Frontiers in Plant Science, 8: 53.

https://doi.org/10.3389/fpls.2017.00053

 

Lee S., Chu M.L., Guzman J.A., and Flanagan D.C., 2022, Modeling soil erodibility and critical shear stress parameters for soil loss estimation, Soil and Tillage Research, 218: 105292.

https://doi.org/10.1016/j.still.2021.105292

 

Liang Y., Liang Y., Lin X., Yamada S., Inoue M., and Inosako K., 2013, Soil degradation and prevention in greenhouse production, SpringerPlus, 2(S1): S10.

https://doi.org/10.1186/2193-1801-2-s1-s10

 

Liebhard G., Guzmán G., Gómez J.A., Winter S., Zaller J.G., Bauer T., Nicolai A., Cluzeau D., Popescu D., Bunea C., and Strauss P., 2024, Vineyard cover crop management strategies and their effect on soil properties across Europe, European Journal of Soil Science, 75: e13573.

https://doi.org/10.1111/ejss.13573

 

Liu J., Yang K., Tariq A., Lu L., Soufan W., and El Sabagh A., 2023, Interaction of climate, topography and soil properties with cropland and cropping pattern using remote sensing data and machine learning methods, The Egyptian Journal of Remote Sensing and Space Science, 26: 579-590.

https://doi.org/10.1016/j.ejrs.2023.05.005

 

Liu X., Li Y., Ren X., Chen B., Zhang Y., Shen C., Wang F., and Wu D., 2020, Long-term greenhouse cucumber production alters soil bacterial community structure, Journal of Soil Science and Plant Nutrition, 20(1): 306-321.

https://doi.org/10.1007/s42729-019-00109-9

 

Lu L., Delrot S., and Liang Z., 2024, From acidity to sweetness: A comprehensive review of carbon accumulation in grape berries, Molecular Horticulture, 4(1): 100.

https://doi.org/10.1186/s43897-024-00100-8

 

Maoz I., Kaplunov T., Raban E., Dynkin I., Degani O., Lewinsohn E., and Lichter A., 2020, Insights on the chemosensory basis of flavor in table grapes, Journal of the Science of Food and Agriculture, 100(4): 1405-1417.

https://doi.org/10.1002/jsfa.10102

 

Martínez-Lüscher J., Sánchez-Díaz M., Delrot S., Aguirreolea J., Pascual I., and Gomès E., 2016, Ultraviolet-B alleviates the uncoupling effect of elevated CO2 and increased temperature on grape berry (Vitis vinifera cv. Tempranillo) anthocyanin and sugar accumulation, Australian Journal of Grape and Wine Research, 22(1): 87-95.

https://doi.org/10.1111/ajgw.12213

 

Matinfar H.R., Maghsodi Z., Mousavi S.R., and Rahmani A., 2021, Evaluation and prediction of topsoil organic carbon using machine learning and hybrid models at a field-scale, CATENA, 202: 105258.

https://doi.org/10.1016/j.catena.2021.105258

 

Moriyama K., Kono A., Matsuzaki R., Azuma A., Onoue N., Sekozawa Y., Sato A., and Sugaya S., 2024, Diversity of flavour characteristics of table grapes and their contributing volatile compounds analysed by the solvent-assisted flavour evaporation method, Horticulture Research, 11(4): uhae048.

https://doi.org/10.1093/hr/uhae048

 

Müller K., Keller M., Stoll M., and Friedel M., 2023, Wind speed, sun exposure and water status alter sunburn susceptibility of grape berries, Frontiers in Plant Science, 14: 1145274.

https://doi.org/10.3389/fpls.2023.1145274

 

Odedeyi T., Kolawole O., Ugoji C., Ayankanmi T., Okhuoya O., Rabbit I., Asante K., and Darwazeh I., 2025, Multi-source data integration and IoT-based sensing for crop yield modelling and optimisation, in 2025 IEEE International Conference on Smart Computing (SMARTCOMP), pp.294-299.

https://doi.org/10.1109/smartcomp65954.2025.00094

 

Ohana-Levi N., Cohen Y., Munitz S., Michaelovsky R., Mintz D., Hagag N., Getz Y., and Netzer Y., 2024, The response of yield, number of clusters, and cluster weight to meteorological factors and irrigation practices in grapevines: A multi-experiment study, Scientia Horticulturae, 325: 112761.

https://doi.org/10.1016/j.scienta.2023.112761

 

Pérémé M., Haddon A., Steyer J.-P., and Jimenez J., 2023, SoilFract: A mechanistic model accounting for the fate of exogenous organic matter in soil carbon and nitrogen cycles, Waste Management, 159: 63-74.

https://doi.org/10.1016/j.wasman.2023.01.015

 

Pervaiz Z.H., Iqbal J., Zhang Q., Chen D., Wei H., and Saleem M., 2020, Continuous cropping alters multiple biotic and abiotic indicators of soil health, Soil Systems, 4(4): 59.

https://doi.org/10.3390/soilsystems4040059

 

Pieri P., Zott K., Gomès E., and Hilbert G., 2016, Nested effects of berry half, berry and bunch microclimate on biochemical composition in grape, OENO One, 50(3): 23-33.

https://doi.org/10.20870/oeno-one.2016.50.3.52

 

Reshef N., Walbaum N., Agam N., and Fait A., 2017, Sunlight modulates fruit metabolic profile and shapes the spatial pattern of compound accumulation within the grape cluster, Frontiers in Plant Science, 8: 70.

https://doi.org/10.3389/fpls.2017.00070

 

Sadasivan M., Krithika K., Shinty P., and Deepa A., 2025, Predictive crop yield modeling and soil quality classification using machine learning, Indian Journal of Science and Technology, 18(32): 2875-2883.

https://doi.org/10.17485/ijst/v18i32.1135

 

Safaee S., Libohova Z., Kladivko E.J., Brown A., Winzeler E., Read Q.D., Rahmani S., and Adhikari K., 2024, Influence of sample size, model selection, and land use on prediction accuracy of soil properties, Geoderma Regional, 37: e00766.

https://doi.org/10.1016/j.geodrs.2024.e00766

 

Schmidt J.C., and Kerkez B., 2023, Machine learning-assisted, process-based quality control for detecting compromised environmental sensors, Environmental Science and Technology, 57(13): 5305-5314.

https://doi.org/10.1021/acs.est.3c00360

 

Selvanarayanan R., Rajendran S., Algburi S., Khalaf O.I., and Hamam H., 2024, Empowering coffee farming using counterfactual recommendation based RNN driven IoT integrated soil quality command system, Scientific Reports, 14(1): 56954.

https://doi.org/10.1038/s41598-024-56954-x

 

Shah A., Thaker T., Shukla V., and Ranpura P., 2024, Efficient predictive modeling of resilient modulus in stabilized clayey soil using automated machine learning, Construction and Building Materials, 447: 137678.

https://doi.org/10.1016/j.conbuildmat.2024.137678

 

Šimanský V., Wójcik-Gront E., Jonczak J., and Horák J., 2023, Optimizing soil management for sustainable viticulture: Insights from a Rendzic Leptosol vineyard in the Nitra wine region, Slovakia, Agronomy, 13(12): 3042.

https://doi.org/10.3390/agronomy13123042

 

Straffelini E., Pijl A., Otto S., Marchesini E., Pitacco A., and Tarolli P., 2022, A high-resolution physical modelling approach to assess runoff and soil erosion in vineyards under different soil managements, Soil and Tillage Research, 223: 105418.

https://doi.org/10.1016/j.still.2022.105418

 

Sun K., Fu L., Song Y., Yuan L., Zhang H., Wen D., Yang N., Wang X., Yue Y., Li X., and Wang K., 2021, Effects of continuous cucumber cropping on crop quality and soil fungal community, Environmental Monitoring and Assessment, 193(7): 436.

https://doi.org/10.1007/s10661-021-09136-5

 

Wang H., Yang M., Martinez-Luscher J., Hilbert-Masson G., Gomès E., Pascual I., Fan P., Kong J., Liang Z., Xu Z., Wang Y., and Dai Z., 2025, Effects of elevated temperature and shaded UV-B radiation exclusion on berry biochemical composition in four grape (Vitis vinifera L.) cultivars with distinct anthocyanin profiles, Food Research International, 218: 116823.

https://doi.org/10.1016/j.foodres.2025.116823

 

Wang S., Yan H., Wang L., Chen X., Wang J., and He F., 2025, Focus on flavors: Physico-chemical, primary metabolite, volatile, and sensory analysis of seven white table grape cultivars in China, Food Chemistry: X, 31: 103157.

https://doi.org/10.1016/j.fochx.2025.103157

 

Wassay M., Khalid B., Ashraf M., Riaz T., Khan N., and Maqbool R., 2026, Geo-intelligent agriculture: Integrating GIS, remote sensing, and IoT for real-time soil and crop health monitoring and predictive farm management, Agricultural Research Reports, 3(2): 465.

https://doi.org/10.54219/arr.03.2.2025.465

 

Wilson A., Ferrandino A., Giacosa S., Novello V., and Guidoni S., 2024, The effect of temperature and UV manipulation on anthocyanins, flavonols, and hydroxycinnamoyl-tartrates in cv Nebbiolo grapes (Vitis vinifera L.), Plants, 13(22): 3158.

https://doi.org/10.3390/plants13223158

 

Wood A., Soto S., Gambetta G., Schober D., Jeffers E., and Coulson T., 2024, Modelling the climate changing concentrations of key red wine grape quality molecules using a flexible modelling approach, OENO One, 58(4): 1-22.

https://doi.org/10.20870/oeno-one.2024.58.4.8086

 

Wu Y., Yang Z., and Liu Y., 2023, Internet-of-Things-based multiple-sensor monitoring system for soil information diagnosis using a smartphone, Micromachines, 14(7): 1395.

https://doi.org/10.3390/mi14071395

 

Zha Q., Zhong H., Tang M., Yin X., Sun P., Jiang A., Xi X., and Wu J., 2024, Effects of temperature and light during the veraison period on grape berry growth, Plant Stress, 12: 100642.

https://doi.org/10.1016/j.stress.2024.100642

 

Zhang D., Zhang Y., Lin K., Wang B., Shi X., and Cheng W., 2021, Comparison of sugars, organic acids and aroma components of five table grapes in Xinjiang, IOP Conference Series: Earth and Environmental Science, 792(1): 012029.

https://doi.org/10.1088/1755-1315/792/1/012029

 

Zhang K., Cao J., Yin H., Wang J., Wang X., Yang Y., and Xi Z., 2024, Microclimate diversity drives grape quality difference at high-altitude: Observation using PCA analysis and structural equation modeling (SEM), Food Research International, 191: 114644.

https://doi.org/10.1016/j.foodres.2024.114644

 

Zhang Y., Lavallee J.M., Robertson A.D., Even R., Ogle S.M., Paustian K., and Cotrufo M.F., 2021, Simulating measurable ecosystem carbon and nitrogen dynamics with the mechanistically defined MEMS 2.0 model, Biogeosciences, 18(10): 3147-3171.

https://doi.org/10.5194/bg-18-3147-2021

 

Zhao H., Li T., Zhang Y., Hu J., Bai Y., Shan Y., and Ke F., 2017, Effects of vermicompost amendment as a basal fertilizer on soil properties and cucumber yield and quality under continuous cropping conditions in a greenhouse, Journal of Soils and Sediments, 17(11): 2718-2730.

https://doi.org/10.1007/s11368-017-1744-y

 

Zhao Y., Mao X., Zhang M., Yang W., Di H.J., Ma L., Liu W., and Li B., 2020, Response of soil microbial communities to continuously mono-cropped cucumber under greenhouse conditions in a calcareous soil of north China, Journal of Soils and Sediments, 20(5): 2446-2459.

https://doi.org/10.1007/s11368-020-02603-5

 

Zhou D., Li J., Xiong R., Saimaiti A., Huang S., Wu S., Yang Z., Shang A., Zhao C., Gan R.-Y., and Li H.-B., 2022, Bioactive compounds, health benefits and food applications of grape, Foods, 11(18): 2755.

https://doi.org/10.3390/foods11182755