A Framework for Soil Health Modeling Under Climate Extremes

Thesis Abstract

Soil health is increasingly threatened by climate extremes, which disrupt soil biological activity, nutrient cycling, structure, and water retention, thereby compromising crop productivity and ecosystem resilience. Despite advances in soil health indicators, there remains a gap in integrated frameworks that model soil health dynamics under simultaneous drought, heat waves, and intense rainfall events using process-based and empirical approaches. This study aims to develop a comprehensive framework for soil health modeling under climate extremes that integrates soil physical, chemical, and biological indicators with climate variability to predict resilience and functional performance of agroecosystems. Specific objectives are (i) to identify and quantify the key soil health indicators sensitive to climate extremes across contrasting soil types and land-use practices; (ii) to develop a modular modeling framework that couples soil hydraulic properties, organic matter dynamics, microbial functional profiles, and microbial–plant–soil feedbacks under projected extreme weather scenarios; (iii) to calibrate and validate the framework using multi-site field data and controlled-environment experiments; (iv) to evaluate scenario-based outputs for different management interventions (cover cropping, reduced tillage, organic amendments) on soil health trajectories; and (v) to provide decision-support metrics for farmers and policy makers that link soil health indicators to yield stability and environmental sustainability. The methodology adopts an explanatory sequential mixed-methods design. The population comprises temperate and semi-arid agroecosystems with representative soils (sandy loam to silty clay) and perennial and annual cropping systems. A stratified random sample of 60 field plots across four sites will be used, with 15 plots per site, augmented by controlled microcosm experiments (n=24) to simulate extreme events under standardized rainfall and temperature regimes. Data collection instruments include soil physico-chemical analyses (texture, bulk density, total organic carbon, permanganate oxidizable carbon, cation exchange capacity, micronutrient pools), soil hydrological measurements (infiltration rate, water holding capacity), microbial community profiling via 16S/ITS amplicon sequencing, enzymatic activity assays (?-glucosidase, dehydrogenase, phosphatase), and plant–soil interaction indicators (plant biomass, root length density, mycorrhizal colonization). Climate inputs will be derived from downscaled regional climate models for representative extreme scenarios (RCP/SSP-based) to drive the model. Data analysis will combine (i) structural equation modeling to identify causal pathways among soil health indicators under extreme events, (ii) machine learning approaches (random forest, gradient boosting) to detect nonlinear responses and interactions, and (iii) process-based modeling to couple hydraulic, carbon, and nutrient cycles with microbial functional dynamics. Model testing will employ cross-site validation, with performance metrics including RMSE, R-squared, and information criteria, and sensitivity analyses will ascertain the influence of key parameters. Ethical considerations adhere to best practices in soil sampling and data management. Expected findings include robust identification of indicator sets most responsive to drought, heat, and heavy rainfall, and the development of a modular framework capable of integrating soil physics, chemistry, and biology with climate drivers. It is anticipated that the framework will reveal threshold behaviors in soil structure and microbial activity that precede declines in nutrient availability and plant growth, as well as the conditions under which management practices mitigate adverse effects. The study is expected to demonstrate that cover crops and organic amendments maintain higher microbial functional diversity and enzyme activities under extremes, translating into improved infiltration, water retention, and yield stability. The contribution to knowledge lies in delivering a validated, scalable soil health modeling framework that explicitly accounts for climate extremes, bridging process-based understanding with empirical patterns, and offering quantifiable metrics for resilience assessment. The main conclusion will emphasize the utility of an integrated, modular soil health model as a decision-support tool for optimizing management under climate risk, while acknowledging uncertainties in climate projections and microbial responses. Recommendations include incorporating soil health modeling into adaptive management plans at regional and farm scales, prioritizing data collection on microbial functional genes and soil texture–structure interactions, and developing user-friendly interfaces for stakeholders to explore scenario-based outcomes and resilience indicators.