Simulated winter snow, soil thermal regime, and artificial drainage dynamics under climate change.

This dataset provides simulated winter subsurface drainage outputs from the calibrated Root Zone Water Quality Model–Simultaneous Heat and Water (RZ-SHAW) and four machine-learning models (Cubist, LSTM, Multilinear, SVM) for five agricultural research sites across Eastern Canada (Alfred, Harrow, Kentville, Ottawa-Greenbelt, and St-Emmanuel). The simulations span from 1950 to 2100 under the high-emission climate change scenario RCP8.5, derived from the Canadian Regional Climate Model Large Ensemble (CANRCM4 LE). Included in this repository are: Winter drainage volume's future projections from 1950 to 2100 the RZ-SHAW and machine-learning models. RZ-SHAW simulated future projections from 1950 to 2100 in snow cover, soil thermal conditions, subsurface drainage volume and frequency, evaporation, surface runoff, and soil water storage. This dataset also contains simulated annual crop yield projections (kg/hectare) from 1950 to 2100 the five sites Alfred (corn), Harrow (corn and soybean), Kentville (corn), Ottawa (corn), St-Emmanuel (corn). Simulations were performed using the RZ-SHAW model under RCP8.5 climate change scenario. Each site's data includes annual grain yield values and Sen's slope trend analysis with Mann-Kendall significance test results. Users should note the following important limitations when interpreting these yield projections: Simplified Crop Growth Parameters: The crop module was implemented primarily to account for crop residue effects on winter soil conditions, not for yield prediction. No calibration was performed for crop growth parameters. No Nutrient Stress: Simulations were run without nitrogen or phosphorus limitations for simplicity, which may overestimate yields in scenarios where nutrient availability would be constraining. Fixed Management Practices: Planting and harvest dates were held constant throughout the simulation period (1950-2100) based on historical averages, which becomes increasingly unrealistic under future climate conditions. No CO₂ Fertilization Effect: Atmospheric CO₂ concentration changes were not incorporated, potentially underestimating photosynthetic responses in future climate scenarios. Limited Stress Responses: The model's representation of heat stress, drought tolerance, and other physiological responses to extreme conditions may not fully capture crop responses under future climate scenarios. No Adaptation Strategies: The simulations do not account for potential adaptations such as cultivar changes, shifting planting dates, or irrigation implementation that would likely occur in response to changing conditions.