Process-based modeling of temperature and water profiles in the seedling recruitment zone: Part I. Model validation

Process-based modeling provides greater spatial and temporal information of the soil environment in the shallow seedling recruitment zone across field topography where measurements of soil temperature and water may not sufficiently describe the zone. Hourly temperature and water profiles within the 75 mm recruitment zone for 75 days after seeding were simulated for Canadian Prairie conditions from the process-based Simultaneous Heat and Water (SHAW) model using local and non-local microclimatic data. Measured and modeled soil cover and spring wheat vegetative cover were used to parameterize the model. Heat and water transfer was simulated through surface residue, early vegetation and soil. Simulations were evaluated using model efficiency, root mean square deviation, and components of mean squared error. The greatest amount of error in simulated soil temperature was lack of correlation in the fluctuation pattern over time, followed by bias of the simulation. Soil temperature simulations had model efficiency of 0.87, overestimation of 0.4 degrees C, and a RMSD of 2.1 degrees C averaged across all topographical factors and soil depths. Simulations of soil water had low model efficiency and RMSD of 0.55 MPa. Average absolute bias for soil water was 0.27 MPa which reflected predominantly positive bias at the soil surface and 0-25 mm soil layer and negative bias in the 25-50 and 50-75 mm soil layers. Process-based modeling using microclimatic information was shown to provide representative simulations of the soil environment for all depths of the seedling recruitment zone. (C) 2013 Elsevier B.V. All rights reserved.