Evaluation of Winter Hydrology Performance of Three Field-Scale Models

Abstract

Highlights EPIC, SHAW, and DRAINMOD models were evaluated for the simulation of winter hydrology. Energy-based models can better simulate late-winter and early-spring hydrology under winter conditions. Effective simulation of soil temperature and soil hydraulics in winters were identified as potential areas of development in temperature-based models. Abstract. The deterioration of Lake Erie's water quality is one of the major concerns in North America. A considerable percentage of annual phosphorus runoff occurs during the non-growing season in cold agricultural regions such as those in the Great Lakes region. Consequently, without accurate simulation of water flow during cold periods, reliable modeling of sediment and nutrient loads to surface water bodies is not achievable. Three hydrological models (EPIC, SHAW, and DRAINMOD) were evaluated for their capacity to predict winter tile flow and to highlight the significant processes that have a larger effect on runoff simulation at a field site in Southern Ontario, Canada. The SHAW model adequately predicted both soil temperature at 10 cm depth (R2 = 0.95; 2013-2014) and winter tile flow (2012-2014, Nov-Apr; R2 = 0.52; PBIAS = 7; NSE = 0.49). In the case of tile flow, DRAINMOD exhibited comparable results to the SHAW model for the same period (R2 =0.55, PBIAS = -28, NSE = 0.58). EPIC was not able to perform satisfactorily in simulating the tile flow during winter conditions, which was attributed to the model’s erroneous prediction of soil temperature from air temperature. It was determined that energy-based models like DRAINMOD and SHAW can better simulate late-winter and early-spring hydrological conditions. Keywords: Agricultural runoff, Canadian winter hydrology, Hydrological models, Soil temperature, Tile flow.