EPIC estimates of soil water, nitrogen and carbon under semiarid temperate conditions
The Environmental Policy Integrated Climate (EPIC) model is an important support tool for environmental management. Previous tests of the model have determined that it is suitable for long-term yield estimation, but it is less precise in assessing annual yield variability. To determine the reasons for the discrepancies between estimated and measured yields, we tested the ability of EPIC version 5300 to predict soil water and soil nitrogen dynamics, using data from a long-term spring wheat (Triticum aestivum L.) rotation experiment in the semiarid prairie region of Canada. Potential evapotranspiration (PET) estimates varied among methods tested: Priestley-Taylor and Penman-Monteith methods resulted in PET means that were about twice those obtained with the Hargreaves and Baier-Robertson methods. The higher PET means were associated with an excessive estimation of net radiation. We used the Baier-Robertson method to generate the other estimates reported herein. EPIC generally overestimated total soil water, but it still allowed clear differentiation among rotation phases and times of the year, and provided adequate estimates of water during the critical shot-blade stage. Water estimates by soil layer were also generally overpredicted, especially at depths from 0.15 to 0.60 m, but we were able to differentiate among rotation phases and times of the year. Precision of these latter estimates was generally low, accounting at most for 27% of the variability, and varied by soil layer, rotation phase and time of the year. Nitrate-N estimates tended to be lower than measured values, especially at depths below 0.3 m and during vegetative growth phases. However, the estimates also allowed us to distinguish among the rotation phases and times of the year. Total N and organic C were satisfactorily estimated by EPIC. In general, EPIC provided adequate long-term estimates of the environmental quality indicators tested. Key words: Environmental quality, environmental modelling, sustainability, spring wheat, fallow, potential evapotranspiration methods