Impact of Accurate Evapotranspiration Estimates on DRAINMOD Simulation in North Dakota

Abstract. Subsurface drainage is an essential practice for farmlands with shallow water table. An accurately designed subsurface drainage system can improve soil and water environment as well as the crop yield. Computer models, such as DRAINMOD, are often used to simulate the field hydrology or design the subsurface drainage systems under different scenarios. The accuracy of the model simulation is highly depending on the accuracy of the input parameters. As an important input parameter to DRAINMOD, evapotranspiration (ET) is estimated using temperature based reference ET method and a stress adjustment factor, while an inaccurate ET estimation can cause errors in DRAINMOD simulation. This paper will evaluate the effects of accurate ET estimation on DRAINMOD prediction through comparisons of various reference ET and actual ET estimates. Reference ET was calculated using data from a nearby weather station. Actual ET was measured by an eddy covariance system. Drainage and hydrology data, including water table, precipitation, and drainage flow were measured in the field in 2009 and used in DRAINMOD simulation. With different reference ET values, different monthly adjustment factors were obtained for each reference ET method in order to match the simulated with the measured water tables. After compared with the actual ET by the eddy covariance method, we found that without adjustment factors, Jensen-Haise, modified Penman (NDAWN) and ASCE-EWRI reference ET methods showed the best agreement than Thorthwaite with an R² of 0.63, 0.58, and 0.58, respectively. The default reference ET method showed poor agreement without adjustment factors with an R² of 0.05 and 0.10 for B1 and B2 locations, and improved to 0.8 and 0.70 by using adjustment factors.