Modeling transpiration and soil water content from a corn ( Zea Maize L.) field: 20 min vs. daytime integration step

Abstract

Simulations of transpiration (T) and soil water depletion using daytime and 20 min interval models of similar general structure were compared. The performance of the models were evaluated using the measurements obtained from a full-canopy corn field subjected to wet and dry irrigation treatments. The wet treatment was well irrigated throughout the season while irrigation for the dry treatment was stopped when full canopy was reached. 20 min interval evapotranspiration (ET) in the wet treatment was determined using the Bowen ratio method and was used as a crop maximum ET input to the models. For both treatments, sap flow measurements averaged over 20 min intervals were performed continuously for 2 to 9-day periods, for a total of 40 days. These devices measure crop water uptake, which can be equated to T by neglecting plant storage. Sap flow per unit green leaf area was measured simultaneously on 7–10 plants per treatment and expressed on a per unit ground area basis. In addition, soil water content and root density were measured throughout the season in both treatments. The daytime and the 20 min (with output integrated for daytime periods) models performed similarly in the simulation of T for both irrigation treatments. Overall, the root mean square error (RMSE) fluctuated from 0.28 to 0.38 mm day −1, and the relative error (RE=100×RMSE/observed mean) from 7.2 to 10.5%. Also, both models performed similarly in simulating soil water content. No advantage was found from the use of the shorter time step model. The use of a daytime integration step appears sufficient unless short-time interval details are needed as model output or for use by other components of a larger model.