LAUSR

Abstract Evapotranspiration (ET) plays an important role in water and energy balance at the surface-atmosphere interface. It is widely reported that near-surface soil water content (SWC) or soil water potential (SWP) significantly affects ET and this parameter has been incorporated into the FAO Penman-Monteith (FAO-PM) model for prediction of ET during crop growth seasons. However, there is little information on the effect of SWC or SWP on prediction of ET during soil freeze-thaw cycles in winter. We present an experiment conducted at a demonstration farm with a crop of winter wheat, over two winters near Beijing, China. A lysimeter system equipped with a weather station was used to measure the ET flux and meteorological data. Unfrozen soil water content (USWC) and soil temperature (Tsoil) were measured using dielectric tube sensors (DTS) and digital temperature sensors, respectively. SWP was determined by measured USWC and a soil moisture characteristic (SMC) curve derived from the soil freezing characteristic (SFC) curve and the Clapeyron equation in frozen soil. Detailed measurements in year 1 showed that the FAO-PM model exhibited a complex error pattern, underestimating ET from unfrozen soil but overestimating ET from stable frozen soil. To address these errors, we define a freezing stress index (Ksf) as a function of SWP. Incorporation of Ksf as a modifier of the standard crop coefficient in the FAO-PM model improved prediction of ET (RMSE declined from 0.424 to 0.187 mm day−1, in year 1). These data revealed a correlation between SWP near the soil surface (