In this study, we couple the Weather Research and Forecasting Model (WRF) with the Advanced Canopy-Atmosphere-Soil Algorithm (ACASA), a high complexity land surface model, to investigate the impact of canopy representation on regional evapotranspiration. The WRF-ACASA model uses a multilayer structure to represent the canopy, consequently allowing microenvironmental variables such as leaf area index (LAI), air and canopy temperature, wind speed and humidity to vary both horizontally and vertically. The improvement in canopy representation and canopy-atmosphere interaction allow for more realistic simulation of evapotranspiration on both regional and local scales. The coupled WRF-ACASA model is compared with the widely used intermediate complexity Noah land surface model in WRF (WRF-Noah) for both potential (ETo) and actual evapotranspiration (ETa). Two LAI datasets (USGS and MODIS) are used to study the model responses to surface conditions. Model evaluations over a diverse surface stations from the CIMIS and AmeriFlux networks show that an increase surface representations increase the model accuracy in ETa more so than ETo. Overall, while the high complexity of WRF-ACASA increases the realism of plant physiological processes, the model sensitivity to surface representation in input data such as LAI also increases.