Evaluation of four models to estimate evapotranspiration under well-watered conditions for a maize agro-ecosystem in northern Italy

To ensure efficient management and planning of irrigation water resources, an accurate computation of actual evapotranspiration (ET) from cropped surfaces is needed. ET for well irrigated crops may be evaluated using models adopting a Penman-Monteith type formula (direct) or models based on a “crop coefficient approach” (indirect). This study evaluates the performance of different models for the estimation of ET for a maize agro-ecosystem grown in the Padana Plain (Northern Italy): 1) the original “onestep” PM model; 2) the “two-step” Shuttleworth model combining one-dimensional models of crop transpiration and soil evaporation; 3-4) the indirect “single crop coefficient” and “double crop coefficient” methods proposed by FAO-56, the latter allowing the separation of soil evaporation and crop transpiration. Latent heat fluxes measured in 2006 by eddy-covariance are used to evaluate the models accuracy. Results show that: i) the calculation of ET by direct methods is more accurate, confirming what found by many authors in the literature; ii) the PM “one-step” model underestimates ET during the initial and development growth stages, while it shows a good performance for LAI>3; iii) the SW “two-step” model successfully simulates ET; iv) the “single” and “double crop coefficient” FAO-56 models with generalized crop coefficients overestimate ET for the entire agriculture season; v) further research is needed to make the direct models suitable for operational use.