An operational model to estimate hourly and daily crop evapotranspiration in hilly terrain: validation on wheat and oat crops

In this paper, we present an operational model to estimate the actual evapotranspiration (ET) of crops cultivated on hilly terrains. This new model has the following three characteristics: (1) ET modelling is based on a Penman-Monteith (PM) type equation (Monteith 1965) where canopy resistance is simulated by following an approach already illustrated by Katerji and Perrier (Agronomie 3(6):513-521, 1983); (2) the estimation of ET, by means of the PM equation, is made by using meteorological variables simulated on sloped sites as input; (3) these variables are simulated by using simple relationships linking the variables measured at a reference site on plane to the topographic characteristics of the site (slope, orientation, altitude as difference between reference, and sloped sites). This approach presents two advantages if compared with previously proposed models: Not only computation steps are greatly simplified but also error sources due to the simulation of climatic variables in sloped sites and the ET estimation are well distinguished. This model was validated at hourly and daily scales at four sites cultivated with wheat and oats offering a wide range of slope and orientation values: a reference site on plane, site 1 (9A degrees sloping, NW orientation, 7 m from the reference site in plane), site 2 (6A degrees, SE, 12 m) and site 3 (1A degrees, SE, 18 m). At hourly scale, the new model performed well at all sites studied. The observed slope of the linear relationships between estimated and measured ET values ranged between 0.93 and 1.03, with coefficients of determination, r (2), between 0.80 and 0.98. At daily scale, the slopes of the linear relationships between measured and estimated ET for the sites on plane and the sloped sites were practically the same (0.98 A +/- 0.01); however, the coefficient of determination r (2) observed in the site on plane was clearly greater (0.98) than that observed in the sloped sites (0.83). The presented analysis does not show any significant systematic effect of topography (slope and orientation of the plots) on the good performance of the proposed model for the ET estimation. Furthermore, we observed that coefficients of determination tend to decrease with the increase in the slope of the site, which translates into increased inaccuracy of the climatic variables simulation, in particular the net radiation, as the slope of the site increases. The proposed model allows to verify the different steps for calculating the fluxes, to identify the eventual sources of error and to make the needed corrections. For this reasons, the proposed model seems to be particularly operational, i.e. a useful tool for estimating fluxes on hilly terrains.