Résumé
The FAO56 booklet (Allen et al., 1998) is still providing the preferred method to compute unstressed evapotranspiration (ET) and transpirations (T) rates for a wide range of applications in hydrology, agronomy and climatology. Even the recent works by key authors (Lhomme et al., Shuttleworth et al., to name a few) base their algorithms on the widespread availability of the crop coefficient libraries of maximum basal crop coefficient values and typical plant phenological cycles. However, unstressed evapotranspiration depends chiefly on the actual vegetation development, therefore on the accurate growth/senescence cycle, rather than, say, the maximum basal crop coefficient itself, which is often set to 1.2 for a vegetation fully covering the ground. This is why remote sensing has gained popularity to derive these maximum ET and T values from LAI and fraction cover (fc) time series retrieved from visible/near infrared reflectances. Allen et al. (2009) provide a comprehensive translation of the FAO56 method into unstressed ET by expressing crop coefficients according to remotely sensed fc only. Since times series of fraction cover and LAI, as well as maximum vegetation height, are theoretically sufficient to constraint the energy budget equation, it is about time to provide a more comprehensive effort to parameterize unstressed ET and Tusing state-of-the-art dual source energy budget models. Amongst the challenges faced by most existing method is the fact that the net radiation is computed by assuming that the equilibrium surface temperature is close to the air temperature at the reference level. Herewe test on four different datasets (two in South of France, one in Morocco, and one in South India) the relevance of four algorithms: Allen et al. (2009), Lhomme et al., (2015), Boulet et al. (2015) and a new parameterization based on Boulet et al. (2015) with the Priestley-Taylor equilibrium rates as a baseline. Results show that for the wettest environments the new formulation is the most accurate for monitoring ET after large rainfalls (an indicator of unstressed conditions), while for the driest the SPARSE formulation based on Penman-Monteith (Boulet et al., 2015) is the most accurate. Both Allen et al. (2009) and Lhomme et al. (2015) tend to overestimate unstressed ET because the equilibrium surface temperature is often a few degrees higher than the air temperature, leading to a higher net radiation and lower sensible flux in absence of moisture limited conditions. The study also provides some baseline unstressed T for the case of a drip irrigation system where the interrow is often dry and advects sensible heat to increase T.