Similarities and differences of two evapotranspiration models with routinely measured meteorological variables: application to a cropland and grassland in northeast China

Han, Songjun; Xu, Di; Wang, Shaoli; Yang, Zhiyong

Similarities and differences of two evapotranspiration models with routinely measured meteorological variables: application to a cropland and grassland in northeast China

2014

Han, Songjun | Xu, Di | Wang, Shaoli | Yang, Zhiyong

Local actual evapotranspiration can be estimated with routinely measured meteorological variables using the Penman–Monteith model with surface resistance parameterized via the Katerji and Perrier approach (Agronomie 3(6):513–521, 1983; PM–KP model), or the nonlinear complementary relationship (CR) model proposed by Han et al. (Hydrol Process 26:3973–3981, 2012). A comparative study was carried out to evaluate the consistencies and differences of two models, as well as the performances of them for a cropland and grassland in northeast China. The departure of the actual evapotranspiration from the potential evaporation is described as a function of the ratio of the surface resistance to the aerodynamic resistance in the Penman–Monteith model, but the ratio of the aerodynamic term to the radiation term in the CR models. The two ratios are connected using a semi-empirical linear function by the Katerji and Perrier approach. The nonlinear CR model can be regarded as replacing the linear function by a power function after mathematical processing. On the other hand, the PM–KP model can be also considered as a CR-type model. On the daily basis at a maize cropland and degraded grassland in semiarid Northeast China, the nonlinear CR model with locally calibrated parameters performed better with data occupying all the growth stages, but the performances of the two models are similar during the early-, mid-, and late-season stages, respectively. On the half-hourly basis, the PM–KP model and the nonlinear CR model both performed well. It is deduced that on the daily basis the nonlinear CR model is more suitable for the cropland and grassland, but further comparisons are needed on the hourly basis.

Afficher plus [+]