Evapotranspiration simulated by Penman-Monteith and Shuttleworth-Wallace models over a mixed plantation in the southern foot of the Taihang Mountain, northern China

Evapotranspiration (ET) is one of the key processes of terrestrial energy/water transfer and carbon exchange, and it is also an important factor affecting regional climate and global carbon cycle. It is important to simulate ET over forest ecosystem to evaluate the role of forests in surface water cycle. In this paper, the Penman-Monteith (P-M) model and the Shuttleworth-Wallace (S-W) model were used to simulate ET over a mixed plantation in the southern foot of the Taihang Mountain, northern China. The mixed plantation is mainly composed of Quercus variabilis, Platycladus orientalis and Robinia pseudoacacia. In order to evaluate the model applicability, ET simulated by the S-W model and the P-M model was compared with the measured one obtained by the eddy covariance technique. In addition, the sensitivities of P-M and S-W models to their resistance parameters were analyzed and the diurnal and seasonal variations in ET were explored. ET simulated by the S-W model and the P-M model was about 6% and 21% lower than the measured ET, respectively. Compared with the S-W model, ET simulated by the P-M model was less than the measured one. Compared with the P-M model, the correlation coefficient, agreement index, root mean square error, mean absolute error between measured ET and simulated ET by the S-W model were lower. The fitted curve of the relation between ET simulated by the S-W model and measured ET was closer to the 1:1 line than P-M model. Accordingly, ET simulated by the S-W model was higher than P-M model. It was suggested that the S-W model was more suitable for simulating ET in this region, and ET simulated by the P-M model was low. ET simulated by the P-M model and the S-W model was most sensitive to the change of canopy resistance. The second sensitive parameter was the aerodynamic resistance from canopy to reference height for the S-W model, and it was insensitive to the aerodynamic resistance from soil to the canopy. Transpiration and evaporation were most sensitive to the canopy resistance and soil surface resistance, respectively. There were no significant differences between the diurnal and seasonal variations in ET simulated by the S-W model and the P-M model, but the simulation accuracy of the S-W model was higher than P-M model. ET simulated by the P-M model in 2009 was obviously lower than the measured one. The ratio of E to ET was 11.3%, and E accounted for 10% of total ET simulated by the S-W model. Therefore, the possible reason was that soil surface resistance was not taken into account in the P-M model. The reason of the deference between simulated ET by the S-W model and the P-M model and measured ET may be that the empirical coefficients in the models are different for any vegetation types, especially how to choose the suitable resistance parameters. Moreover, meteorological factors are different in varied regions, which will influence the model accuracy. Accordingly, in order to improve the model accuracy, the climate and environment situation in study area should be taken into account to find a suitable resistance parameter of the model.