Abstract. Within the Budyko framework, the controlling parameter (ω in the Fu equation) is widely considered to represent landscape conditions in terms of vegetation coverage (M); however, some qualitative studies have concluded that climate seasonality (S) should be incorporated in ω. Here, we discuss the relationship between ω, M, and S, and further develop an empirical equation so that the contributions from M to actual annual evapotranspiration (ET) can be determined more accurately. Taking 13 catchments in the Loess Plateau as examples, ω was found to be well correlated with M and S. The developed empirical formula for ω calculations at the annual scale performed well for estimating ET by the cross-validation approach. By combining the Budyko framework with the semi-empirical formula, the contributions of changes in ω to ET variations were further decomposed as those of M and S. Results showed that the contributions of S to ET changes ranged from 0.1 to 74.8 % (absolute values). Therefore, the impacts of climate seasonality on ET cannot be ignored, otherwise the contribution of M to ET changes will be estimated with a large error. The developed empirical formula between ω, M, and S provides an effective method to separate the contributions of M and S to ET changes.

No abstract

Potential evapotranspiration represents the maximum evaporative capacity for a region, which underscores the importance of analysing its changes and attribution to provide information for water resource management. This study focused on the northern Loess Plateau of China, investigated the changes in potential evapotranspiration (ET 0 ) using data from 34 stations during 1960-2013 and conducted its temporal and spatial attribution through the differentiation equation method. The results show that the annual mean ET 0 is 1004.9 mm with a decreasing trend at the rate of 0.33 mm year −1 . The lowest values were found in the northeast and southwest regions, while the greatest ET 0 was in the northwest region. An 'evaporation paradox' existed as a whole, but not obviously. The ET 0 was most sensitive to actual vapour pressure (e a ), followed by solar radiation (R s ), mean temperature (T mean ) and wind speed (U 2 ). The inter-annual variation in the sensitivity of ET 0 in relation to climatic factors showed that ET 0 became more sensitive to U 2 and T mean but less sensitive to R s and e a from 1960 to 2013. Temporally, the increasing T mean resulted in the increase of ET 0 ; while its effect was weakened by the other three factors, especially by decreasing U 2 . The contributions of climatic factors to the ET 0 spatial variation trends were also analysed along two transects, but the dominant factors were different for each transect. Furthermore, the possible impacts of human activities on ET 0 changes were also discussed.