Trends in pan evaporation and reference and actual evapotranspiration across the Tibetan Plateau

The Tibetan Plateau is one of the areas of the world where humans have had a relatively minor impact. The plateau thus provides ideal conditions for investigating evapotranspiration (In this paper, evapotranspiration terms are defined as follows: (1) “Actual evapotranspiration” includes evaporation from water and soil and transpiration from the vegetation of a specific region; (2) “potential evapotranspiration” includes the maximum quantity of water capable of being evaporated from the soil and transpired from the vegetation of a specific surface; (3) “reference evapotranspiration” includes the maximum quantity of water capable of being evaporated from the soil and transpired from a hypothetical reference grass with an assumed height of 0.12 m, a fixed surface resistance of 70 s m −1 , and an albedo of 0.23; and (4) “pan evaporation” means evaporation from open circular pans with a diameter of 20 cm.) issues such as temporal trends of evapotranspiration, the pan evaporation paradox, and the complementary relationship hypothesis. We examined Penman‐Monteith reference evapotranspiration and pan evaporation from a 20‐cm pan by using a data set from 75 meteorological observatories across the plateau during the period 1966 to 2003. Actual regional evapotranspiration was estimated in 16 catchments across the plateau during the period 1966 to 2001. Reference evapotranspiration and pan evaporation significantly decreased at 47 and 38% of observatories, respectively, though air temperature at most of sites significantly increased ( P < 0.05); wind speed and sunshine hours significantly decreased at 85 and 43% of observatories ( P < 0.05). The annual reference evapotranspiration and pan evaporation averaged from all the observatories significantly decreased ( P < 0.05) while actual annual evapotranspiration averaged from all the catchments increased ( P < 0.1), indicating the existence of a pan evaporation paradox on the plateau. The analysis with a recovered stationary series method showed that decreasing trend in reference evapotranspiration was due to a decrease in wind speed and a decrease in net total radiation, and the increase in air temperature, however, showed little correlation with the declining trends in reference evapotranspiration and pan evaporation. Regional actual evapotranspiration and reference, Penman potential evapotranspiration, or pan evaporation exhibit complementary behavior, which, however, does not support Bouchet’s complementary hypothesis, perhaps because of the very low vapor pressure deficit. The current study suggests that the Bouchet’s complementary relationship needs to be reconsidered at high elevations.