Premium
Calibration‐Free Complementary Relationship Estimates Terrestrial Evapotranspiration Globally
Author(s) -
Ma Ning,
Szilagyi Jozsef,
Zhang Yongqiang
Publication year - 2021
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2021wr029691
Subject(s) - evapotranspiration , fluxnet , environmental science , precipitation , climatology , eddy covariance , atmospheric sciences , water cycle , water balance , forcing (mathematics) , calibration , radiative forcing , climate change , meteorology , ecosystem , mathematics , geology , geography , statistics , ecology , oceanography , geotechnical engineering , biology
Abstract While large‐scale terrestrial evapotranspiration (ET) information is essential for our understanding of the Earth's water and energy cycles, substantial differences exist in current global ET products due partly to uncertainties in soil‐ and vegetation‐related parameters and/or precipitation forcing. Here a calibration‐free complementary relationship (CR) model, driven purely by routine meteorological forcing (air and dew‐point temperature, wind speed, and net radiation), mainly from ERA5, was employed to estimate global ET rates during 1982–2016. Modeled ET agrees favorably with (a) monthly eddy‐covariance measurements of 129 global FLUXNET sites, and; (b) water‐balance‐derived ET of 52 basins at the multiyear mean and annual scales. Additional evaluations demonstrate that the CR is very competitive, in comparison with other 12 widely used global ET products. The 35‐years mean global land ET rate from the CR is 500 ± 6 mm yr −1 (72.3 ± 0.9 × 10 3 km 3 yr −1 ) with more than 70% of the land area exhibiting increasing annual ET rates over the study period. Globally, CR ET significantly increased at a rate of 0.31 mm yr −1 during 1982–2016, suggesting a 2.2% increase in global land ET over last 35 years. Model inter‐comparisons indicate that global annual CR ET values and their trend are close to the median of not only the 12 ET products chosen but also that of 20 CMIP6 models. Since this calibration‐free CR model requires no precipitation (except in sea‐shore deserts for a subsequent ET correction), vegetation or soil data, it could be incorporated into complex hydrological and/or climate models, thereby facilitating large‐scale hydrological and climate simulations.