Robust historical evapotranspiration trends across climate regimes.

Evapotranspiration (ET) links the hydrological, energy and carbon cycles on the land surface. Quantifying ET and its spatio-temporal changes is also key to under- standing climate extremes such as droughts, heatwaves and flooding. Regional ET estimates require reliable observation- based gridded ET datasets, and while many have been de- veloped using physically based, empirically based and hy- brid techniques, their efficacy, and particularly the efficacy of their uncertainty estimates, is difficult to verify. In this work, we extend the methodology used in Hobeichi et al. (2018) to derive two new versions of the Derived Optimal Linear Combination Evapotranspiration (DOLCE) product, with ob- servationally constrained spatio-temporally varying uncer- tainty estimates, higher spatial resolution, more constituent products and extended temporal coverage (1980–2018). Af- ter demonstrating the efficacy of these uncertainty estimates with out-of-sample testing, we derive novel ET climatology clusters for the land surface, based on the magnitude and variability of ET at each location on land. The new clus- ters include three wet and three dry regimes and provide an approximation of Köppen–Geiger climate classes. The veri- fied uncertainty estimates and extended time period then al- low us to examine the robustness of historical trends spa- tially and in each of these six ET climatology clusters. We find that despite robust decreasing ET trends in some re- gions these do not correlate with behavioural ET clusters. Each cluster, and the majority of the Earth’s surface, shows clear robust increases in ET over the recent historical period. The new datasets DOLCE V2.1 and DOLCE V3 can be used for benchmarking global ET estimates and for examining ET trends respectively.

Figure 4. Spatial pattern of ET climate trends in DOLCE V3 over 1980–2018 derived using Mann–Kendall and Sen’s slope methods. Grid cells in white correspond to unreliable ET trends either because (i) the confidence interval of the slope encompasses a mix of negative and positive values or (ii) trend slopes computed for multiple, different random samples of ET within the interval ET ± uncertainty do not agree in sign.