Derived Optimal Linear Combination Evapotranspiration (DOLCE): a global gridded synthesis ET estimate

Abstract

Abstract. Accurate global gridded estimates of evapotranspiration (ET) are key to understanding water and energy budgets, as well as being required for model evaluation. Several gridded ET products have already been developed which differ in their data requirements, the approaches used to derive them and their estimates, yet it is not clear which provides the most 10 reliable estimates. This paper presents a new global ET dataset and associated uncertainty with monthly temporal resolution for 2000-2009. Six existing gridded ET products are combined using a weighting approach trained by observational datasets from 159 FLUXNET sites. The weighting method is based on a technique that provides an analytically optimal linear combination of ET products compared to site data, and accounts for both the performance differences and error covariance between the participating ET products. We examine the performance of the weighting approach in several in-sample and out-15 of-sample tests that confirm that point-based estimates of flux towers provide information at the grid scale of these products.We also provide evidence that the weighted product performs better than its six constituent ET product members in three common metrics. Uncertainty in the ET estimate is derived by rescaling the spread of participating ET products so that their spread reflects the ability of the weighted mean estimate to match flux tower data. While issues in observational data and any common biases in participating ET datasets are limitations to the success of this approach, future datasets can easily be 20 incorporated and enhance the derived product.

Global observed and modelled impacts of irrigation on surface temperature

“…The DOLCE data set provides three tiers of composite ET estimates at a 0.5 × 0.5° spatial resolution, which are derived from a trained weighted combination of different global ET products, including MOD16, GLEAM, and the ET products of the Max Planck Institute for Biogeochemistry (MPIBGC; Jung et al ., ). The three DOLCE tiers differ in the number of ET products included, and the number of flux towers employed for the weight training algorithm (Hobeichi et al ., ).…”

“…Due to documented uncertainties in the ET estimation in MOD16 (Decker et al ., ; Hobeichi et al ., ), two other ET products are also included in our analysis: the Global Land Evaporation Amsterdam Model (GLEAM) and the Derived Optimal Linear Combination Evapotranspiration (DOLCE). GLEAM products provide three different estimates of terrestrial ET (Martens et al ., ) at a 0.25 × 0.25° spatial resolution.…”

“…Surprisingly, the DOLCE tier 1 and tier 2 data sets, which contain three other three ET products besides GLEAM, do show a consistent pattern in ET change compared with the MOD16, and support our results about evaporative cooling in the western United States and Great Plains. It should be noted that neither the evaporation models of GLEAM and MOD16 nor the weight training algorithm in DOLCE explicitly considers irrigation (Mu et al ., ; Martens et al ., ; Hobeichi et al ., ). Therefore, the potential effects of irrigation on ET may not be included in these data sets and cannot be detected in our algorithm.…”

Observed and modelled historical trends in the water‐use efficiency of plants and ecosystems

“…We do not include leaf gas exchange data because, while they are commonly used to estimate instantaneous leaf‐level water and carbon exchanges (Long & Bernacchi, ; Wingate, Seibt, Moncrieff, Jarvis, & Lloyd, ), they are not useful for estimating long‐term trends because of the practical impossibility of adequate long‐term sampling. We do not include estimates of WUE based on remotely sensed vegetation greenness and carbon and water fluxes (Hobeichi, Abramowitz, Evans, & Ukkola, ; Parazoo et al, ), because none of the data sets used are solely driven by remotely sensed observations—they rely upon models and invoke substantial assumptions. Nevertheless, given the increasing interest in using remote sensing for investigating spatial changes in forest carbon and water balance, we have addressed them in the Supporting Information (Text S1).…”

Multimodel‐based analyses of evapotranspiration and its controls in China over the last three decades

“…Therefore, it is critical to fully understand the variations and drivers of ET, especially in the context of climate change (Fisher et al, ; Seneviratne et al, ). In the past few decades, using ET estimates from remote sensing, land surface models (LSMs), site observations, and reanalysis data, many researchers have examined the spatiotemporal variations of ET at regional and global scales (Elhag, Psilovikos, Manakos, & Perakis, ; Vinukollu, Sheffield, Wood, Bosilovich, & Mocko, ; Mueller et al, ; Li et al, ; Li, Wang, Chen, Yang, & Wang, ; Mao & Wang, ; Feng, Su, Ji, Zhi, & Han, ; Hobeichi, Abramowitz, Evans, & Ukkola, ; Zhang et al, ). However, our knowledge about the drivers of long‐term variations in ET and the controls of ET at regional and global scales are still insufficient, due to the substantial uncertainties in the ET estimates (Mueller et al, ; Chen et al, ; Wang et al, ; Liu et al, b; Mao & Wang, ; Li et al, ; Wartenburger et al, ).…”