A Bayesian Three-Cornered Hat (BTCH) Method: Improving the Terrestrial Evapotranspiration Estimation

Abstract

In this study, a Bayesian-based three-cornered hat (BTCH) method is developed to improve the estimation of terrestrial evapotranspiration (ET) by integrating multisource ET products without using any a priori knowledge. Ten long-term (30 years) gridded ET datasets from statistical or empirical, remotely-sensed, and land surface models over contiguous United States (CONUS) are integrated by the BTCH and ensemble mean (EM) methods. ET observations from eddy covariance towers (ETEC) at AmeriFlux sites and ET values from the water balance method (ETWB) are used to evaluate the BTCH- and EM-integrated ET estimates. Results indicate that BTCH performs better than EM and all the individual parent products. Moreover, the trend of BTCH-integrated ET estimates, and their influential factors (e.g., air temperature, normalized differential vegetation index, and precipitation) from 1982 to 2011 are analyzed by the Mann–Kendall method. Finally, the 30-year (1982 to 2011) total water storage anomaly (TWSA) in the Mississippi River Basin (MRB) is retrieved based on the BTCH-integrated ET estimates. The TWSA retrievals in this study agree well with those from the Gravity Recovery and Climate Experiment (GRACE).

Can Indirect Evaluation Methods and Their Fusion Products Reduce Uncertainty in Actual Evapotranspiration Estimates?

“…The BTCH method is used to generate an ET product (ET BTCH_ref ) by the weighted average of multiple products according to the accuracy of products (He, Xu, Xia, et al., 2020). In this paper, the ET BTCH_ref is applied as the reference value to evaluate the uncertainty in each ET product.…”

“…However, several studies show that the ET estimate generated by the AA method is not superior to a single ET product, whereas other ensemble methods can improve the ET estimate. These ensemble methods include the Bayesian Three Cornered Hat (BTCH), Bayesian Model Averaging, Support Vector Machine, Weighting Approach, and others (He, Xu, Xia, et al, 2020;Hobeichi et al, 2018;Yao et al, 2014Yao et al, , 2017. Among these ensemble methods, the BTCH method is most suitable for ungauged basins because it can integrate ET products without any prior knowledge, while other methods rely on the support of observations (He, Xu, Xia, et al, 2020).…”

“…These ensemble methods include the Bayesian Three Cornered Hat (BTCH), Bayesian Model Averaging, Support Vector Machine, Weighting Approach, and others (He, Xu, Xia, et al, 2020;Hobeichi et al, 2018;Yao et al, 2014Yao et al, , 2017. Among these ensemble methods, the BTCH method is most suitable for ungauged basins because it can integrate ET products without any prior knowledge, while other methods rely on the support of observations (He, Xu, Xia, et al, 2020). In summary, while numerous studies have attempted to reduce the uncertainty in ET estimates from different products and combinations of products by using a broad range of methods and metrics, it urgently requires a comprehensive comparison of the methods that can reconcile conclusions.…”

Improving global gross primary productivity estimation by fusing multi-source data products

“…The objective of this study is to adopt a new approach to combine the strengths of 45 GPP products without relying on any prior knowledge to produce an improved GPP dataset that achieves consensus on a relative majority of 45 GPP products. The Bayesian-based three-cornered hat (BTCH) method ( He et al., 2020 ) is applied to integrate 45 sets of 4 different types of monthly GPP products into this long-term dataset called BTCH-GPP on the pixel scale, with a period of 1980–2018 and a spatial scale of 0.5°. And we evaluate the performance of the fused dataset in the spatial and interannual variability of GPP from several eddy covariance towers and discuss the effectiveness of the fusion method.…”

Improve the Performance of the Noah‐MP‐Crop Model by Jointly Assimilating Soil Moisture and Vegetation Phenology Data

“…Accurate predictions of water and carbon fluxes in croplands are essential for determining crop yield, hydrologic components, and irrigation schedules and researching climate change (He, Xu, Xia, et al., 2020; Lokupitiya et al., 2016; Williams, Gornall, et al., 2016; Xu, Guo, et al., 2018). Process‐based land surface models (LSMs) have been developed to represent the complex land/hydrological interactions in earth systems (Chen & Dudhia, 2001; Dai et al., 2003; Lawrence et al., 2019; McDermid et al., 2017).…”