Estimation of clear‐sky land surface longwave radiation from MODIS data products by merging multiple models

Abstract

[1] The surface longwave radiation budget plays an important role in the Earth's climate system. Remote sensing provides the most practical way to map surface longwave radiation on a large scale and at a high spatial resolution. In this paper, we evaluate both surface downward longwave radiation (DLR) and upwelling longwave radiation (ULR) models under clear-sky conditions from MODIS data products. There are multiple DLR models available with variable uncertainties, and the Bayesian Model Averaging (BMA) method is incorporated in this study to combine the predictive distribution of these models for better accuracy. The integrated estimates for DLR based on the BMA method have lower root-mean-square errors (RMSEs) and higher coefficients of determination (R 2 ) than the best individual model. The RMSEs decreased by approximately 10 W/m 2 at two forest sites and by approximately 4 W/m 2 at other sites. The R 2 value increased at each site by more than 0.05. Two models for calculating the surface upwelling longwave radiation (ULR) are also evaluated at 16 sites. The results show that both the land surface temperature (LST)-emissivity method and the direct method, the Wang-U model underestimate the clear-sky ULR. The validation results show that the surface net longwave radiation (NLR) estimated using DLR estimates based on the BMA method and ULR estimates based on the LST-emissivity method is the most accurate.Citation: Wu, H., X. Zhang, S. Liang, H. Yang, and G. Zhou (2012), Estimation of clear-sky land surface longwave radiation from MODIS data products by merging multiple models,

Multi-model ensemble prediction of terrestrial evapotranspiration across north China using Bayesian model averaging

“…Multimodel ensemble approaches have therefore become increasing popular in climate change projections (Tebaldi et al , ; Yang et al , ), land surface component simulation (i.e. soil moisture, Guo et al , ; surface longwave radiation, Wu et al , ), groundwater assessment (Neuman, ), hydrologic streamflow predictions (Duan et al , ; Vrugt and Robinson, ; Zhang et al , ), and terrestrial ET estimation (Ershadi et al , ; Yao et al , ). With the ensemble techniques, the SA approach focused on point predictions, while the BMA approach was mainly concerned with producing bias‐corrected probabilistic forecasts (Diks and Vrugt, ).…”

Surface-sensible and latent heat fluxes over the Tibetan Plateau from ground measurements, reanalysis, and satellite data

“…RF realizes the concept of ensemble learning from the group of regression trees through processes of boosting and bagging. An additional commonly used method, particularly in multimodel fusion, is Bayesian model averaging (BMA), which was implemented to optimize weights to integrate predictive distributions from multi-models such as the mapping of longwave radiative fluxes from multiple models (Raftery et al, 2005;H. Wu et al, 2012).…”

Using Bayesian model averaging to estimate terrestrial evapotranspiration in China

“…The contribution of each individual model in the BMA method is weighted by its posterior weight of evidence (Ellison, 2004). BMA has been widely used to study the climate change (Duan and Phillips, 2010), improve the predictions accuracy of hydrology (Duan et al, 2007), weather (Raftery et al, 2005;Wu et al, 2012), forest biomass and economics (Fernandez et al, 2001). Previous studies indicated better estimations of BMA than other multi-model ensemble methods (Viallefont et al, 2001;Ellison, 2004;Raftery et al, 2005;Sloughter et al, 2007).…”