Evaluation of Satellite Estimates of Downward Shortwave Radiation over the Tibetan Plateau

Kun Yang,Rachel T. Pinker,Yaoming Ma,Toshio Koike,Margaret M. Wonsick,Stephen J. Cox,Yuanchong Zhang,Paul Stackhouse
DOI: https://doi.org/10.1029/2007jd009736
2008-01-01
Abstract:The state‐of‐the‐art satellite products of downward shortwave radiation over the Tibetan Plateau against ground observations are evaluated in this study. The satellite products include the International Satellite Cloud Climatology Project‐Flux Data (ISCCP‐FD) as produced at the NASA Goddard Institute for Space Studies (GISS) from the ISCCP D1 data, the Global Energy and Water Cycle Experiment‐Surface Radiation Budget (GEWEX‐SRB) results as derived at the NASA Langley Research Center (LaRC) from the ISCCP DX data, and a University of Maryland product derived with a modified version of the University of Maryland Surface Radiation Budget (UMD‐SRB) model as implemented with METEOSAT‐5 observations. These products are at different spatial and temporal resolutions, and the evaluation is performed at their native resolutions. Comparisons indicate that, in this region of great variation in elevation, using hourly, spatially homogeneous, and high resolution satellite data (UMD‐SRB) compares more favorably with surface measurements than products that use three hourly, sparse subsamples at coarse resolutions (ISCCP‐FD and GEWEX‐SRB). Discrepancies among the satellite products are usually larger in highly variable terrain (such as in the Himalayas region) and smaller for nonvariable terrain (such as in the central Plateau). This suggests that errors of satellite products are spatially dependent over the Tibet. Therefore caution needs to be exercised when extending comparison results based on limited in situ data from accessible sites to the entire Plateau. Attention should be also given to the quality of input parameters besides cloud properties, as there are large discrepancies among the satellite products for clear‐sky radiation.
What problem does this paper attempt to address?