Spatial Downscaling of Monthly TRMM Precipitation Based on EVI and Other Geospatial Variables Over the Tibetan Plateau From 2001 to 2012

• Main Authors: Yuli Shi, Lei Song
• Format: Article
• Language: English
• Published: International Mountain Society 2015-05-01
• Series: Mountain Research and Development
• Subjects: precipitation; random forests; Moderate Resolution Imaging Spectroradiometer (MODIS); EVI; spatial downscaling; Tropical Rainfall Measuring Mission (TRMM); Tibetan Plateau
• Online Access: http://www.bioone.org/doi/full/10.1659/MRD-JOURNAL-D-14-00119.1

Recent developments in hydrological modeling and biomass retrieval in complex mountain areas have heightened the need for accurate precipitation data at high spatial resolution. The Tropical Rainfall Measuring Mission (TRMM) provides rainfall estimates for certain climate models in mountain ranges where rain gauges are lacking. TRMM precipitation estimates, however, inherently have large uncertainties because of their coarse spatial resolution. In this study, we investigate a statistical downscaling calibration procedure to derive high-spatial-resolution (1-km) precipitation maps for the Tibetan Plateau using the satellite-based data set Enhanced Vegetation Index (EVI) from the Moderate Resolution Imaging Spectroradiometer, a digital elevation model from the Shuttle Radar Topography Mission, and the TRMM 3B43 product. Spatial downscaling from 0.25° to 1 km was achieved by using the nonparametric statistic relationships between precipitation and EVI, altitude, slope, aspect, latitude, and longitude. An additive method was used to calibrate the downscaled precipitation data. The best 1-km resolution annual precipitation data for 2001–2012 over the Tibetan Plateau were generated through downscaling and additive calibration for most cases. The results show that the method improves the accuracy of rainfall estimates. Monthly 1-km precipitation data were also obtained by disaggregating 1-km annual downscaled estimates with monthly fractions of annual total precipitation. Monthly precipitation predictions are in good agreement with rain gauge data. The calibration of the monthly product with rain gauge data significantly reduced the bias value. Overall we conclude that the methodology is useful for areas with varied climate conditions and complex topography. These results have practical implications for calculating hydrological balances, mapping aboveground biomass, and assessing regional climate change.