Abstract:
Selected land surface variables from the MERRA reanalysis, including daily maximum and minimum air temperatures, atmosphere vapor pressure deficit, incident solar radiation (SWrad), and surface soil moisture were evaluated against global satellite microwave (AMSR-E) retrievals, earlier generation atmospheric analysis (GEOS-4), and in-situ biophysical measurements. Relative to GEOS-4, MERRA is generally warmer and drier for low and middle latitude regions (< 50ｰN) associated with reduced cloudiness and increased SWrad. MERRA and AMSR-E show favorable agreement for temperature predictions with overall accuracy (RMSD) of 4ｰC. Surface (ﾜ 2cm depth) soil moisture estimates from MERRA and two AMSR-E products are moderately correlated (R2 ~ 0.2) for middle latitude regions with low to moderate vegetation biomass. MERRA surface soil moisture also corresponds favorably with in situ observations (R2 > 0.2, p < 0.001) in the mid-latitudes, where accuracy is directly proportional to the quality of MERRA precipitation. In the high latitudes, MERRA shows inconsistent seasonal water and heat dynamics relative to in-situ observations. A simple soil heat transfer model is developed in order to partially correct overestimation in MERRA surface temperatures in the high latitudes, which may cause bias in biophysical modeling depending on soil temperature inputs. The upcoming SMAP mission is expected to improve MERRA type reanalysis accuracy especially where surface meteorological observations are sparse and in cold land regions subject to seasonal freeze/thaw transitions.