Application of GSMaP and MODIS/SeaWiFS Downward Surface Short Wave Radiation in the Land Simulation System: Yesterday's Earth at EORC (YEE)

In the Global Precipitation Measurement (GPM) era, data assimilation is being recognized as a powerful tool for interpolating intermittent satellite data. Japan Aerospace Exploration Agency (JAXA) Earth Observation Research Center (EORC) established a research group on water cycle (W-RG) as one of the cross-cutting research groups over several satellite oriented missions, and the W-RG has started developing offline simulation system on water cycles over global land, which will be a basis for real-time data assimilation in the future, targeting GPM mission and the Global Change Observation Mission (GCOM) series. Current system does not have capability of data assimilation, but there are proposals to introduce the function in the system. The pilot system, called as the Yesterday’s Earth at EORC (YEE), is based on the “Today's Earth” from Institute of Industrial Science (IIS), The Univ. of Tokyo (Yoshimura et al., 2008), and simulates energy and water balances over global domain using Iso-MATSIRO, an extended version of MATSIRO (Minimal Advanced Treatments of Surface Interaction and Runoff) land surface model (LSM) (Takata et al., 2003). In addition to water balance components such as evapotranspiration, runoff, and soil moisture, river discharge is also calculated using Total Runoff Integrating Pathways (TRIP) (Oki and Sud, 1998). Currently, the system is just an offline simulation, using forecast data from the Japan Meteorological Agency (JMA) and satellite data as external forcing. Three experiments to use Iso-MATSIRO land surface model for 2003-2004. The first one is an experiment using JRA-25 as input forcing parameters (YEE-JRA25), and the second one is same as JRA25-EX but replace model rainfall to the Global Satellite Mapping for Precipitation (GSMaP) (Aonashi et al., 2009; Ushio et al., 2009) product (YEE-GSMaP), and the last one replace downward surface short wave radiation (SWR) product (Frouin and Murakami, 2006) by MODIS and SeaWiFS data (YEE-MODIS). Results were compared to in-situ data in daily and monthly basis. They were also compared to retrievals of them by the Advanced Microwave Scanning Radiometer for EOS (AMSR-E). In most of validation sites, satellite-based SWR data shows better agreement with in-situ measurement compared to that from JRA-25. Generally, output parameters from YEE-MODIS show better agreement with in-situ data and AMSR-E retrievals, except for snow depth in snowmelt season. Snow depths calculated by YEE-JRA25 and YEE-GSMaP show good correspondence to in-situ measurement for seasonal cycle. Results of YEE-MODIS and YEE-JRA25 are similar during mid-winter, however, YEE-MODIS overestimates during snow melt season. In future, we plan to replace fixed database such as Normalized Difference Vegetation Index (NDVI), and assimilation of forecast parameters such as soil moisture or snow depth are also discussed to improve YEE.