Construction Of High Spatial And Temporal Resolution Forcing Data And Its Application In Noah-MP Land Surface Model

The exchange process of water, heat and momentum between land surface and atmosphere is called land surface process. Using land surface model of different complexity to simulate is one of the important methods for land surface process study, and high precision, high spatial, high temporal resolution atmosphere forcing data can effectively improve land surface model simulation. In this paper, an independent experiment is designed to test the fusion result of multigrid variational analysis method; using multigrid variational analysis method to combine the ERA-Interim reanalysis data and observations from over 40000 ground meteorological stations in China, combined with two kinds of CMORPH precipitation products and shortwave radiation products of FY-2 satellite, we construct high spatial and temporal resolution (1 hour,0.0625°) CMA Land Data Assimilation System (CLDAS) version 2.0 atmosphere forcing data,based on hourly observing data from national meteorological stations,we evaluate and compare CLDAS2.0 atmosphere forcing data, ERA-Interim reanalysis data and the similar operational product of America (Global Land Data Assimilation System atmosphere forcing data); using CLDAS2.0 atmosphere forcing data to drive Noah-MP land surface model, we can get hourly soil temperature and moisture simulations from 2013 to 2014 in China, and based on hourly observations from automatic meteorological stations, we evaluate the simulation results. The conclusions are summarized as follows:(1) Using multigrid variational analysis method, we fuse ERA-Interim reanalysis data and hourly surface temperature, pressure, humidity, wind speed observations from over 40000 meteorological stations except international exchange meteorological stations, then we design an independent experiment to evaluate the fusion result of multigrid variational analysis method using observation data from international exchange meteorological station. The results show that the multigrid variational analysis method can effectively acquire observation information, and if background field and observations do not fit in, it can make effective correction on the background according to the observation data, which provides the basis of using the multigrid variable analysis method to construct high spatial and temporal resolution atmosphere forcing data.(2) Using hourly observations from the national meteorological station, we evaluate and compare CLDAS2.0 atmosphere forcing data, ERA-Interim reanalysis data and Global Land Data Assimilation System atmosphere forcing data. The results show that all three kinds of data are able to reflect the characteristics of spatial distribution of observation, and the fusion of high density observation data makes CLDAS2.0 atmosphere forcing data’s quality is not only superior to the ERA-Interim reanalysis data, but also better than similar operational products of America. At present, CLDAS2.0 atmosphere forcing data which is constructed in this paper has been put into operation in China Meteorological Administration, and well applied in lots of institutes, like Beijing Normal University, Institute of Atmospheric Physic (Academy of Sciences).(3) Using hourly soil moisture and soil temperature observation data from automatic meteorological station which are built by China Meterological Administration, we evaluate Noah-MP land model simulation, then we divide China into eight research areas based on climate characteristics, and analyse temporal variation characteristics of regional averaged soil moisture and soil temperature,the results show that:Noah-MP land surface model can well simulate the spatial distribution of 0-10cm soil moisture, bias is mainly distributed within ±0.06m3/m3, and the correlation coefficient is mainly distributed in the range of 0.5 or more, in the aspect of regional scale, Noah-MP can well simulate the temporal change of soil moisture in each research areas and bias in all study area is within ±0.03m3/m3; correlation between Noah-MP land surface model’s 10cm,40cm soil temperature simulation and observation is very strong, the correlation coefficients in the most of the stations are greater than 0.8, the bias is mainly distribute within ±2 ℃, in the aspect of regional scale, Noah-MP can well simulate temporal change in all research areas; however, soil temperature of northeast area in winter is underestimated.