Study On The Method Of Constructing High Resolution Climate Change Scenarios Over Ningxia

Although General Circulation models (GCMs) represent the main features of the global atmospheric circulation reasonably well, their ability in reproducing regional scale climatic details is rather limited because of their low-resolution. As a result, there is a need to develop tools for downscaling GCM predictions of climate change to regional and local scales. High-resolution GCMs, regional climate models(RCM) nested in GCMs, statistical downscaling and spatial interpolation are the four main tools for downscaling. Many studies have been made to regional climate models China, but the resolution is not so high for the province application, so the re-downscaling is needed. But it takes so long time to construct the same scenarios in a very high resolution by RCM that the high resolution scenario can't be used in time. However, the statistical downscaling and the spatial interpolation can be applied to construct high resolution scenarios. This study focuses on research of statistical downscaling technique and spatial interpolation coupled with PRECIS, and their application of projecting history local climate scenarios over Ningxia.First 15-year (1979-1993) ECWMF reanalysis data was employed as quasi- observational boundaries to drive PRECIS at 25km and 50km horizontal resolution to simulate climate over Ningxia from 1979 to 1993. Second, after choosing the predictand and predictand area, empirical relationships are derived among daily maximum temperature, minimum temperature, precipitation and radiation obtained of PRECIS-50 and PRECIS-25 simulation at 1979-1989 by Multiple Linear Regression function. Subsequently, the statistical downscaling models are applied to PRECIS-50 and PRECIS-25 simulation at 1990-1993 to construct statistical result at PRECIS-25 resolution. Third, after choosing the interpolate method, the method of Moving Least Squares (MLS) is employed to interpolate PRECIS-50 simulation in 1990-1993 to PRECIS-25 grids. Finally, compare the result of two method of downscaling.The main results of this research can be summarized in the following items:1) PRECIS-25 simulation is better than PRECIS-50's over Ningxia and is similar with observation;2) Statistical downscaling model coupled with PRECIS can downscale the PRECIS-50 simulation to 25km×25km resolution, and the result is similar with PRECIS-25 simulation, but the correction by the mean daily bias is not efficient, so a further study on the correction is needed;3) The difference between different interpolate result is small, but the MLS is the best method relatively; although the location and topography are involved to the method of MLS, the interpolate result is not improved;4) Compared with MLS, statistical downscaling model coupled with PRECIS is better. The PRECIS-50 simulations' resolution is increased and the PRECIS-50 simulations are improved by statistical downscaling model coupled with PRECIS, but the interpolation can increase the resolution only. And the statistical model is simpler than MLS, so it takes shorter runtime.