Impacts of Horizontal Resolution and Topography on the Numerical Simulation of East Asian Precipitation

GAO Xue-Jie,XU Ying,ZHAO Zong-Ci
2006-01-01
Abstract:Traditionally General Circulation Models(GCMs) have shown a poor performance in simulating the East Asian monsoon precipitation patterns.The precipitation center simulated by GCMs is usually located too far north(over) central China.This can be found in many GCM simulations,e.g.,the Australia CSIRO model.However,in our previous work,after nesting the above mentioned CSIRO model results to a regional climate model (RegCM2) which has a much higher resolution,the simulation of the precipitation is highly improved.Not only regional details but also the large scale spatial distribution becomes closer to reality.To investigate the role the resolution and topography play in the simulation of East Asian precipitation,a series of experiments with different settings of them are conducted using RegCM2 nested in one-way mode within the CSIRO global coupled atmosphere-ocean model.Two sets of RegCM experiments are performed and intercompared: in one set both topography and resolution change across the range from 45 to 360 km(referred to as R45,R60,R90,…,R360);in the second set only the resolution changes while the topography remains the same as that of the CSIRO GCM(referred to as G60,G90,…,G360).It is found that the simulated East Asian large-scale precipitation patterns are significantly affected by resolution.More specifically,simulated precipitation is increasingly realistic with grid point refinement.For coarser resolutions the precipitation maximum is displaced too far north and west compared to observations,which is similar to the GCM simulation.Spatial correlation coefficients between simulated and observed annual mean precipitation are calculated to more quantitatively analyze the model simulation of precipitation spatial patterns at the 160 China stations.In general,it can be seen that the coefficient increases as the model resolution increases.R45 shows the highest value of 0.81,while R60 is 0.78,R90 is 0.71,etc.The effect of resolution is the most important during the mid to late monsoon months when smaller scale convective processes dominate.For example,although Meiyu front is a synoptic scale system,studies show that the precipitation is mainly caused by the(sub-GCM grid) medium and mesoscale convection systems.Coarse resolution does not allow the description of them and therefore the moisture from the sea is not efficiently removed over the coastal regions and can more easily reach the interior of the continent(central China) and precipitate there.During the winter months,the model reproduces well the large-scale distribution of precipitation over East Asia regardless of the resolution.Summer is the main rain season over China,precipitation is probably mainly caused by mesoscale convective systems.Coarse resolution does not allow the description of these systems and therefore the moisture from the sea is not efficiently removed over the coastal regions and can more easily reach the interior of the continent(central China) and precipitate there.While real topography is important,it is also shown that the simulations using a higher resolution but the coarse CSIRO topography perform better than the simulations using the coarser model resolution with real model topography.For example,the above mentioned coefficient is 0.75 for G60,0.71 for R90,and it is 0.60 for G90,(0.53) for R120,etc.This suggests that over East Asia it is more important to have adequate spatial resolution of the physical and dynamical processes than topography.Lastly,the results indicate that model resolutions of 60 km or higher are needed to accurately simulate the distribution of precipitation over China and East Asia.
What problem does this paper attempt to address?