初、边值条件对区域气候模拟的影响

Effects of Initial/ Lateral Boundary Conditions on Regional Climate Simulations

利用区域气候模式（RegCM2）对1998年夏季风气候进行了模拟,并就初、边值条件对模拟结果的影响情况进行了讨论,结果表明:该模式可以较好地模拟出月际尺度的气候变化,但对降水异常的模拟还需作进一步完善。通过在几个季节采用不同初始场进行的数值模拟发现,从春季开始的积分其结果对初始场的敏感性较高,初始场的差别会对后期模拟产生明显影响。相对来说,从冬季开始的积分,其对初始场的依赖性较小,初始场的差别会在积分过程中逐渐减小,因而在利用RegCM2进行区域气候模拟时宜从冬季开始。另外,通过对采用不同侧边界嵌套方案的模拟效果进行简单讨论,发现当采用较少的缓冲区（5圈）时,海绵边界对温度、比湿及位势高度的模拟要比指数松弛及流入流出边界好,降水的模拟也要比其他方案好一些,但对风场的模拟则不如指数松弛方案。

The simulation of the 1998 summer monsoon climate and the discussion of the effect of initial and lateral boundary conditions on simulation show that regional climate model (RegCM2) can well simulate the climate variations in monthly scale, but it should be further improved in simulating the climatic anomaly. From the simulations by adopting varied initial fields in different seasons, we find that, the simulation starting from spring is the most sensitive to the initial fields, and the deviation of the initial fields will cause a obvious effect on anaphase simulations. Relatively, the simulation starting from winter is not strongly dependent on initial fields, and the bias of initial fields will gradually decline in the later integration process, The regional climate simulation by using RegCM2 beginning from winter is suitable, in this kind of integration, the results of simulation are not notably affected by initial bias. In addition, we simply discuss the effect of using different lateral boundary schemes on climate simulations. The results show that, if adopting smaller buffer zone (five-grid-point-wide buffer zone), employing sponge lateral boundary brings about a better effect than employing exponential relaxation lateral boundary on temperature/ humidity/ geopotential height simulations, and a slight improvment also takes place in the precipitation simulations, but worse effect on the wind field simulations.