GRAPES单柱模式的试验研究

A case study of the GRAPES single column model

中文关键词:  GRAPES，单柱模式，GCSS WG4，物理过程评估

英文关键词:GRAPES, Single column model, GCSS WG4, Physics process evaluation

中文摘要:

      根据GCSS WG4（Global Energy and Water Cycle Experiment Cloud System Study Working Group 4）第3次个例模拟的观测数据，为GRAPES（Global and Regional Assimilation and Prediction Enhanced System）设计了一个可用于检验其整套物理参数化过程对夏季中纬度陆地天气过程模拟的单柱模式试验,并利用该试验考察了不同复杂度的两种陆面过程（CoLM和SLAB）对温、湿度和降水模拟的影响。整个观测时段的模拟表明，模拟的降水与观测的量级一致，位温和水汽混合比没有明显偏离观测，这说明本试验的构造是合理的。考虑到模式系统误差对长期积分结果的影响，随后选取了4个降水子时段分别进行积分。结果表明，使用CoLM方案模拟得到的累积降水量均大于使用SLAB方案的，但使用CoLM方案时出现虚假降水的概率较大。由于区域平均的初始热动力廓线比实际降水发生地区偏干，使用两个方案的模拟均对子时段3的第1个降水事件延迟24 h左右，这对其在子时段3的相关系数都很小有贡献。时间平均的位温和水汽混合比误差分析表明，使用CoLM模拟的子时段1和2的对流层低层偏冷、偏湿，而其他情况下为偏暖、偏干。对流层低层位温的误差与地表气温的误差一致。此外，还发现使用CoLM模拟得到的感热通量偏小，潜热通量偏大，而使用SLAB模拟得到感热通量偏大，潜热通量偏小。对流层中高层，子时段1和4为偏冷、偏湿，对应降水偏少（使用CoLM的模拟在子时段1的降水偏多归因于虚假降水）；子时段2，使用CoLM的模拟为偏暖、偏干，对应降水偏多，使用SLAB的模拟为偏冷、偏干，对应降水偏少；子时段3，使用两个陆面方案的模拟均为偏冷、偏干，对应降水偏多。

英文摘要:

      Based on the GCSS WG4 (Global Energy and Water Cycle Experiment Cloud System Study Working Group 4) Case 3 data，A SCM (Single Column Model) experiment is designed to evaluate physics parameterizations performance of the GRAPES (Global and Regional Assimilation and Prediction Enhanced System) over the mid-latitude land in summer. The influences of two different complexity land surface schemes (CoLM and SLAB) on temperature, moisture, and precipitation simulation are also investigated. In the 29 d simulation, it is shown that the magnitudes of simulated precipitation are similar to those of observation, and there are unobvious departures of simulated potential temperature and water vapor mixing ratio from those of observation, suggesting that the design of the experiment is correct. Then the four precipitation sub period simulations are respectively performed in order to reduce the influence of the model system error. It is found that the simulated accumulated precipitation with the CoLM scheme is more than that with the SLAB scheme for the four sub periods. In the sub-period 3, the first precipitation event is delayed in both simulations, and their correlation coefficients are very small，which may have relation with the fact that area-averaged initial thermodynamic profile is drier than that of the area where precipitation occurred. The analyses of the time averaged errors of the potential temperature and water vapor mixing ratio show that the simulation with the CoLM has cold and moist biases in the lower troposphere in the sub period 1 and 2, while warm and dry biases in other simulations. These biases of potential temperature in the lower troposphere are related with the fact that the simulated surface temperature with the CoLM is lower than that of observed in the sub-period 1 and 2, while higher in other simulations. Moreover, it is found that the simulation with the CoLM simulated more latent heat fluxes and less sensible heat fluxes than observations, which are opposite to the simulation with the SLAB. In the middle and upper troposphere, cold and moist biases correspond to less precipitation for the two simulations in the sub-period 1 and 4 (the simulation with the CoLM has more precipitation in the sub-period 1 due to its false precipitation). Warm and dry biases correspond to more precipitation for the simulation with the CoLM and cold and dry biases correspond to less precipitation for that with the SLAB in the sub-period 2. Both simulations have cold and dry biases which correspond to more precipitation in the sub-period 3.