Regional climate simulations over South America: sensitivity to model physics and to the treatment of lateral boundary conditions using the MM5 model
Regional climate simulations over South America: sensitivity to model physics and to the treatment of lateral boundary conditions using the MM5 model
复制标题
南美洲区域气候模拟：对模型物理的敏感性以及使用 MM5 模型处理横向边界条件的敏感性
作者：
S. Solman;N. Pessacg
本文通过一系列敏感性试验，评价了MM5模式对南美洲暖季季节内变率主模的模拟能力。用不同的积云方案和行星边界层方案对ERA40再分析进行了几次为期3个月的模拟，试图确定模拟拉普拉塔盆地(LPB)和南大西洋辐合带地区干湿交替条件的最佳物理参数组合。将结果与不同的观测数据集进行了比较，并在考虑了目标区域月降水量的空间分布和日降水量统计的情况下进行了模型评估。虽然每个实验都能够捕捉到模拟期间降水的对比行为，但降水在很大程度上被低估了，特别是在LPB地区，这主要是由于在水汽通量辐合中的错误表述。用网格轻推行星边界层上方的风的实验表明，与不对模式区域内的区域环流施加约束的实验相比，其性能更好。总体而言，没有发现任何单一的实验在整个领域和两个相反的月份表现最好。实验的效果取决于感兴趣的地区，是使用Grell(Kain-Fritsch)积云方案与更适合于副热带(热带)纬度的MRF行星边界层方案相结合的模拟。与任何单独的实验相比，敏感度实验的整体表现更好。
In this study the capability of the MM5 model in simulating the main mode of intraseasonal variability during the warm season over South America is evaluated through a series of sensitivity experiments. Several 3-month simulations nested into ERA40 reanalysis were carried out using different cumulus schemes and planetary boundary layer schemes in an attempt to define the optimal combination of physical parameterizations for simulating alternating wet and dry conditions over La Plata Basin (LPB) and the South Atlantic Convergence Zone regions, respectively. The results were compared with different observational datasets and model evaluation was performed taking into account the spatial distribution of monthly precipitation and daily statistics of precipitation over the target regions. Though every experiment was able to capture the contrasting behavior of the precipitation during the simulated period, precipitation was largely underestimated particularly over the LPB region, mainly due to a misrepresentation in the moisture flux convergence. Experiments using grid nudging of the winds above the planetary boundary layer showed a better performance compared with those in which no constrains were imposed to the regional circulation within the model domain. Overall, no single experiment was found to perform the best over the entire domain and during the two contrasting months. The experiment that outperforms depends on the area of interest, being the simulation using the Grell (Kain–Fritsch) cumulus scheme in combination with the MRF planetary boundary layer scheme more adequate for subtropical (tropical) latitudes. The ensemble of the sensitivity experiments showed a better performance compared with any individual experiment.