本研究利用普渡區域模式（Purdue Regional Model ，PRM ）進行10 年夏季
模擬，針對模式結果進行大氣環流及地表變數的驗證，同時為求客觀及數量化，
亦利用氣候統計方法評估模擬結果，以便了解PRM 模擬東亞夏季季風的能力。
除了降雨外，其他模擬變數的誤差均很小，型態相關（pattern correlation ）也相
當高（大於0.8 ）。降雨誤差的進一步分析，發現PRM 的夏季模擬，存在一個明
顯的系統性誤差結構。若將此系統性誤差扣除，降雨的模擬誤差值與型態相關均
大幅提升。在1991~2000 十年的年際變化模擬方面，PRM 模擬的各年夏季雨量
的分布，與觀測雨量分布的情形相當相近，模擬降雨分布的型態相關係數表現的
相當理想（型態相關為0.68~0.94 之間）。模擬東亞地區平均的距平值除了1994 、
1995 、1996 及1998 ，因模擬結果太過接近平均值，表現出距平訊號太弱，呈現
與觀測距平值反號外，其他各年距平的分布均與觀測值呈現同相（即正負號相
同）。PRM 成功的模擬此一現象。在季內振盪方面，模式模擬的30~60 天低頻訊
號，在南海地區的強度與相位均與觀測資料相似。綜觀以上結果，PRM 在東亞
夏季季風的模擬，不論是氣候平均、年際變化或季內振盪，都有相當令人滿意的
表現。 This study used the Purdue Regional Model (PRM) to simulate the East Asian
summer monsoon for the 10 summers from 1991-2000. Objective statistical methods
and comparisons between the simulated and observed circulation and surface
variables were performed to evaluate the performance of the PRM. All simulated
variables except precipitation exhibit small errors and high pattern correlations with
the observed. An analysis of the precipitation errors reveals the existence of a
spatially-coherent systematic error pattern. After removing this pattern, the precipitation errors reduce significantly and the pattern correlations rise from 0.3-0.4
to 0.6-0.8. The signs of the simulated regional-averaged precipitation anomalies are
the same as the observed (e.g., negative anomalies in 1996 and 1998 and positive
anomalies in 1997, 1999 and 2000). This result indicates the well-simulated
interannual variability by the PRM. The PRM also simulated the observed negative
correlation between the 850hPa geopotential high and the precipitation in East Asia
and Taiwan. In the intraseasonal time scale, the PRM well simulated the fluctuations
of the 850hPa vorticity in both amplitude and phase in the South China Sea. Overall
speaking, the PRM is able to simulate the East Asian summer monsoon variability
from sub-seasonal to interannual time scale.
