中文摘要：

利用中尺度气候模式MM5计算黑河山区流域2003年2月11日到6月30日的日降水量、2．0m高度的日平均气温和潜热，并将其嵌套到DWHC模型中。MM5运行周期为10d，积分步长为3S，空间分辨率为3km。保持DWHC模型土壤参数、植被参数、经验参数和可调参数等不变，仅对模型初始参数进行了调整，利用最近距离法（nearest）将MM5输出结果插值到1km×1km格点上，所计算的黑河干流出山口日平均流量与实测序列的NSE=0．79，B=-0．79（％），EV=0．79，R^2=0．8l。利用基于三角网格的立体插值法（cubic）所获结果与此相当，NSE=0．79，B=-0．65（％），EV=0．79，R。=0．80。这说明利用MM5-DWHC嵌套模型来模拟流域日平均流量是可行的。MM5-DWHC嵌套模型在径流模拟方面，比利用地面资料驱动结果要好。MM5-DWHC嵌套模型的计算结果表明，内陆河高寒山区流域存在明显的浅表产流特征，这与地面观测资料驱动结果一致。模型调试结果表明，MM5输出结果存在某种奇异性，且输出的非汛期降水量明显偏大。

英文摘要：

Using Mesoscale Model version 5 （MM5） to calculate the daily precipitation, daily averaged air temperature at the 2.0m heights and daily latent heat, from Feb. 11 to June 30, 2003, of the Heihe mountainous fiver basin and its near area, with a geographical boundary of 96. 786°- 102.284°E, 37. 328°-40. 601°N and an area of 17×10^4km^2, which was much larger than the area of the Heihe mountainous river basin （ 10009 km^2）, the DHWC model was calibrated. The spatial resolution of the MM5 is 3km, the integral timescale is of 3s, and the calculated cycle is about 10d. In the MM5 model, the Grell scheme cumulus parameterization method, the Dudhia option, the explicit moisture scheme （IMPHYS）, the cloud-radiation scheme, MRS PBL option, and the modified Oregon State University Land-surface model （OSULSM） were chosen to use. According to the geographical position of the MM5 results and projection transform methods, the MM5 results were projected into Alberts coordinate, which was the coordinate of the DWHC model, and were interpolated into 1 km×1 km, using nearest and cubic methods. The results showed that, when the nearest method was used, the Nash-Sutcliffe equation value of the dai- ly averaged runoff was of 0.79, the balance error was of -0.79%, and the R^2 value was of 0.81. When the cubic method was used, the Nash-Sutcliffe equation value, the balance error and the R2 value was of 0.79, -0.65 % and 0.80, respectively. Though the evaluation criterion values are not very high, the model results are much better than the model results using the data at the meteorological and hydrological stations, with a Nash-Sutcliffe equation value as 0.61. The model results are not very good because of the lack of the detailed soil and vegetation data. The MMS-DWHC model results also showed that the runoff production processes mainly occurred on the soil surface and in the shallow soil layers. The calibration results showed that, the MM5 results were singular to some extent.