雅鲁藏布江中下游地区湿季降雨量的空间分布模型

对于水文气象站点稀少的青藏高原山区,获得较为精确的降雨空间分布信息是进行降雨径流模拟的前提。该文利用GIS手段和多元回归方法,建立了雅鲁藏布江中下游地区旱、偏旱、正常、偏涝、涝等5种类型年份的湿季(5—9月)降雨量的空间分布模型。结果显示:多种地理地形因子对湿季降雨空间分布均有重要影响。统计分析表明:(1)雅鲁藏布江中下游流域地形复杂、降雨站点稀少,回归分析中的地理、地形因子共线性问题突出,方差扩大因子法、Pear-son相关系数法和因子分析法的组合应用是识别和解决共线性问题的有效方法;(2)影响湿季降雨的局部地形因子表达存在着最优空间尺度,15—25km半径范围内的平均坡度、平均坡向、开放度值与降雨也有较高相关性,但考虑其他局部地形因子后,开放度对降雨空间分布的影响可以忽略。

Spatial precipitation distribution of over the middle and lower reaches of the Yarlung Zangbo River

The Tibetan Plateau has few meteorological stations,so precise spatial precipitation distribution data is difficult to find for hydrological simulations.Multiple linear regressions were used with Geographic Information System techniques to estimate the spatial precipitation distribution in the rainy season(from May to September) in the middle and lower Reaches of the Yarlung Zangbo River for 5 types of hydrological years(i.e.,dry,relatively dry,normal,relatively wet,and wet years).Geographical and topographical factors have significant impacts on the rainy-season precipitation.The results show that significant multicollinearity among the geographic and topographic factors in the regression analysis due to the complex terrain and sparse meteorological stations in the middle and lower Reaches of the Yarlung Zangbo River.Integration of the variance inflation factor(VIF),Pearson correlation,and factor analysis methods may be an efficient way to identify and solve the multicollinearity problem.In addition,local terrain factors were found to give the optimal spatial scales to determine the analyses factors and the best relative slope with the rainy-season precipitation.The terrain slope,geometric features in regions with radii of 15—25 km and the exposures in regions with radius of 25 km around a meteorological station show the highest correlations.The effect of exposure can be neglected relative to the other terrain factors.