淡水河流域降雨時空分析及推估

Abstract

In this thesis, we try to use Geographic Information System (GIS) to exploit the spatial characteristics of rainfall within the Tam-Sui River Watershed, and then move forward to find the best method to interpolate rainfall. Here we build an rainfall interpolator environment database system which records the collected data and locations of those stations in the water shed. This database system also can provide users a convenient and efficient way to analyze the spatial of rainfall within the watershed. We compare four spatial interpolating models by the interpolator system: Thiessen Polygon method, Inverse-Distance Weighting (IDW) method, Kriging method, and Trend method to find the adequateness on stimating monthly rainfall, event rainfall, and typhonic total rainfall. The results show that Kringing is the best when rain fall is concentrated, such as air-mass thunderstrom, whereas Inverse Distance Weighting (IDW) is suitable for uniform distributed rainfall, such as frontal surface rain. In general, Kringing is better than the other methods.