Examination of selected atmospheric and orographic effects on monthly precipitation of Taiwan using the ASOADeK model



Characterization and quantification of precipitation spatial variability are required for many studies, such as those in hydrology, ecology, forestry and agriculture. In mountainous terrains (e.g. Taiwan), interactions of moist air mass with underlying complex topography result in increased but often predictable precipitation spatial variability. When these interactions are quantitatively captured, they can contribute to precipitation mapping. In this paper, we applied a recently developed multivariate regression model [auto-searched orographic and atmospheric effects de-trended kriging (ASOADeK) regression] to examine selected atmospheric and orographic effects: atmospheric moisture gradient (MG), terrain elevation (Z), terrain aspect and related moisture flux direction (MFD), on Taiwan mean monthly precipitation climatology, primarily based on a dense gauge network (about 110 gauges per 10 000 km2). The results indicate that (1) the regression captures over 50% of spatial variability in mean monthly precipitation over the whole Taiwan island, with improved performance when the regression is applied to sectional regions; (2) the atmospheric MG and MFD are significant factors explaining mean monthly precipitation spatial variability in Taiwan, whereas the elevation effect is not as significant; (3) when the data of sectional regions are applied in the regression, the inferred MFDs vary in a certain range, indicating slightly different climate setting between northern and southern parts of Taiwan; (4) the degree of agreement between the ASOADeK regression-inferred MFDs and local observed prevailing winds is systemic, and can be used to explain the contrast of rainfall seasonality patterns between the northern part and the rest of Taiwan; and (5) a gauge network of 50 gauges per 10 000 km2 is sufficient to capture the orographic effect on mean monthly precipitation distribution. Copyright © 2008 Royal Meteorological Society