Abstract
Long-term variability of precipitation characteristics are investigated on the coastal region of Niigata Prefecture, by statistical analysis of daily precipitation series observed at Niigata, Murakami and Takada Meteorological Observatories. Rainfall data at each station is divided into 3 equal terms and analyzed by three hydrologic variables:(1) annual maximum daily value, (2) annual maximum storm total, and (3) annual maximum 3-day value. Snowfall data is divided into 2 equal terms and analyzed by annual maximum daily and storm total values only. Comparison of the frequency analyses across terms shows that within the same Kaetsu district stochastic rainfall has recently increased in Niigata and decreased in Murakami, while the change at Takada located in Joetsu district is small. Correspondingly, the return periods of given rainfall amounts have become shorter in Niigata and longer in Murakami during the most recent term. At Takada the direction of change depends on the hydrologic variable considered. In contrast to the rainfall data, stochastic snowfall has recently decreased in Niigata, Murakami and Takada, although again the change at Takada is small. For all three stations, the return period of a given stochastic snowfall has become longer for the most recent term. Through these results we have shown that local climatic characteristics can be detected and evaluated by statistical analysis of daily precipitation series.
