Investigation of Trend Variations in Annual Maximum Rainfalls in South Korea

摘要

Hydrologic frequency analysis practically assumes that a time series of interest is temporally stationary. However, the assumption is sometimes not valid within the context of climate change. Monotonic trend tests, which are widely used in practice to investigate the presence of significant non-stationarity of hydrologic variables, can determine only whether a trend is significant. This study proposed a practical method for characterizing trend pattern embedded in hydrologic observations through the efficient illustration of changes in trend significance over time. This study analyzed the trend patterns of observed rainfalls in the Han River basin using Mann-Kendall test, and developed checker board plots to show the pattern of trend variations. The results indicate that the time and the magnitude of significant trend would differ by station although most of record periods show a certain trend of increasing annual maximum rainfalls.