STUDY ON JUMPS AND TRENDS IN ANNUAL TEMPERATURE AND PRECIPITATION TIME SERIES IN JAPAN

Abstract

The possible trend and jump in hydrologic time series, due to climatic change, is investigated in this paper. The detection is accomplished by using both parametric and nonparameiric statistical tests. The analysis is made to two regionalized datasets: 115-year annual temperature over 6 climatologic stations and 103-year annual precipitation over 46 rain gauges in Japan. Both long-term trend and jump temperature data are detected, and the results indicate that the annual temperature increases 1.33-Celsius degree due to climatic change in Japan over the past century. Although it still needs further study on the variability of precipitation te series, both two kinds of above variations in annual precipitation seem also exist. The relationship between the current length of hydrologic time series and the possible length of time series required for detecting jump is investigated. These conclusions may provide some useful insights into the issues related to the hydrologic frequency analysis.