A Study of Long-memory Analysis of Rainfall Time Series in Korea

There were two major purposes to be addressed in the analysis. First, it was of interest to evaluate the long-memory analysis. Second, we intended to examine the long-memory of monthly and annual precipitation in South Korea. Overall, the present study adds to the existing literature on possible alternatives. The long-memory in hydrologic time series can be analyzed by testing the null hypothesis that there is only short-term memory in the series. In this study, we have been focused on V/S test that can complement shortcomings of existing methods. And to compare with the V/S test we have been demonstrated long-memory test using the other 4 methods. They are Lo Test, KPSS Test, LobRob Test and Periodogram Test. To explore these methodologies, the AR(1) model is considered. As a results of this approach, the KPSS test and the V/S test provided a useful ground to test long memory. In this overall perspective, the each method was applied to the monthly and annual precipitation. We can identify that there is little evidence of long memory effect in monthly and annual precipitation time series.