P2.1
Relationships between Changes in Annual Frequency of Heavy Precipitation in Japan and ENSO

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 1 February 2006
Relationships between Changes in Annual Frequency of Heavy Precipitation in Japan and ENSO
Exhibit Hall A2 (Georgia World Congress Center)
Hironori Higashi, National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Ibaraki, Japan; and T. Matsuura

Poster PDF (254.0 kB)

Climate changes in global scale such as the El Nino-Southern Oscillation (ENSO) and global warming have significant impact on local and regional hydrological regimes, which affect ecological, social and economical system. A key parameter in basin water managements is the characteristic of local precipitation, and an accurate estimate of the climate-change impacts is needed to prevent flood disasters. Climate-change impacts have been investigated mainly using global circulation and regional climate models. Some of their outputs are useful for mean local climate: the annual, seasonal precipitation, and typhoon track in monsoon Asia changes due to ENSO. However, the models cannot capture the short-term and basin-scale precipitation accurately. This study describes temporal, regional and statistical changes in the observed precipitation in Japan for a period from 1961 to 2002. Relationships between ENSO, which is one of the global climate changes, and the local characteristics of annual precipitation, seasonal precipitation, and annual frequency of heavy precipitation events are discussed.

Daily precipitation records during 42 years (1961-2002) at 125 observational stations of the Japan Meteorological Agency are used to analyze long-term changes in precipitation. Annual and seasonal precipitations in each station are calculated using these data. Heavy precipitation is defined as the 2-day precipitation exceeding threshold. The threshold values are set to be the minimum 2-day precipitation in the annual exceedance series at each station, considering with the usefulness for the prevention of regional-scale flood disasters.

The time series of the annual and the seasonal (May-Oct.) precipitation are analyzed with the Mann-Kendall trend test. The slopes of the trends are calculated by least square linear fitting(Fig. 1). Both the annual and the seasonal precipitations tend to increase in East Japan, and to decrease in Central and West Japan. The annual frequencies of heavy precipitations are also analyzed. The trends of the frequencies are similar to those of the annual and the seasonal precipitations. Almost all heavy precipitation events in each station occurred in Baiu and typhoon seasons (May-Oct.). The annual amounts of the difference between the annual and the heavy precipitations are found not to increase or decrease at any stations. The trends of annual precipitation are mainly the results of the changes in the annual frequency of the heavy precipitations.

To investigate the cause of the trends, the singular value decomposition (SVD) analysis is carried out. First, the data in the SVD analysis are the annual precipitations and the average of sea surface temperature (SST) in Nino 3, Nino 4, and Nino West regions during May-Oct. in each year. The correlation coefficient between the expansion coefficient for the annual precipitation and that for SSTs is very small. The result is similar to many previous studies. Next, the data sets of the annual frequency of the heavy precipitations and SSTs are used(Fig. 2). The change in normalized expansion coefficient for the annual frequency of the heavy precipitations is similar to that for SSTs. The expansion coefficient for SSTs in the first SVD mode, which accounts for 68.2% of the squared covariance fraction (SCF), indicates the periodical changes of ENSO. The 3.6-year period is obviously detected by the spectrum analyses. The expansion coefficient for SSTs in the second SVD mode, which accounts for 26.1% of SCF, shows the trend of SSTs. In the heterogeneous correlation pattern for the second SVD mode, the correlation coefficients between the annual frequency of the heavy precipitations and the expansion coefficient for SSTs tend to be positive in East Japan and to be negative in Central and West Japan. The distribution of positive/negative correlation corresponds to that of the increase/decrease trend of the annual frequency of the heavy precipitations. It is revealed that the changes in the heavy precipitation in Japan are related to ENSO.