097号 page 101/212

このページは 097号 の電子ブックに掲載されている101ページの概要です。
秒後に電子ブックの対象ページへ移動します。
「電子ブックを開く」をクリックすると今すぐ対象ページへ移動します。

概要：
Historical Occurrence of El Nino and its Impact on Food Crop Production at a Regional Level in Asia and the Pacific 91qualitatively and quantitatively, is difficult (World Meteorological Organization,....

Historical Occurrence of El Nino and its Impact on Food Crop Production at a Regional Level in Asia and the Pacific 91qualitatively and quantitatively, is difficult (World Meteorological Organization, 1999).A quantitative definition of El Nino, originally proposed by the Japan MeteorologicalAgency, then modified by the Climate Variability and Predictability (CLIVAR) project,gives five-month running means of SST anomalies in the Nino 3.4 region (5°N-5°S,170°W-120°W) that exceed 0.4°C for six months or more, based on accepted conceptsand designed to be consistent with previous recognized events (Trenberth, 1997). Themultivariate ENSO Index (MEI) was devised to measure the strength of ENSO eventswhich is expressed in the first principal component of six observed variables over thetropical Pacific: sea-level pressure (P), zonal (U) and meridional (V) components ofthe surface wind, sea surface temperature (S), surface air temperature (A), and totalcloudiness fraction of the sky (C) (Wolter and Timlin, 1993). These two indices are wellaccepted for overall monitoring of the ENSO phenomenon. However, if one’s interestin ENSO is to a specific part of the world, it is recommended to establish another indexto fit one’s needs (Trenberth, 1997; comment by Wolter at MEI Web site, 2002).For our study area, southeast Asia and Australia, the SOI is known as a goodindicator of rainfall (Podbury et al., 1998; Yoshino et al., 2000; Nicholls and Beard,2000; Tawang et al., 2002; and Naylor et al., 2002). Consistently negative and rapidlyfalling SOI phase is associated with dry spells (Stone and Auliciems, 1992). Basedon this finding, we specified El Nino years as when the annual average of SOI waslower than the standard deviation (-7.28), in the period of 1961-2000, and the annualaverage of SOI substantially (more than 2.00) declines from that of the previous year.Accordingly, since 1961 we can identify eight El Nino years as 1965, 1972, 1977,1982, 1987, 1991, 1994 and 1997. This specification is mostly identical with those ofCLIVAR and MEI (Trenberth, 1997; Wolter and Timlin, 1998).3. Methods and dataTo clarify the impacts of El Nino-induced abnormal weather conditions on totalcereal production in the region, we ran simple linear regressions of production, areaharvested, and yield against time, incorporating a dummy variable for El Nino years.Production variability was measured for each crop by percentage deviation froma five-year moving average as trend, assuming no technological progress took place inthe five-year period (Gommes, 1998, measured production loss based on seven-year