Abstract

The defining of weather forecasting regions is possible. since the representativeness of regional weather can by reasonably clarified in terms of weather entropy and the use of information ratio. In this paper, the weather entropy and information ratio were derived numerially from using the information theory. The typical weather characteristics were clarified and defined in the homogeneous weather forecasting regions of the southern parts of Korea. The data used for this study are the daily precipitation and cloudiness during the recent five years (1990-1994) at 42 stations in southern parts of Korea. It is divided into four classes of fine, clear, cloudy and rainy. The results are summarized as follows: 1. The maximum value of weather entropy in study area is 2.009 vits in Yosu in July, and the minimum one is 1.624 bits in Kohung in October. The mean value of weather entropy is maximal in July, on the other hand, minimal in October during four season. The less the value of entropy is, the stabler the weather is. While the bigger the value of entropy is, the more changeable the weather is. 2. The deviation from mean value of weather entropy in southern parts of Korea, with the positive and the negative parts, shows remarkably the distributional tendency of the east (positive) and the west (negative) in January but of the south (positive) and the north (negative) in July. It also clearly shows the distributional tendency of the east (postive) and the west(negative) in the coastal region in April, and of X-type (southern west and northern east: negative) in Chiri Mt. in October. 3. In southern parts, the average information ratio maximaly appear 0.618 in Taegu area in July, whereas minimally 0.550 in Kwangju in October. Particularly the average information ratio of Pusan area is the greatest in April, but the smallest in October. And in Taegu, Kwangju, and Kunsan, it is the greatest in April, January, and July, but the smallest in Jyly, July, and pril. 4.The narrowest appreance of weather representativeness is in July when the Kwangju is the center of the weather forecasting. But the broadest one is in April when Taegu is the center of weather forecasting. 5. The defining of weather forecasting regions in terms of the difference of information ratio most broadly shows up in July in Pusan including the whole Honam area and the southern parts of Youngnam when the Pusan-Taegu is the basis of the application of information ratio. Meanwhile, it appears most broadly in January in Taegu including the whole southern parts except southern coastal area.

한국은 아시아대륙의 북동쪽에 위치한 반도로서 삼면이 바다로 둘러싸인 독특한 주위지형을 이루고 있고, 산맥 등의 산악효과와, 반도의 속성인 바다-육지의 열용량 차이 및 종관규모의 대기운동의 영향 등을 받는다. 따라서, 반도의 대기흐름 양상은 지역차가 현저하므로 정확한 일기예보 및 일기예보 구역설정이 매우 어렵다. 그러나 각 지역의 일기대표성은 매월의 일기엔트로피와 情報此(information ratio) 개념을 기초로 하면 수리적, 지리적 요인 및 계절변화 등을 포함한 보다 恒常的인 모습으로 밝힐 수 있다. 이것은 각 지역이 갖는 일기엔트로피란 어느 기간중 매일 완전한 일기예측이 그 지역에 부여될 때의 일일 평균정보량이고 일기의 불확정성의 測度로서 일기일수의 치우친 비율을 나타내므로 결국 일기엔트로피 값의 대소는 일기출현의 특성을 잘 나태낸다. 그러므로 날씨 및 기후는 일기출현율과 그 월변화의 특성을 잘 반영하고 있다고 할 수 있다. 따라서 본 논문은 남부지방의 날씨 및 기후특성을 명확히 반영하고 잇는 탁월일기의 출현 다소와 월변화에 대해 정보이론을 이용하여 일기엔트로피와 情報此를 수량적으로 추출하고 그 공간스케일의 시간적 변동을 살펴 대표지점으로서 부산, 대구, 광주, 군산 지역의 일기대표성을 밝히고 그에 따라 남부지방의 일기예보구역을 설정한 것이다.

Keywords

• climatic characteristics;
• weather entropy;
• information ratio;
• information theory;
• defining forecasting regions

• 氣候特性;
• 일기엔트로피;
• 情報比;
• 情報理論;
• 預報區域設定;