Investigations for typhoon disasters in South Korea and comparative research on flood disaster mechanism between Japan and South Korea

2002年，台风“罗莎”给朝鲜半岛带来了历史上最大的降雨，造成了韩国有史以来最严重的洪灾。2003年，台风MAEMI袭击了韩国南部地区。损失估计为5.6万亿韩元，这是韩国有史以来最大的损失。2004年，10个台风袭击了日本，这是有观测以来的最高记录。今年，活跃的降雨锋也在新泻-福岛和福井地区造成了严重的洪水灾害。如前所述，两国近年来连续遭受严重洪灾。本研究从自然科学和社会科学的角度，对气候、地形、生活方式、人口密度等方面相似的两国洪涝灾害发展机制进行了比较研究。同时，我们研究小组开发的分布式水文模型应用于韩国流域，以澄清水文模型在两国的可转移性。此外，为了提高短期洪水预报的准确性，提出了一种基于分布式降雨径流模型的实时洪水预报方法。两国都有先进的地形、水文信息等数据库系统。实时水文观测系统包括雷达降水临近预报。为了充分利用水文观测信息，提出了一种考虑预报误差的随机实时分布式洪水预报方法。

In 2002, Typhoon RUSA brought historical rainfall to the Korean Peninsula, which caused the largest-ever flood disaster thought South Korea. In 2003, Typhoon MAEMI hit the south parts of South Korea. The damage was estimated to 5,600 billion won, which is the largest-ever loss in South Korea. In 2004, 10 typhoons hit Japan, which is the highest record since the observation began. In this year, active rainfall fronts also caused heavy flood disasters at Niigata-Fukushima and Fukui areas. As stated, both countries suffered severe flood disasters consecutively in resent years.In this research, flood disaster development mechanisms in both countries with similar climate, topography, life style, population density etc. were investigated comparatively through physical and social scientific view points. At the same time, a distributed hydrologic model developed in our research group was applied to Korean catchments to clarify the transferability of the hydrologic model in both countries. In addition, to improve the accuracy of short-term flood predictions, a real-time flood forecasting method incorporating a distributed rainfall-runoff model is newly developed. In both countries, advanced database systems for topography, hydrologic information etc. are well developed. A real-time hydrologic observation system including radar rainfall nowcasting is also available. To fully utilize the hydrologic observation information, a stochastic real-time distributed flood prediction method with a rainfall forecasting considering its prediction error is proposed.