Search

In recent, abnormal phenomena in nature occurs due to climate change, which result in increase of flood frequency and rainfall intensity by change of meteorologic and hydrological factors. Especially, flood inundation in urban watershed frequently happens on casualties expected and property damage. Scientific and technological adaptation are acutely needed for unexpected damage in nature, and that, it is considerably important to develop proper model of hydrological forecasting over the short periods and to build a real-time watershed monitoring system. One of important things for solving a variety of problems is to connect real-time watershed monitoring system with short term hydrologic forecasting model. In this research, Suyoung stream in Busan was determined as a study area. After that, real-time watershed monitoring system at ten minute interval has been established for flood warning and forecast in the field. The data from this system was applied to short-term hydrologic forecasting model based on the artificial neural network. When flood happens in urban, input data of forecasting model on each rainfall event was used. In conclusion, after establishing the model being possible to forecast water level from 10 to 60 minutes at 10 minute interval, as a result of simulating the model, the model performances are considerably good as RMSE 0.02~0.6 and more than 0.9. It makes possible effectively to manage the watershed in urban.

Studies for predicting climatic factors using GCM-RCM methods have limited applications to various area. It also showed different simulated results by each of runoff models. It is possible to make some problems for reliability of results in case of applying to runoff model, as long as uncertainty of effect assessments to runoff models is great. In this study, prediction of hydrological factors over the mid-term with a concept of non-scenarios on the 7 sites in Han river was conducted based on observed hydrological data from the past to the present. This study is not used to GCM scenarios and applies to statistical models like Sen test of trend analysis methods based on non-scenarios in seven points of Han river. The trend and elasticity analysis were performed after detecting the characteristics in each of the sites using statistical methods with observed annual rainfall and runoff from 1966 to 2001. Using the results, prediction of hydrological factors (2013, 2015, 2017, 2020) was carried out by trend analysis and elasticity method. The prediction result using the elasticity method of the results of this study was more reliable than result from trend analysis.

Climate change occurring globally destroys the balance of equilibrium of water cycle connected to precipitation, ground water recharge, outflow, storage, and evapotranspiration. IPCC said that change of precipitation pattern generated variation of climate system in the weakest part among the impacts of future climate change on human environment so researches on the tendency of runoff and precipitation for the variation of characteristics are carried out globally. These have studied a wide of tendency for analyzing the variation of precipitation characteristics due to climate change in Korean. By collecting observed precipitation and runoff data from 1966 to 2001 at 24 points in the country, extracting at the units of year, these were used as basic data for predicting mid-term runoff by 2020. The result of mid-term hydrological prediction from observed data was the better than other and more effective to predict. This study on the basis of long term observed hydrological data contrary to GCM-RCM method on the primary purpose is to develop hydrological prediction model using non-scenario methods over the mid-term.