Application multivariate regression in water resources availability forecasting upstream of red river

Abstract

      Water resources availablity of a basin is highly influenced by many factors such as local rainfall, terrain,  and water use. Water resources availability prediction plays a key role in managing this important resources to prevent natural disaster such as floods, droughts and support the development of all life aspects such as economic, tourism, ecology, etc. This paper presents an application of multivariate regression in water resources availability forecasting upstream Red River. In this method, mathematical equations represent the relationship between climatic factors (rainfall, temperature, evaporation, humidity, atmospheric circulation, the high air surface pressure in the Pacific, ENSO phenomenon, water storage period of early season, end of season) and predictor (maximum flow, flood volume, minimum flow, flow volume in dry season). Based on these equations, in year 2011-2012, for volume of flood season and volumen of dry season prediction, the confidence level of forecast is 65%; for Q_max year and Q_min year, it is 80%.