Meteorological Drought Forecasting Using Stochastic Models

摘要

Drought is a normal part of climate and occurs in virtually all regions of the world. It is one of the major weather related disasters which is likely to continue for months, possibly years. It can affect large areas and may have serious environmental, social and economic impacts. Therefore drought forecasting plays an important role in the mitigation of impacts of drought on water resources systems. Accurate drought forecasts would enable optimal operation of reservoirs system during drought and would be helpful for decision maker. Due to the stochastic behaviour of drought, linear stochastic models known as Autoregressive Integrated Moving Average (ARIMA) and multiplicative Seasonal Autoregressive Integrated Moving Average (SARIMA) models were used to forecast droughts based on the procedure of model development. From the existing simple and most commonly used indices which used for the estimation of drought, the Standardized Precipitation Index, (SPI), seems to have a widespread applicability. The primary reason is that SPI is based on rainfall alone, so that drought assessment is possible even if other hydro-meteorological measurements are not available. The stochastic models presented in this study were fitted to the SPI index using SPI data sets relative to 47 years (1961-2007) in the Ruhr river basin in Germany. The predicted results show reasonably good agreement with the observed data. Results show that the forecast accuracy decreases with increase in lead-time, so the models could be used to forecast droughts up to 2 months of lead-time with reasonably accuracy.