Nationwide Prediction of Drought Conditions in Iran Based on Remote Sensing Data

摘要: Iran is a country in dry part of the world and extensively suffers from drought. Drought natural, temporary, iterative phenomenon that caused by shortage rainfall, which affects people's health well-being adversely as well impacting society's economy politics with far-reaching consequences. Information on intensity, duration, spatial coverage drought can help decision makers to reduce vulnerability drought-affected areas, therefore, lessen risks associated episodes. One major challenges modeling (and short-term forecasting) unavailability long-term meteorological data for many parts country. Satellite-based remote sensing dataathat are freely availableagive information vegetation conditions land cover. In this paper, we constructed artificial neural network model forecast) based satellite imagery. To end, standardized precipitation index (SPI) was used measure severity. A number features including normalized difference (NDVI), condition (VCI), temperature (TCI) were extracted NOAA-AVHRR images. The received these input outputted SPI value (or condition). Applying stations available, showed it could forecast an accuracy up 90 percent. Furthermore, TCI found be best marker among satellite-based features. We also multilayer perceptron better than radial basis function networks support vector machines forecasting conditions.