نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجوی دکتری /اقلیم شناسی دانشگاه اصفهان و عضو گروه اقلیم شناسی کاربردی پژوهشکده اقلیم شناسی
2 عضو هیات علمی /دانشکده کشاورزی دانشگاه فردوسی مشهد
3 عضو هیات علمی /پژوهشکده اقلیم شناسی و رییس پژوهشکده اقلیم شناسی و مرکز ملی اقلیم
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Seasonal rainfall forecasts can effectively be used for resources planning and management - e.g., reservoir operations, agricultural practices, and flood emergency responses. Effective planning and management of water resources in the short term requires a general view of the upcoming season. In the long term, this needs realistic projections of scenarios for future variability and changes. In this paper, 33 years of rainfall data in the Khorasan region, northeastern Iran was analyzed. The study area is situated at 31°-38°N, 74°- 80°E. This synoptical data was trained by the Mamdani fuzzy Inference system and the artificial neural network. For performance evaluation, predicted outputs were compared with the actual rainfall data. First, synoptical relationships were investigated i.e. Sea Level Pressure (SLP), Sea Surface Temperature (SST), Sea Surface Pressure Difference (DSLP), Sea Surface Temperature (DSST) and Air Temperature at 850 hpa, Geopotential Height at 500 hpa, and Relative Humidity at 300 hpa. Models were then calibrated for the period of 1970 to 1992. Finally, the rainfall is predicted. Simulation results revealed that the Mamdani fuzzy Inference system and artificial neural networks are both promising and efficient. The root mean square for Mamdani fuzzy Inference system and the artificial neural network were 52 and 41 millimeters, respectively.
کلیدواژهها [English]