Document Type : Research Paper
Authors
1 Associate Professor, Faculty of Agriculture, Isfahan University of Technology, Iran
2 MSc Student, Faculty of Agriculture, Tarbiat Modares University, Iran
3 Professor, Faculty of Agriculture, Isfahan University of Technology, Iran
Abstract
The most important factor of drought analysis is rainfall. All resources which are supported by rainfall, such as, surface streams; groundwater and lakes, are affected by deficit of rainfall. Analysis of past rainfall statistics indicates a pattern. Forecasting future rainfall can be used in water resources planning. This opportunity can be provided by time series patterns. The aim of this research is drought forecasting of Isfahan Province, by fitness of various models with rainfall data. Studied models are included in monthly, monthly-seasonally, annual-exponential, moving average (3, 5, 7) and sinusoidal. 24 meteorology stations have been studied in the time series analysis. The at-site statistic period ranges between maximum 44 years (1957-2000) to minimum 11 years (1987-1997). By regarding ARIMA model fitness, the results indicate that the studied stations are divided to three groups as high, medium and low quality. These groups are included 9, 5, 4 stations, respectively. Although Isfahan and Kashan have been owned long-term rainfall data, they have less suitable fitness than the other stations. Therefore, by only having long-term rainfall data, it cannot be expected the better fitness of time series. Using ARIMA for annual moving average modeling can be effective in drought pattern analysis and have been presented the good patterns.
Keywords
)