Journal
MODELING EARTH SYSTEMS AND ENVIRONMENT
Volume 6, Issue 1, Pages 63-71Publisher
SPRINGER HEIDELBERG
DOI: 10.1007/s40808-019-00655-2
Keywords
Statistical analysis; Risk; RBF and ANFIS models; Simulation; Fuzzy
Categories
Funding
- Mohaghegh Ardabili University
Ask authors/readers for more resources
Drought is one of the natural hazards that has hit Iran over the past decades with serious environmental hazards, including those in southern parts of Iran. Researches done in the southern region of Iran in the field of statistical modeling of drought are rare. Therefore, the aim of this study is to fuzzify the SMS index, modeling and forecasting droughts in the southern part of Iran. For this study, 29-year-old temperature and precipitation data were used in 28 synoptic stations in the southern part of Iran during the period 1908-2018. In this study, three drought indexes SPI, MCZI, SET were separately calculated and combined and the fuzzy index SMS was obtained. Then, in ANFIS and RBF neural network models were compared and modelled in MATLAB software and simulated for the next 16 years, and finally, using the TOPSIS multivariate decision-making model, the drought-affected areas for the coming years, 16 next years, they were prioritized. The findings of the study showed that the new fuzzy index of the three indicators reflected drought with acceptable accuracy. In assessing the two models of ANFIS and RBF, the RBF model with a RMSE value of 1.15 and R-2 values of 0.9161 have the highest accuracy than the ANFIS model for prediction. According to the SMS fuzzy index, stations such as Kerman, Yasuj and Abadan with drought of 0.69, 0.97, and 0.89, respectively were exposed more to future drought. Also, based on Topsis model, central and northern stations such as Koohrang and Safashahr with drought of 0.19 and 0.21, respectively, were subjected lower to drought in the following years.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available