Journal of Meteorology and Atmospheric Science

Forecasting Drought Situation by Standardized Precipitation Evapotranspiration Index and Artificial Intelligence Models(Case Study: Kerman Synoptic Station)

10.22034/jmas.2022.356011.1184
Volume 4, Issue 3
Autumn 2021
Pages 260-271
Journal of Meteorology and Atmospheric Science

Document Type : Original Article

Authors

Mahshid Amiranipur 1
mohammad najafzadeh 2
Sedigheh Mohamadi 3

1 Student/University of Advanced Industrial and Technological Education in Kerman

2 Professor/Department of Water Engineering, Faculty of Civil Engineering and Mapping, Graduate University of Industrial and Advanced Technology, Kerman

3 Department of Ecology, Research Institute of Advanced Science and Technology and Environmental Sciences, Graduate University of Industrial and Advanced Technology

Abstract
Drought is one of the natural calamities that causes a lot of damage to natural ecosystems. According to the results of the researches, the things that play a special role in the occurrence and continuation of drought are precipitation, temperature, evaporation, wind and relative humidity, but precipitation is the most important in The occurrence of drought. In this research, the data of maximum temperature, minimum temperature, average temperature, relative humidity, precipitation and wind speed of Kerman synoptic station were used in a period of 29 years (from 1991 to 2020). Standardized evapotranspiration (SPEI) and artificial intelligence methods such as tree model (MT) and spline adaptive multivariate regression (MARS) are presented. The results of Kerman drought prediction using the standardized precipitation-evaporation and transpiration index showed that the frequency of wet and dry periods is high in short-term time scales and this frequency decreases with the increase of the time period. In the period of the first 6 months of 1992, the highest value of the drought index (SPEI=2.48) and the most severe droughts occurred in the period of 24 months ending in 2009 with the lowest value of the index (SPEI=-5.53). The performance of MARS artificial intelligence model according to the values of R, RMSE and MAE statistical indicators in the training stage (R=0.989, RMSE=0.148 and MAE=0.105) and testing (R=0.950, RMSE = 0.290and MAE = 0.158), it is more suitable than the MT model.

Keywords

Drought
SPEI index
Artificial intelligence
Model tree
Multivariate adaptive regression spline
Subjects

Home

Guide for Authors

Submit Manuscript

Reviewers

Contact Us