Mendeley  

Zotero  

Kheyri R, Mojarrad F, Masompour J, Farhadi B. Evaluation of drought changes in Iran using SPEI and SC-PDSI. MJSP 2021; 25 (1) :143-174
URL: http://hsmsp.modares.ac.ir/article-21-46543-en.html

Introduction
Drought is one of the extreme climatic phenomena that occurs in all regions of the world, especially in arid and semi-arid regions with different intensities. In recent decades, drought has caused great damage to the economic, social, agricultural and water resources sectors. Numerous studies have been conducted in this field, each of which has achieved remarkable and practical results, and in most cases, have provided constructive solutions to adapt to this natural disaster. The purpose of this study is to monitor drought in the country using Self-Calibrating Palmer Drought Severity Index (SC-PDSI) and Standardized Precipitation Evapotranspiration Index (SPEI) indices, and to evaluate the validity of data from two global drought databases for the study of these two indices.
Methodology
In the present study, data on precipitation, minimum temperature, maximum temperature, average temperature, wind speed and sunshine hours of 101 stations in a 24-year period (1992-2015) have been used. After reviewing the data, two indices, SC-PDSI and SPEI were calculated using SC-PDSI and SPEI packages in R software, and then percentage maps of years with each drought class were drawn with ArcGIS software. The Penman-Monteith method was used to calculate evapotranspiration in the SPEI index. In the next step, by comparing the computational values of the indices with the values provided by the two global databases of the Climate Research Unit (CRU) of the University of East Anglia at https://crudata.uea.ac.uk/cru/data/drought/ and The Global SPEI database at https://spei.csic.es, which provide drought indices with a resolution of 0.5 by 0.5, the accuracy and validity of the data of these two databases were investigated. It should be noted that the computational values of these two indices in 101 stations were compared with the data of 99 points obtained from the database, which were the closest distances to the observation stations. These points were determined using the Euclidean distance method.
Results and Discussion
SPEI values in 3, 6, 12, and 24-month time periods showed that in all four time periods, more than 60% of the years were in the normal class, and less than 5% of the years were in the extreme drought class. The maximum cores of the two normal and extreme drought classes are located in the eastern and western parts of the country, respectively. According to the SC-PDSI, the highest percentage of years are in the normal class, and less than 4% of the years are in the extreme drought class, with the exception of a few very limited maximum cores nationwide. According to this index, the concentration of drought maximum cores has been related to the western, northwestern and to some extent the center of the country. It is noteworthy that there is a maximum core of 13% in the vicinity of Bandar Anzali. The location of the rainiest station in the country (Bandar Anzali) in the extreme drought class indicates a severe negative anomaly in the rainfall of this station. The beginning of the drought period in the SC-PDSI index was in 1999 and in the four time periods of the SPEI index in 1999 and 2000. Comparison of shows that with the exception of a few stations, there is no specific coordination between these two data sets. Thus, based on the calculated values, in all stations of the country, a higher percentage of years (48-76% of years) is in the normal class. However, according to the database values, in most stations, 50% of the years are in the normal class, and the rest of the years have been affected by mild to extreme droughts. Database index values also place more years in the extreme drought class than computational values. Comparison of computed and database extracted SPEI values also shows a similar inconsistency. According to the calculated values, more than 60% of the years are in normal condition in all stations; but in the database, the percentage of normal years at the stations fluctuates a lot and its range varies from 20 to 80% in different periods. Although the extreme drought class has formed numerous maximum cores in all periods, it has been concentrated in the western, southwestern and central regions of the country in less than 5% of the years. In the eastern half of the country, the number of years with extreme drought has been low and even zero in some cases. What can be deduced from the analysis of database values is the existence of maximum cores in the western and southern regions of the country in periods of 6, 12 and 24 months, which shows that under 20% of the years of these regions are under the control of extreme drought. In general, the values of the two indices in both databases have always been higher than the calculated values and have shown the drought to be more severe.
Conclusion
Based on the general results of this research, more than 60% of the years are in normal condition in all stations in both indices. Normal cores are located in the eastern parts of the country and drought cores are located in the western half of the country and rainy areas. The pattern observed in the maps indicates that severe and extreme droughts have occurred in parts of the country that have relatively better rainfall conditions, and changes in the amount of rainfall have led to more severe droughts.
For example, although the western half of the country receives relatively more rainfall than other regions, a slight decrease in rainfall and an increase in evapotranspiration have caused extreme droughts in at least 5% of the years. Another obvious example is the existence of the core of the extreme drought at Bandar Anzali station. Comparison of the results obtained from the calculations and the database shows that there is no coordination and spatial overlap between them in terms of drought severity and there are obvious differences in the percentage of years of each drought class. However, despite the differences between the calculated and database values, there is good temporal coordination between them.