Introduction

Undoubtedly, one of the areas in which drought has a great impact is the water resources of each country, and the areas that have the capability of drought are more limited and sensitive to water resources and require measures before, during and after the drought(Lain and Adamowisky,2012). While the dependence rate of agricultural lands equipped for irrigation on underground water sources is 37.8% on average in the world, the said coefficient is 46.2% in the Middle East region and especially in Iran it is equal to 62.1%. This is a confirmation of the high share of the agricultural sector of groundwater consumption in Iran compared to other countries and even the Middle East region (FAO, 2009). About 85% of the surface of Iran has a dry desert, semi-arid and ultra-arid climate (Ghafouri, 2003). Also, the annual rainfall of Iran is 240-260 mm and less than one third of the average annual rainfall of the world (870 mm). The phenomenon of drought during its occurrence period affects the underground water resources, which unfortunately has been less noticed. One of the important effects of drought is related to the drop of underground water, which due to the lack of rain and snow, there is a sharp decrease in the nutrition of the aquifers, and with the excessive exploitation of these aquifers, the condition of severe drop. Water resources are provided at the water level.



Materials and methods

The research area is located in the north of Lake Urmia and the watersheds overlooking the city of Tusuj in the coordinates of 45° 18’ to 45° 33’ east longitude, 38° 20’ to 38° 24’ north latitude. According to different classification methods, the climate of the region is cold, semi-arid, and it is considered a Mediterranean rainfall regime.

In this study, one of the important drought indices that is widely used in drought monitoring is the standardized precipitation index (SPI), which was presented by McKay et al.(1993) to determine the probability of drought occurrence. To model the monthly rainfall data, one of the suitable statistical distributions for this task is to use the gamma distribution. The reason for using this distribution is the nature of the statistical distribution itself, in which values close to zero have a frequency with greater probability, and the same nature is also valid for monthly rainfall data (the existence of months without rainfall). In this study, the correlation between the SPI index and the GRI standard groundwater index is investigated and the effect of drought on the GRI index is evaluated with a regression model.

Discussion and results

According to the results of the correlation coefficient table, the GRI index has a significant positive correlation with the SPI drought index in the time scale of 48 months for Tasuj station. For Sharafkhane station, the SPI index has a significant negative correlation at a significance level of 0.05 in the time scale of 12 and 24 months. For Khoi station, in all three time scales of 12, 24 and 48 months, there is a significant negative correlation at the significance level of 0.05 and even 0.01. The 48-month drought index of Tasouj, 12-month Sharafkhane and 24-month Khoi, which have the most significant correlation with the GRI index, were entered into the regression model as factors influencing drought to determine the total changes in GRI that were affected by the drought conditions of three stations, to be determined. Therefore, three drought indexes as three independent variables and GRI index as a dependent variable form a linear multivariate regression model.

According to the results, the correlation coefficient (R) and the coefficient of determination of the model (R Square) are 0.706 and 0.499, respectively, and this indicates that 49.9 percent of the changes in the model are determined or covered by independent variables. A value of 0.000 for Sig. In the analysis of variance (ANOVA) table, it shows the significance and validity of the model. By having regression coefficients, it is possible to reach the relationship between the GRI index and the drought indices of the stations. which is mentioned in relation 1. All coefficients of the model are significant (Sig.=0.000) and indicate the validity of these coefficients. Therefore, the regression relationship of the model can be shown as follows.

GRI=0.240+0.923×SPI48(tasouj)-0.235×SPI12(sharaf)-0.559×SPI24(khoy) (1)



Conclusion

The results of the study of the effect of drought on the underground waters of the Tusuj aquifer area indicate a significant correlation between the 24-month drought profile of the Tusuj station, 12-month Sharafkhane and 24-month Khui stations with the GRI standard groundwater profile of the Tusuj aquifer area. The downward trend of the GRI groundwater standard index is consistent with the downward trend of the drought of the Tasuj rain gauge station in three time scales of 12, 24 and 48 months. Of course, despite the fluctuations of the SPI drought index, especially the short-term increase and the occurrence of insignificant drought in some years in Sharafkhaneh and Khoi stations, but due to the accumulation of drought and the dominance of drought years over drought years on the long-term climate memory The region has been affected and has caused a decrease in groundwater and a downward trend in the GRI standard index.

For a more accurate statistical analysis, a suitable regression model was determined between the GRI index as a dependent variable and three drought index variables of 48 months of Tasouj, 12 months of Sharafkhane and 24 months of Khoi as dependent variables. which is given in relation (1) as the result of the regression equation. By using this relationship, it is possible to calculate the values of the corresponding GRI standard index by having the SPI drought index values.