Journal of Spatial Analysis

---

Drought is the most important natural disaster, due to its widespread and comprehensive short and long term consequences. Several meteorological drought indices have been offered to determine the features. These indices are generally calculated based on one or more climatic elements. Due to ease of calculation and use of available precipitation data, SPI index usually was calculated for any desired time scale and it’s known as one of the most appropriate indices for drought analysis, especially analysis of location. In connection time changes, most studies were largely based on an analysis of trends and changes in environment but today special attention is to the variability and spatial autocorrelation. In this study we tried to analyze drought zones in the North West of Iran, using the approach spatial analysis functions of spatial statistics and detecting spatial autocorrelation relationship, due to repeated droughts in North West of Iran and the involvement of this area in the natural disaster.

In this study, the study area is North West of Iran which includes the provinces of Ardebil, West Azerbaijan and East Azerbaijan. In this study, the 20-year average total monthly precipitation data (1995-2014) was used for 23 stations in the North West of Iran. In this study, to study SPI drought index, the annual precipitation data of considered stations were used. According to the statistical gaps in some studied meteorological stations, first considered statistics were completed. The correlation between the stations and linear regression model were used to reconstruct the statistical errors. Stations annual precipitation data for each month, were entered into Excel file for the under consideration separately and then these files were entered into Minitab software environment and the correlation between them was obtained to rebuild the statistical gaps. Using SPI values drought and wet period’s region were identified and zoning drought was done using ordinary kriging interpolation method with a variogram Gaussian model with the lowest RMS error. Using appropriate variogram, cells with dimensions of 5_×5km were extended to perform spatial analysis on the study area. With the establishment of spatial data in ARC GIS10.3 environment, Geostatistic Analyze redundant was used to Interpolation analysis Space and Global Moran's autocorrelation in GIS software and GeoDa was used to reveal the spatial relationships of variables.

The results showed that most studied stations are relatively well wet and this shows the accuracy of the results of the SPI index. Validation results of the various models revealed that Ordinary Kriging interpolation method with a variogram Gaussian model best explains the spatial distribution of drought in North West of Iran. So, using the above method the stations data interpolation related to SPI index in North West of Iran was done. The results showed that Moran index values for the analysis of results of standardized precipitation index (SPI) in all studied years, is more than 0.95. Since Moran’s obtained values are positive close to 1, it can be concluded that drought, in the North West of Iran during the statistical period has high spatial autocorrelation cluster pattern of 90, 95 and 99 percent. Results also showed that in all the years of study, Moran's global index is more than 0.95 percent. This type of distributed data suggests that spatial distribution patterns of drought in North West of Iran changes in multiple scales and distances from one distance to another and from scale to another and this result shows special space differences in different distances and scales in this region of the country. Results also showed that drought in North West of Iran in 2008 is composed of two parts: Moderate drought in parts of West and North West region (stations of Maku, Khoy, Salmas, Urmia, naghadeh, Mahabad and Piranshahr) and severe drought in the southeastern part of the study area (stations: Sarab, Khalkhal, Takab, Tabriz and Mianeh). So the pattern of cluster drought in the North West of Iran in 2008 is on the first and fourth quarter. The results of this index showed that drought and rain periods are similar in the studied stations. The results of the application of Moran's index about identifying spatial distribution of drought patterns showed that The values of the different years during the period,  have a positive a positive coefficient close to 1 (Moran's I> 0.959344) and this shows that the spatial distribution of drought is clustered. The results of the standard score Z values and the P-Value proved the clustering of spatial distribution of drought.

The results of the analysis of G public value, In order to ensure the existence of areas with clusters of high and low values showed that The stations of Maku, Khoy, Salmas, Urmia, naghadeh, Mahabad, Piranshahr and Parsabad follow the moderate drought pattern in the region and are significant at the 0.99 level. Jolfa station also has a mild drought of 0.95 percent confidence level and for Sardasht station is significant in 0.90 percent. High drought pattern in Sarab, Khalkhal, Takab, Tabriz and Mianeh stations was significant in 0.99 percent level and also for Ardabil, Sahand and Maragheh stations very high drought pattern was significant in 0.95 percent level and for Meshkinshahr and Ahar high drought pattern is significant in 0.90 percent. By detection of clusters of drought and rain in the North West of Iran using Moran’s spatial analysis technique and G general statistics a full recognition of the drought affected areas in this region can be obtained and take the necessary measures in its management 

---

Drought is a natural occurrence that occurs repeatedly or alternately and is likely to occur in almost every kind of climatic event. Also, the distinction between this phenomenon and other natural disasters is that unlike other disasters, this phenomenon gradually over a relatively long period of time to act and its effects may be delayed after a few years and more than any other natural disaster appears. Several indicators have been presented to decide the characteristics of hydrological and meteorological drought. These indicators are generally based on one or more climatic elements. The SPI and SWI indicators are similar in terms of ease in calculations and results, and use monthly precipitation data and monthly spatial data rates. The simultaneous effect of meteorological droughts on groundwater levels rarely happens. Therefore, the present study investigates the effect of meteorological droughts on the groundwater level of Marand plain and calculates the time delay of drought on groundwater level.
The study area in this study is Marand Plain in East Azarbaijan Province. In this research, we used meteorological data (average monthly rainfall) of 7 rain gauge stations during the statistical period (1980-2012), and the monthly water level data of 23 piezometric wells during the statistical period (2001-2011). The correlation between stations and piezometric wells and linear regression method was used to reconstruct the statistical defects, then SPI and SWI indices were used to study the rainfall and groundwater changes process and the analysis of drought conditions in the meteorological and underground watersheds. The SPI index is basically calculated for periods of 3, 6, 9, 12, 18, 24, and 48 months. Also, the standardized water level indicator (SWI) has been used as a criterion for assessing occur drought and wet years in the Marand plain. The purpose of the SWI index is to allow zoning of groundwater level fluctuations at the study area. Extraction of drought and wet year intensities in different scales and basin zonation for drought maps in Marand plain was first calculated by entering the monthly values in DIP software, SPI values for 12-month time series. SWI values were calculated from monthly data of piezometric stationary level surfaces, such as SPI values, with the help of DIP, Minitab and Excel software. Geostatistical Analyst was also used to decide the weather drought and groundwater drought periods for the ArcGIS software.
 
 
The results of the SPI values showed that meteorological drought is not of a definite local place, while groundwater droughts have not occurred randomly in the area and its concentration in the west of the aquifer is more than the east. Considering the increase in the area under cultivation, to compensate for the water needs of agricultural lands, an increase in the harvesting of underground water table has occurred in order to compensate for the need for water, indicating a tangible relationship between the rainfall and the level fluctuation in the Marand plain. Therefore, considering the increase in the area under cultivation during the years of drought in the region, the best correlation between them was -0.720 with a delay of 5 months, in order to investigate the effects of drought on the surface of the station, which was significant at 1% level It illustrates the impact of groundwater resources with a 5-month delay. Also, the results of the survey of monthly data of Marand plain surface during the statistical period (2001-2011) showed that the groundwater level of the plain had a negative trend that fell by about 2 meters.