An analysis of the migration of cities to the center of the province; Causes and trends (Case study: West Azerbaijan province)

Forecasts indicate a significant increase in population in urban centers. Larger urban centers and metropolitan areas, such as magnets, often attract this population through various forms of migration. This growing flood of migration to large cities, which is usually faster and more intense in third-world countries, creates many problems at the source and destination, including the need for more urban and public services in the destination and the shortage and loss. One of the main destinations of immigrants from villages and small towns is to move to the city center of the province, a big city that acts as a metropolis in the region.One of the provinces in Iran where such migration flows are observed is the West Azerbaijan province, where research on the trends and reasons for the migration of its cities to the center of the province seems necessary. In this research, the following questions have been answered:What has been the process of emigration to the province's center during the period of 85-90 and 90-95?Spatial distribution of migratory data on which of the following spatial distribution patterns was consistent?What factors affect the emigration of the city population to the center of the province?

The present study is a descriptive-analytical research with an applied approach. The data used from library resources such as statistical yearbooks of the province were extracted from the Statistics Center of Iran, and in their analysis, spatial statistics and inferential statistics were used. Spatial statistics analysis is performed in the ArcGIS software environment. In this regard, first, the immigration رده بندی maps of the cities were drawn according to the statistics published by the National Statistics Organization in the two periods of 85-90 and 90-95.Data analysis was performed in two parts: trends and causes of migration. Regarding the migration process of cities to the center of the province, first, by Moran spatial autocorrelation statistics, the type of spatial distribution of migration data was investigated and hot spots of migration were identified by using hot spot analysis. Then, the factors affecting the migration of cities to the center of the province in the form of independent variables were modeled using geographically weighted regression (GWR) analysis.

Salmas city, with 3776 people, had the highest, and Takab city, with 333 people, had the lowest number of emigrants to the center of the province. The rate of emigration from other cities was between these two cities.According to the value obtained for the Moran index, which is positive and significant, the migration data of cities have spatial autocorrelation and cluster spatial distribution pattern and are not randomized or scattered. Hot spot analysis was used to identify places with high-value clustering or hot spots. The results showed that in the period of 90-95, high amounts of emigration in Khoy and Salmas counties had formed hot spots.However, the correlated t-test showed no statistically significant changes in the rate of emigration in the mentioned period. The variables considered and related to the theoretical foundations and research background in the form of geographically weighted regression analysis have been affected by the dependent variable, the migration rate to the province's center. In this regard, five variables of population density, distance of origin and destination, the number of urban services, the urbanization coefficient at the origin, and the number of job applicants at the origin entered the regression model. The model's results were evaluated with goodness of fit (GOF) indicators. The value of R2 for the whole model is 0.773438, and the value of Local R2, for other complications (cities), varies from 0.772427 to 0.773186, which is close to the value of 1, and in general, shows the goodness of the regression model.The regression model has been able to explain about 65% of the changes in the dependent variable. In the next step, by performing Moran autocorrelation analysis on the residual values of the regression model, the residual spatial distribution pattern was investigated; the value of Z obtained for the residual distribution is negative, and no cluster distribution is observed in the residuals, and the residues follow the pattern of random spatial distribution.The effect of each variable on the dependent variable of the migrant was evaluated; the effect of the population density variable in the northern cities of the province was more than in the southern cities of the province. The effect of the origin and destination distance variable is more evident in the south and southeast of the province than in the north. Changes in the effect of the variable number of job applicants at the origin, from northwest to southeast of the province, have an increasing trend. The effect of the urbanization coefficient variable at the origin from north to south of the province shows a decreasing trend. Contrary to the effect of the variable of having urban services at the origin, the trend is increasing from the north of the province to the south and southeast of the province.

In this section, the main research questions were answered. Migrant data has a clustered distribution pattern. The trend of migration of cities to the center of the province during the periods of 90-85 and 90-95, in general, does not show statistically significant changes.Regarding the causes of migration to the center of the province, it can be said that the five variables of population density, a distance of origin and destination, level of urban services, urbanization coefficient at the origin and the number of job applicants at the origin, explain 65% of the dependent variable changes.