Sahar Abdollahi; Hashem Dadashpoor
Abstract
Introduction: Land use changes in coastal areas of Gilan Province in recent decades have caused problems such as forest and wetland degradation, soil erosion, biodiversity reduction, and increased environmental pollution. This region is important in terms of its unique features and human use of this ...
Read More
Introduction: Land use changes in coastal areas of Gilan Province in recent decades have caused problems such as forest and wetland degradation, soil erosion, biodiversity reduction, and increased environmental pollution. This region is important in terms of its unique features and human use of this environment for a variety of residential, industrial, and recreational activities. Therefore, it is necessary to be aware of the changes and factors influencing them, and predict the changes process in the future to prevent irreparable damages to the environment. The purpose of this study was to analyze the land use changes in Gilan Province during a 20-year period (1996-2016) and predict changes for the next 30 years based on the integration of the artificial neural network of multilayer perceptron and Markov chain model using the Land-change Modeler. Material and methods: Landsat 5 and 8 (TM) and (OLI-TIRS) satellite images were used for the years 1996, 2006 and 2016. Land cover maps were prepared for five classes in the forest, grass, agriculture, water, and residential resources using the Maximum Likelihood method. Land use changes and then modeling the transmission potential were explored using multilayer perceptron algorithm of artificial neural network using 13 independent variables and obtained 7 sub-models for modeling land use change for 2016 and then using Markov chain method, land use map for the year 2016 was predicted with a coefficient of Kappa 0.98. Finally, the land use pattern of Gilan Province was simulated for 2046. Results and discussion: The results obtained from the analyses of land use changes in the first period (1996-2006) indicated that residential land use with the 7702.72 hectares increased the most among other users. In these changes, agricultural land use had the largest share, where 7663 hectares of this land turned into residential areas. In the second period (2006-2016), residential land use, as in the previous period, with the annual change of 633.7 hectares, had the most significant change in this period. In the whole study period from 1996 to 2016, the residential land reached from 12157.57 hectares in 1996 to 26197.59 hectares in 2016, which agricultural lands had the largest share in the conversion of the built-up areas. Conclusion: The process of land use change suggests that this trend has begun from the past and will continue in the future. So, the results of the detection of changes from the predicting land use for the next 30 years would indicate an increase in residential use and a decrease in the area of agricultural lands, forests, and grasslands. According to these results, timely and accurate evaluation of these changes lead to better decision making and planning.
