Iranian Remote Sensing and GIS Society / Shahid Beheshti University Iranian Journal of Remote Sensing and GIS 2008-5966 16 4 2024 11 21 Land Use Changes Modeling and Predictions Using CA-ANN Hybrid Model in Khamir and Qeshm Mangrove Forests Land Use Changes Modeling and Predictions Using CA-ANN Hybrid Model in Khamir and Qeshm Mangrove Forests 95 114 104488 10.48308/gisj.2024.233850.1189 FA Parvaneh Sobhani Assistant Professor, Department of Environmental Science, Natural Resources Faculty, Lorestan University, Khorramabad, Iran 0000-0001-9878-3768 Afshin Danehkar Professor, Department of Environmental Science, Natural Resources Faculty, University of Tehran, Karaj, Iran 0000-0003-0641-9286 Journal Article 2023 11 20 Introduction: Currently, one of the most important environmental issues in the world is land use change and increasing unsustainability in natural ecosystems and protected areas. Mangrove forests grow in tropical coastal areas and are one of the most vulnerable and endangered ecosystems in the world. Land use changes generally affect the integrity of these natural ecosystems and also lead to habitat conversion and pose a threat to plant and animal life. The direct consequences of these changes and their effects include a decrease in the health and area of mangrove forests, intensification of global warming and climate change, a decrease in coastal water quality, a decrease in biodiversity, fragmentation of coastal habitats, and the destruction of biological resources. Therefore, spatio-temporal monitoring of land use changes and modeling and predicting the trend of these changes can contribute to integrated management and proper planning in mangrove forests. Accordingly, in the present study, the spatial-temporal changes in land use of Khamir and Qeshm mangrove forests during the years 1989-2023 were investigated. In addition, to model and predict the trend of these changes, a combined CA-ANN model was investigated based on the descriptive variables of altitude, slope, population density, distance from settlements, distance from city center, and distance from roads for the year 2060. Materials and Methods: Khamir and Qeshm mangrove forests (Hara Protected Area) with an area of 86,258 ha located in Hormozgan province. In this study, the spatial-temporal changes in land use in this region were investigated using a set of Landsat satellite images (1989-2023) in the Google Earth Engine (GEE) web-based system. In addition, to model and predict these changes, a combined artificial neural network and cellular automata (CA-ANN) model was examined based on descriptive variables of altitude, slope, population density, distance from settlements, distance from city center, and distance from roads, and a map of the possible trend of land use changes for the year 2060 was also prepared. Finally, the ordinary least squares (OLS) regression model was used to analyze the impact of these variables on the trend of land use changes in the region. Results and Discussion: According to the results, the mangrove forests of Khamir and Qeshm show a decreasing trend in 2023 compared to 1989. The results of the prediction of land use changes also showed that tidal zones and bare lands will increase in 2060, while mangrove forests and aquatic areas will decrease. Also, the most significant changes are related to the reduction of mangrove forests (in the northern and southeastern regions of the region) and the increase of other unnatural areas (tidal zones and bare lands in the area of Bandar Khamir in the north and northeast of the region, as well as on the outskirts of the villages of Laft to Goran). Accordingly, considering the values of the probability of transition, mangrove forest covers are susceptible to transformation into other unnatural areas. In addition, the results of the regression model analysis showed that the most important descriptive variables affecting land use changes include distance from settlements and roads due to greater accessibility and the possibility of high development of human activities in these natural habitats. In this regard, the continued increase in land use changes in the Khamir and Qeshm mangrove forests will lead to the destruction and extinction of a large area of these valuable biological reserves in the south of the country. Conclusion: Given that the Khamir and Qeshm mangrove forests are considered areas under environmental protection and management boundaries, the implementation of proposed projects and the construction of any infrastructure and development in this area must be carried out by management plans (zoning) and environmental assessments. On the other hand, land use changes outside the management boundaries of the area must be limited to minimize the reduction in the integrity and fragmentation of the habitat in these natural habitats. Accordingly, reducing the adverse effects of land use changes in this area requires appropriate planning and integrated management in the proper utilization of these natural resources. The findings of this study can also help stakeholders create an opportunity to develop appropriate strategies to protect the mangrove forests of Khamir and Qeshm and restore these habitats. Introduction: Currently, one of the most important environmental issues in the world is land use change and increasing unsustainability in natural ecosystems and protected areas. Mangrove forests grow in tropical coastal areas and are one of the most vulnerable and endangered ecosystems in the world. Land use changes generally affect the integrity of these natural ecosystems and also lead to habitat conversion and pose a threat to plant and animal life. The direct consequences of these changes and their effects include a decrease in the health and area of mangrove forests, intensification of global warming and climate change, a decrease in coastal water quality, a decrease in biodiversity, fragmentation of coastal habitats, and the destruction of biological resources. Therefore, spatio-temporal monitoring of land use changes and modeling and predicting the trend of these changes can contribute to integrated management and proper planning in mangrove forests. Accordingly, in the present study, the spatial-temporal changes in land use of Khamir and Qeshm mangrove forests during the years 1989-2023 were investigated. In addition, to model and predict the trend of these changes, a combined CA-ANN model was investigated based on the descriptive variables of altitude, slope, population density, distance from settlements, distance from city center, and distance from roads for the year 2060. Materials and Methods: Khamir and Qeshm mangrove forests (Hara Protected Area) with an area of 86,258 ha located in Hormozgan province. In this study, the spatial-temporal changes in land use in this region were investigated using a set of Landsat satellite images (1989-2023) in the Google Earth Engine (GEE) web-based system. In addition, to model and predict these changes, a combined artificial neural network and cellular automata (CA-ANN) model was examined based on descriptive variables of altitude, slope, population density, distance from settlements, distance from city center, and distance from roads, and a map of the possible trend of land use changes for the year 2060 was also prepared. Finally, the ordinary least squares (OLS) regression model was used to analyze the impact of these variables on the trend of land use changes in the region. Results and Discussion: According to the results, the mangrove forests of Khamir and Qeshm show a decreasing trend in 2023 compared to 1989. The results of the prediction of land use changes also showed that tidal zones and bare lands will increase in 2060, while mangrove forests and aquatic areas will decrease. Also, the most significant changes are related to the reduction of mangrove forests (in the northern and southeastern regions of the region) and the increase of other unnatural areas (tidal zones and bare lands in the area of Bandar Khamir in the north and northeast of the region, as well as on the outskirts of the villages of Laft to Goran). Accordingly, considering the values of the probability of transition, mangrove forest covers are susceptible to transformation into other unnatural areas. In addition, the results of the regression model analysis showed that the most important descriptive variables affecting land use changes include distance from settlements and roads due to greater accessibility and the possibility of high development of human activities in these natural habitats. In this regard, the continued increase in land use changes in the Khamir and Qeshm mangrove forests will lead to the destruction and extinction of a large area of these valuable biological reserves in the south of the country. Conclusion: Given that the Khamir and Qeshm mangrove forests are considered areas under environmental protection and management boundaries, the implementation of proposed projects and the construction of any infrastructure and development in this area must be carried out by management plans (zoning) and environmental assessments. On the other hand, land use changes outside the management boundaries of the area must be limited to minimize the reduction in the integrity and fragmentation of the habitat in these natural habitats. Accordingly, reducing the adverse effects of land use changes in this area requires appropriate planning and integrated management in the proper utilization of these natural resources. The findings of this study can also help stakeholders create an opportunity to develop appropriate strategies to protect the mangrove forests of Khamir and Qeshm and restore these habitats.

https://gisj.sbu.ac.ir/article_104488_89308c589b23db7c2337db49ef50042c.pdf