Document Type : Research Article
Authors
1
Faculty of Geography ; University of Tehran; Tehran; Iran
2
Professor, Department of Natural Geography, Geomorphology, Faculty of Geography, University of Tehran
3
Associate Professor, Department of Natural Geography, Geomorphology, Faculty of Geography, University of Tehran
Abstract
Urban environmental pollution in coastal areas significantly impacts water and soil quality, biodiversity, public health, and tourism activities. This study aims to identify and map pollution risk zones along the coasts of Anzali, Rezvanshahr, and Astara using a hybrid approach combining the Analytical Hierarchy Process (AHP) and the Random Forest (RF) machine learning algorithm. Fourteen critical criteria—including urban wastewater, population, human activity, coastal degradation, natural attractions, and tourism infrastructure—were selected from an initial set of 24 indicators through comparative analysis. Using expert judgment, the relative weights of each factor were calculated in Expert Choice software, and GIS-based spatial layers were generated in ArcGIS Pro. The pollution risk map was created by integrating the weighted layers. The RF model was trained using 70% of the high-risk zones as training data and 30% for testing. Model validation using ROC–AUC analysis showed excellent accuracy for the RF model (AUC = 0.97). The results revealed that Anzali County had the highest proportion of very high-risk zones, while Rezvanshahr and Astara also exhibited significant high-risk areas. The findings highlight the combined impact of human and natural drivers on coastal pollution and offer valuable insights for environmental management, spatial planning, and policy interventions to mitigate future risks.
Extended Abstract
Introduction
Coastal areas require special attention due to the vulnerability of ecosystems and the dependence of human activities on them. Today, beaches have become one of the most popular destinations for tourists, and the presence of various coastal features creates diverse potentials for attracting visitors. Tourism, as a rapidly growing phenomenon, has become one of the largest industries in the world. The development of tourism and related recreational activities has a significant impact on the development patterns of regions and helps provide direct and indirect economic, social, cultural, and environmental benefits to host communities. However, at the same time, this development raises concerns and threats regarding environmental degradation and the destruction of the natural, historical, and cultural heritage of local residents. In fact, mass tourism does not equate to cleanliness or environmental sustainability. In this regard, some of the most significant impacts of tourism-related activities on the physical coastal environment include soil erosion, landslides, water quality degradation, shoreline alteration, and litter pollution. Litter pollution has become an undeniable threat to the sustainability of coastal ecosystems, and the threat posed by non-biodegradable plastic waste to coastal environments is increasingly evident. Pollution in the Caspian Sea poses a serious threat to aquatic life and human health. Unfortunately, coastal zones and estuaries are highly exposed to pollution, and this process endangers the survival of marine species. Urban environmental pollution in coastal areas can lead to a decline in water and soil quality, a reduction in biodiversity, health problems, and negative impacts on tourism and recreational activities.
Material and Methods
The main objective of this study is to identify and zone the coastal areas of the three counties of Anzali, Rezvanshahr, and Astara in terms of urban environmental pollution risk using the Analytic Hierarchy Process (AHP) and the Random Forest (RF) machine learning algorithm. In this regard, effective criteria were first identified. From an initial set of 24 criteria, 14 were selected: natural attractiveness, coastal degradation, river, road, human activity, urban drainage, population, urban waste, protection and management, facilities, security, natural landscapes, hotels, and the influence of tourism on population. Information layers for these were prepared in the ArcGIS Pro environment. Using the AHP method, the relative impact of each criterion was determined, and 70% of the high-risk areas were used as training data and 30% as test data for training the Random Forest algorithm. The predictive accuracy of the model results was evaluated using the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC).
Results and Discussion
The results of validating the zoning maps using test data showed that the accuracy of the analytic hierarchy process and random forest models was 0.812 and 0.97, respectively. These results show that the random forest model has high accuracy in zoning coastal pollution. The results also show that the security criterion is the most significant in training the machine learning algorithm.
The results of coastal zone pollution zoning in the analytic hierarchy process and random forest model showed that in Anzali County, 0.82% and 10.10% were in the very low-risk class, 0.55% and 16.55% in the low-risk class, 0.6% and 0.36% in the medium-risk class, 73.89% and 27.35% in the high-risk class, and 20.6% and 24.9% in the very high-risk class. In Rezvanshahr County, 0.57% and 12.55% were in the very low-risk class, 7.09% and 28.71% in the low-risk class, 21.27% and 22.68% in the medium-risk class, 66.64% and 98% in the high-risk class, and 0.45% and 0% in the very high-risk class. In Astara County, 0.02% and 26.73% were in the very low-risk class, 6.30% and 23.21% in the low-risk class, 23.87% and 31.20% in the medium-risk class, and 6.44% and 16.51% in the high-risk class.
Conclusion
Pollution from industrial and domestic wastewater entering the coastal waters is one of the major challenges on the northern coasts of Iran. In Bandar Anzali, this problem has been exacerbated due to the lack of adequate wastewater management infrastructure and poor enforcement of environmental laws. The general conclusion for the three cities of Astara, Bandar Anzali, and Rezvanshahr shows that all three regions face specific environmental pollution challenges caused by a combination of natural and human factors. Managers and planners can effectively reduce the risks of pollution in coastal areas by using the results of this research. Furthermore, these results will help facilitate the development and implementation of practical solutions.
Keywords
Main Subjects
)
Send comment about this article