Investigating environmental factors for locating mangrove ex-situ conservation zones using GIS spatial techniques and the logistic regression algorithm in mangrove forests in Iran

Abstract

Along the coastal strip of tropical and subtropical regions, mangrove forests extend as one of the most productive coastal ecosystems. Iran is considered one of the most western coastal habitats of these communities in Asia. Unfortunately, in recent years, despite their economic and ecological importance, these forests have been threatened mainly due to human activities, which necessitate the need for protecting and developing these ecosystems. Ex-situ conservation, which aims to transfer a species to right place for preservation and development, is one of the most effective ways of protecting these forests. It is obvious that the successful transfer of a species in order to protect it as ex-situ conservation requires a comprehensive environmental assessment. The present paper employs a spatio-statistical methodology to delineate areas enriched with environmental parameters and weather factors necessary to support a mangrove ex-situ conservation plan in Hormozgan Province, Iran. With extensive effort in reviewing the related studies in the literature, we elicted a collection of 14 environmental and weather indicators critical to maintaining mangrove stands in a sustainable ecological environment. For screening the number of indicators to environmental variables with highest importance in our research, we used the Delphi method to quantify expert attitudes in order to develop a suitability index collection. Accordingly, an integrated application of GIS spatial analysis and a logistic regression algorithm were used to build an explicit spatial predictive model that evaluates the targeted area in terms of identified factors, thereby quantifying the potential of suitable zones for locating mangrove ex-situ conservation sites. Predictive performance of the developed model was evaluated based on Kappa index and omission error rate. According to our results, land proneness to a mangrove plantation has a high positive correlation with environmental attributes such as temperature fluctuation, land form, and mean tidal level, and has a negative correlation with slope and wave height. Our results should be considered advantageous for decision makers and planners in the field of mangrove ecosystems.