Detection of spatio-temporal changes of vegetation in coastal areas subjected to soil erosion issue

Coastal soil erosion can be recognized as the most alarming environmental issue since, causing shoreline retreat, reduces the area available for plant habitat survival, highly influencing their health status, and, consequently, limiting their ability in beach front properties protection. A deep knowledge of vegetation changes is required to identify the proper strategy to be adopted to face soil erosion problems in coastal areas. Therefore, the current paper is aimed to quantitatively examine the spatio-temporal changes suffered by the vegetation in the coastline of Siponto in Apulia Region (Southern Italy) covering a time period of about forty years. LANDSAT images from 1975, 2006, 2011 and 2018 were collected, atmospherically corrected and, finally, processed to generate binary classification maps of vegetation by applying the Composite Vegetation Index, a novel index based on the interpolation of Red, Green and Near-Infrared bands, suitable for catching both cellular and metabolic features of vegetation. Then, the generated binary classification maps were compared using the Vegetation Index Differencing technique, a post-classification change detection technique. The results showed an increase in vegetation extension cover and density overall the entire examined period. That phenomenon appeared more and more prominent between 1975 and 2006, where an increment of vegetated areas extension of about 88% were registered. Combining of the novel vegetation index, developed ad-hoc in the current research, and Vegetation Index Differencing approach shows promising results in vegetation classification and comparison over the time. Indeed, the method allows the fast vegetation extraction, great processing time saver. Nevertheless, spatial resolution of Landsat Images limits the classification of small and low-density vegetated areas. Therefore, future work should plan to test the proposed approach at a more detailed scale.