Investigation of accumulation zones in bays along the northern Maltese coastline

Abstract

One of the most important, rapidly expanding, and deteriorating global environmental issues is marine litter. While many studies have been conducted on the accumulation and dispersion mechanisms of surface marine litter, studies which assess the accumulation zones of benthic marine litter are still lacking, especially in the Maltese islands. This study attempts to investigate if a relationship between bathymetry and benthic litter accumulations exists. Google Earth Engine was used to obtain the pseudo Satellite Derived Bathymetries (pSDB) for 10 scenes using a Blue:Green band ratio and a Blue:Red band ratio. Each scene was then modelled using two modelling techniques: Linear Regression model and Random Forest model and the results were then compared with in-situ LIDAR data to identify which combination of month, band ratio and modelling technique achieved the lowest Mean Average Error (MAE). The scene from July 2020 using the Blue:Green band ratio with the Random Forest modelling technique proved to be the most accurate and was utilised as a basis to interpolate SDB at each bay. Litter data from a previous study was then projected over the SDB interpolations, and the sample points of the litter data were compared to the bathymetry of the bay. Scatter plots comparing litter count with depth of the studied bays revealed that shallower depths had little to no litter counts, but once the 2-metre depth is exceeded, litter count substantially increases. Several visual instances also showed similar occurrences, with small amounts of litter in shallow depth and larger amounts of litter in deeper depths. However, when a Spearman’s Rank correlation was applied to the two variables, no statistical relationship could be identified.