This dataset is an output of the ÔÇ£Mediterranean Sensitive HabitatsÔÇØ project (MEDISEH). It shows under a raster form modelled spatial distributions of Posidonia oceanica across the Mediterranean Sea. Posidonia oceanica is endemic to the Mediterranean Sea, where it is the dominant seagrass, covering about 50,000 km2 of coastal to offshore sandy and rocky areas down to depths of about 45 m. P. oceanica is a protected species according to EU legislation (Habitat directive), the Bern and Barcelona Conventions and several national legislations. While its distribution is well documented along the European shores of the western Mediterranean Sea, limited information is available about the southern shore and the eastern Mediterranean Sea. In order to bridge this information gap, one of the goals of Task 1.3 of the MEDISEH project was to model P. oceanica distribution across the whole Mediterranean basin. A Random Forest (i.e. a Machine Learning technique) was ÔÇ£trainedÔÇØ on data from regions where information was available and then used to predict the probability of occurrence of P. oceanica where needed. The raster has a spatial resolution of 0.004166 decimal degrees, and the values are in the [0,1] interval (occurrence probabilities).

This dataset is an output of the ÔÇ£Mediterranean Sensitive HabitatsÔÇØ project (MEDISEH). It shows under a raster form modelled spatial distributions of Posidonia oceanica across the Mediterranean Sea. Posidonia oceanica is endemic to the Mediterranean Sea, where it is the dominant seagrass, covering about 50,000 km2 of coastal to offshore sandy and rocky areas down to depths of about 45 m. P. oceanica is a protected species according to EU legislation (Habitat directive), the Bern and Barcelona Conventions and several national legislations. While its distribution is well documented along the European shores of the western Mediterranean Sea, limited information is available about the southern shore and the eastern Mediterranean Sea. In order to bridge this information gap, one of the goals of Task 1.3 of the MEDISEH project was to model P. oceanica distribution across the whole Mediterranean basin. A Random Forest (i.e. a Machine Learning technique) was ÔÇ£trainedÔÇØ on data from regions where information was available and then used to predict the probability of occurrence of P. oceanica where needed. The raster has a spatial resolution of 0.004166 decimal degrees, and the values are in the [0,1] interval (occurrence probabilities).