Environmental properties of coastal vegetated ecosystem locations for machine learning models

This dataset has been created to train and develop a machine learning (ML) model that finds suitable areas to restore and expand coastal vegetated ecosystems (CVE), including seagrass meadows, salt marshes, and mangrove forests. There are two types of datasets. The first type (see Spatial Distribution) of datasets represents the spatial extent of the CVEs. The second type (see Spatial Absence of CVEs) represents areas that are not covered by certain types of CVEs or for which their presence is not known. These datasets cover areas such as: Baltic Sea, North Sea, Mediterranean Sea, Black Sea, Colombia, Australia, Indonesia, Malaysia, East Asia (China, Japan, and South Korea), America (Caribbean, Pacific, Atlantic coast, Gulf coast, and Canada), Africa (Mozambique, Tanzania, Kenya coast, and Gulf of Guinea coast). Data for the model were derived from various databases that provided information on physical, geological, hydrological, and biological aspects of the CVE habitats and coastal areas. With this data, we can use the ML model to analyse which areas without CVEs are potentially suitable for restoration and expansion. The data consists of information such as: habitat, species, location (latitude, longitude), topography (slope, depth), tides, wave energy, chlorophyll-a level, KD490 level, temperature (sea temperature, air temperature, minimum and maximum air temperature), salinity, nutrients (iron, nitrate, and phosphate), tidal range, and tidal stream, eastward, northward velocity, current speed, and sediment type.