In this post, Associate Editor Nathalie Butt discusses a recent paper ‘Habitat mapping of coastal wetlands using expert knowledge and Earth Observation data’ by Maria Adamo, Cristina Tarantino, Valeria Tomaselli, Guiseppe Veronico, Harini Nagendra and Palma Blonda.

Habitats at risk

What do we usually think of when we hear ‘wetland’ or ‘coastal wetland’? A flat marshy area by the sea somewhere, perhaps with a bird or two flying overhead across a grey sky? Actually, wetlands are much more exciting than that. They can be as productive as coral reefs or rainforests, and support the livelihoods of more than one billion people around the world. However, they are highly threatened, as environmental degradation is more prevalent in these ecosystems than in any other. Coastal wetlands include salt marsh and mangroves which often include rare and threatened plant species, and provide crucial habitat for a large number of birds, insects and amphibians. They occur across all continents, except for Antarctica, and can cover vast tracts of coastline.

In Mediterranean Europe, coastal wetlands include pond systems and marshes, wet meadows of rush, sedge and reed communities, lagoon systems with annual pioneer salt marsh communities, and in drier areas trees such as Aleppo pine, Pinus halepensis. They support many birds, including migratory and over-wintering species, and there is a high level of endemism amongst amphibian and reptile species.

Greater Flamingo Phoenicopterus roseus in a coastal wetland in southern France (Wikimedia commons).

European Union biodiversity legislation

The European Union Habitats Directive was adopted in 1992, and protects 1,000 species and 200 habitats. Because of the rate at which wetland habitats are being lost, their conservation is a key part of the European Union (EU) Biodiversity Strategy: EU Member States are required to protect, maintain, or restore the species, habitats and communities listed in the Strategy. Ecosystem services are also highlighted as important within the strategy, as wetlands globally contribute hugely to ecosystem services (around 40%) including water supply, flood mitigation and nutrient retention. In order to protect and manage these ecosystems effectively, managers and policy makers need accurate information on both the habitats themselves and habitat changes.

Juncus maritimus
Juncus maritimus, a typical Mediterranean saltmarsh species (Wikimedia commons).

Remote sensing data

One way to provide information on where habitats are and how they are changing is through satellite data. Satellite Earth Observation (EO) data, for example, LANDSAT and WorldView-2, are used to map land cover. These remotely sensed data are available at high temporal and spatial resolutions, and are therefore useful for habitat mapping and monitoring. While land cover maps can be automatically generated from these satellite data, and land cover is generally used as a proxy for habitat type, this may be missing some of the picture. Land cover maps produced this way do not account for biodiversity as well as ground-truthed habitat maps do. The problem is that different habitats may correspond to the same land cover class, but the satellite data do not provide sufficient information to allow us to discriminate between them.

Translating land cover maps to habitat maps requires information on specific environmental attributes, which can be lacking, especially in poorly-studied areas. In their recent paper, Adamo et al. demonstrate how, by combining EO data with ‘expert ecological rules’ on coastal wetland habitats, different habitats within the same land cover class can be differentiated, without the need for costly field campaigns to collect environmental attribute data. The authors used two morphologically similar Italian coastal wetland habitats to demonstrate how their method can translate the land cover classification (the same for both habitats) into accurate habitat maps. To do this, they included a range of additional habitat-specific information, such as topological and zonation patterns (spatial data), and plant growth stage and seasonal water regimes (temporal data), to produce accurate habitat maps.

fig 4a
The two habitats outlined in red correspond to the same LC class, but can be discriminated by spatial (topological) relationships.

Over time, changes in the individual habitats can be detected and quantified by comparing frequently produced maps based on these ecological expert rules. The use of this new method will enable more site-specific observation, and will support site managers in long-term habitat mapping. In this way, changes in habitat extent, fragmentation, and other processes, can lead to more effective conservation planning and management, especially where such monitoring is required in order to inform restoration of habitat, such as under the European Habitats Directive.