The executive summary from a report for the Crown Estate detailing the results of a project to examine the use of aerial images to assess the extent of seaweed habitats.
Large brown habitat-forming seaweeds, including fucoids and kelps, are highly productive, essential components of marine ecosystems and create conditions for diverse understory communities of flora and fauna. These seaweed habitats are of great economic value, and it is estimated that kelps alone provide ecosystem services worth billions of pounds. These habitats are increasingly threatened by rising CO2 levels and associated warming. However, a lack of knowledge of the distribution and abundance of these seaweed communities is a major impediment to understanding the impact of such threats and this is compounded by the difficulty of accessing seaweed-dominated areas. In order to overcome the practical difficulty and cost of direct data gathering, there is a need for indirect methods that can be used for assessment, such as remote sensing and distribution modelling. There are a variety of remote sensing options for coastal assessments, ranging from multi-spectral satellite sensors through to high resolution aerial imagery. In order to assess the utility of a variety of satellite and aerial imagery to quantify the extent of seaweed habitats, a pilot project was undertaken along the Thanet coast, North East Kent. This coastline has a detailed history of study, a series of conservation zones and approximately 1 km2 of Fucus habitat which has remained relatively stable over the last decade. Four types of imagery were tested: Channel Coastal Observatory (CCO), Bing maps, RapidEye and Landsat 8). Aerial imagery from the CCO produced the best estimate of habitat extent. These images are the highest resolution available and are coastal-specific surveys taken at low tide. Lower quality images produced smaller habitat estimates, partly due to the images being taken during sub-optimal tidal conditions. It is recognised that there are limitations of what can be achieved with these images. Complete habitat recognition cannot be automated fully, as local knowledge of the types of expected habitats is required to tune any model to region-specific conditions. However, aerial images can be extremely useful for determining distributions of large habitats at local scales.
Yesson C, Ash L & Brodie J (Feb 2015) Using aerial images to quantify the extent of coastal seaweed habitats. A report for the Crown Estate. (PDF)