Intertidal ground-truth surveys
The intertidal ground-truth survey has two aims, to validate and ground-truth the remote sensing survey.
The validation is required to check both horizontal and vertical accuracy of the Digital Terrain Model (DTM) using dGPS RTK (Real Time Kinematic) ground surveys of reference sites and conspicuous objects selected from (aerial) orthophotographs. This can be done at any state of the tide as the reference sites are usually located in the upper reaches of the tidal zone or on dry land. The procedure is explained in the Case Study Validating the digital terrain model.
An initial interpretation of the remotely sensed data (satellite imagery, aerial photographs, LiDAR etc) is made using unsupervised classification or manual contouring to identify different facies and their borders, which are represented on a draft physical map. Borders may be distinct, or gradual transitions from one facies type to another.
The ground-truth survey is then planned to target different facies and borders, using a series of transects and point sample locations. The majority of surveying will use direct observation, noting the nature of the substrates and the communities they support. Quantitative sampling should be undertaken at a selection of sites representing distinct facies types to provide detailed information on the community composition and nature of the sediments (through granulometric analyses).
Intertidal ground-truth sampling sites and transects overlain on an orthophotograph at a study site in le Croisic, Brittany.
If automated interpretation and classification techniques are to be used to analyse the remotely-sensed data, then care must be taken to adequately sample locations that will be used as ‘training zones’ in supervised classification. Sufficient samples and observations should be collected to allow the resulting data to be divided into two parts, one data set being used to make the interpretation (i.e. to define ‘signatures’ to be used in the supervised classification) and the other to subsequently validate that interpretation.
Unlike sub-tidal ground truth surveys, the intertidal work can be far more flexible and ‘adaptive’, Because the ground is easily seen, decisions can be made during the survey to adjust the sampling plan to make sure the sampling sites are located in the most representative areas. Ad-hoc sampling is also possible, for example running a transect across a gradual change in sediment type that is recognised ‘in the field’ but was no obvious in the remotely sensed data.
A likely optimal sequence for intertidal ground truth surveys is summarised in the flow diagram, and further details are provided in the Case Study: Optimising intertidal zone field work.
The latter example goes on to explain that, the knowledge gained from the physical ground-truth sampling is used to make a more refined interpretation of the remotely sensed data to produce a morpho-sediment map. If the remotely sensed data includes digital imagery, automated or semi-automated classification techniques can be used. If only photographs are available, the interpretation relies on informed expert judgement. Knowledge gained from the biological ground truthing is used to further refine the map, identifying and delineating bio-physical habitats according to the classification system being used (e.g. EUNIS). The final map is then validated using the validation data set.
In shallow water areas, the ground truthing strategy is similar, with the exception that sampling and observations are more difficult to perform. It can frequently be the case that other remote sensing techniques, such as shallow water sidescan or multibeam sonar, can be run to confirm the presence or lower depth range of kelp, seaweed and seagrass and to give cross referenced information on the nature if the seabed. Divers can be used for further observation and grabs can collect sediment samples.

All material variously copyrighted by MESH project partners 2004-2010

| List Access Keys | Mapping European Seabed Habitats (MESH) - Home | Overview | News | Work areas in the MESH Project | Partners | Contact Us | Search | Site Map | Partner Extranet |