Classifying Wetland and Coastal Environments

Correlations between wetland habitat, coastal environment, lidar-derived elevation, and EM-derived conductivity suggest that lidar and EM data could be used to improve the accuracy of coastal habitat classification and partly automate the process. One approach would be to combine photographic, elevation, and conductivity data in a common spatial environment, using elevation and conductivity as a supplement to aid classification of ambiguous habitat signatures on aerial photographs.

A more quantitative approach would be to establish statistical elevation and conductivity characteristics for all possible habitat and coastal environment types, then use measured elevations and conductivities to classify locations according to proximity of each measurement to average elevation and conductivity for each habitat or environment. Because the statistical characteristics (average, range, and standard deviation) could be calculated for each unit, probabilities of accurate classification could be assigned for each point. Because elevations and conductivities are quite distinct between upland and fresh habitats and estuarine and marine habitats, probabilities of misclassification at this level would be low. Probabilities of misclassification among habitats with more elevation and conductivity overlap, such as between some estuarine and marine units and saline environments, would be higher.