Mapping macrophytic vegetation in shallow lakes using the Compact Airborne Spectrographic Imager (CASI)

Abstract

1. The ecological status of shallow lakes is highly dependent on the abundance and composition of macrophytes. However, large-scale surveys are often confined to a small number of water bodies and undertaken only infrequently owing to logistical and financial constraints. 2. Data acquired by the Compact Airborne Spectrographic Imager-2 (CASI-2) was used to map the distribution of macrophytes in the Upper Thurne region of the Norfolk Broads, UK. Three different approaches to image classification were evaluated: (i) Euclidean minimum distance, (ii) Gaussian maximum likelihood, and (iii) support vector machines. 3. The results show macrophyte growth-habits (i.e. submerged, floating-leaved, partially-emergent, emergent) and submerged species could be mapped with a maximum overall classification accuracy of 78% and 87%, respectively. The Gaussian maximum likelihood algorithm and support vector machine returned the highest classification accuracies in each instance. 4. This study suggests that remote sensing is a potentially powerful tool for large-scale assessment of the cover and distribution of aquatic vegetation in clear water shallow lakes, particularly with respect to upscaling field survey data to a functionally relevant form, and supporting site-condition monitoring under the European Union Habitats (92/43/EEC) and Water Framework (2000/60/EC) directives.