Integrating catchment land cover data to remotely assess freshwater quality: a step forward in heterogeneity analysis of river networks

Abstract

Attempts to obtain information from geospatial data in freshwater ecology is a highly challenging task requiring the development of new concepts and adequate tools. Conventionally, river networks are represented as collections of vectors, but they can also be thought of as a succession of raster cells corresponding to the digital elevation model of the landscape they traverse. Based on the principle that each cell in the river raster collects environmental influences from its upstream drainage basin, we defined a remote measure of the potential of pollution named RWQ (Remote Water Quality). We used the CORINE Land Cover categories found in the catchment area of each cell in the river network grouped by ecological relevance and weighted by their respective areas in the catchment. To refine the index to account for the proximity of potential pollution sources, we tested successive buffers of 1 km up to the full catchment of each investigated point, concluding that the RWQ calculated for the full catchment is the most suitable index. For implementation, we developed RIVERenhancer, a free Python-based ArcGIS tool making possible the enhancement of raster river networks with data extracted from various files. The reliability of RWQ was tested with the aid of in situ measurements of chemical and biological water quality obtained from several sources in Danube basin (Romania and Hungary). The strong correlation with field data shows that this index can be considered a surrogate to depict the quality of freshwater habitats and to analyse network heterogeneity. The strength of this concept comes from taking advantage of the dendritic nature of river networks, opening new directions of operations for large scale approaches concerning important issues in global ecology, biogeography and conservation.