Exploring Spatial Patterns of Pan-European Hydrological Signatures and their Links With Catchment Characteristics by Taking Advantage of Large Open Datasets.

Abstract:

The increasing availability of open hydrological and physiographic data over large spatial domains opens the door for a more thorough investigation of dominant flow generating mechanisms across scales using a large number of catchments. This study aims at exploring and understanding the physical controls on spatial patterns of pan-European flow signatures. This understanding will ultimately enhance our ability to predict hydrological variables in ungauged catchments.

In this study, similarities in signatures are explored and compared to similarities in catchment characteristics to distinguish coherence in spatial patterns. In total, some 50 characteristics variables (physical, human alteration and climate) have been computed for more than 1500 stream gauges across Europe. For the same gauges, 15 selected signatures have been calculated for different time-periods (5, 10, 15, 20, 25 and 30 years) of continuous daily flow measurements. Relationships between characteristics and signatures are subsequently explored through correlation analyses to find the best explanatory variables for each signature and to build regression models for predictions in ungauged basins. Significant relationships are observed between some signatures and predictors like land-use area percentages (agriculture, open areas), topography and climatic indices.

Two types of classification (based on catchment characteristics or flow signatures) are applied and the obtained patterns are compared. Regression models are built for each class and compared to the general models built without classification. Attention is drawn to human alteration when looking at outliers or differences between modeled and observed patterns. Finally, the regression models are applied for 35 000 watersheds, mostly ungauged, across Europe (on average 250 km2) to create a map of flow regimes across the European continent. Dominant flow generating processes are analyzed for each class to understand the spatial pattern.