Quantitative morphodynamic typology of rivers: a methodological study based on the French Upper Rhine basin

Abstract

This study sought to establish a quantitative functional hydrogeomorphological typology of river reference types in the French Upper Rhine basin that would meet the requirements of the European Water Framework Directive, defined in terms of hydromorphological quality and prospects for river restoration. Four ecoregions (i.e. hydrogeomorphic units) were delineated by expert opinion and validated by independent variables (i.e. 1·5-year peak discharge, comparison of 10-year daily flow and low flow, valley bottom morphology and specific stream power). A data set of 31 quantitative and qualitative variables for 187 field sites was established and analysed. Agglomerative hierarchical clustering (AHC) of the quantitative variables and principal component analysis (PCA) provided hierarchy and grouping of variables. Multiple correspondence analysis (MCA) confirmed these results, discriminating sites into seven groups, but did not lead to a functional typology due to important overlapping between groups and variability within groups. A definitive quantitative typology was obtained through AHC and discriminant analysis with cross-validation computed separately in each ecoregion. Results for two ecoregions, 'marly calcareous loess-covered hills' and 'crystalline Vosges mountains', are given as examples. Finally, the classification of the sites is extended to the whole river network by river sectorization, each homogeneous river section being classified to one category on the basis of the preceding classification of sites. Methodological implications regarding river classification are also given.WFD only suggests possible criterion for river classification (altitude, geology, energy, valley and channel morphology, sediment transport etc.). Each country has to elaborate its own operational method.A large array of methods for hydrogeomorphological typologies can be found in the literature. This range of approaches is related to various objectives, which are often linked to specific management problems, various scales and different geographical settings (Kondolf et al., 2003). Many authors have tried to highlight the principles and results of these numerous approaches through several bibliographic syntheses (Hawkes

Ecoregions and stream morphology in eastern Oklahoma

“…For example, a region with similar lithology, climate, geomorphology, and land use history would support similar channel characteristics (Kondolf et al, 2003). This hierarchical approach is widely accepted as a measure to assess the factors that influence stream morphology and the ecological characterization of rivers (Vannote et al, 1980;Lotspeich, 1980;Brussock et al, 1985;Biggs et al, 1990;Montgomery, 1999;Brierley and Fryirs, 2000;Thompson et al, 2004;Schmitt et al, 2007). A landscape that has been regionalized by both physical and biological components (i.e., ecoregions) provides a broad-scale hierarchy by which investigations into fluvial studies begin.…”

Comparison of statistical algorithms for detecting homogeneous river reaches along a longitudinal continuum

“…As shown by previous authors (Mosley, 1987;Schmitt et al, 2007), 325 geomorphic typology is based on a set of variables to try to summarise the 326 complexity, often based on exploratory multivariate statistics. Gurnell (1997) 327 studied the hydraulic thresholds along a fluvial continuum using a discrim- introduces these patterns instead of alternative statistical assumptions.…”

Spatial disaggregation and aggregation procedures for characterizing fluvial features at the network-scale: Application to the Rhône basin (France)

“…Numerous classification schemes have been established in fluvial geomorphology (Kellerhals et al, 1976;Montgomery and Buffington, 1997;Brierley and Fryirs, 2000;Schmitt et al, 2007) and ecology (Naiman et al, 1992;Beechie et al, 2006;Chessman et al, 2006;Schmitt et al, in press) as a means of reducing the network-scale diversity and complexity of fluvial features into a series of discrete and labeled groups (Kondolf et al, 2003). Studies have recognized that no single universal classification scheme exists despite a large consensus on the factors determining forms and processes in rivers for a range of spatial and temporal scales (Kellerhals et al, 1976;Frissell et al, 1986).…”