Remote sensing of fluvial environments: Riverscape characterization of in-stream hydraulic habitat heterogeneity

Anthropogenic or natural disturbances to river ecosystems (driven by river regulation or land use and climate changes) often lead to long term changes in morpho-sedimentological dynamics and, in turn, to gradual modifications to riverine habitats over extended river lengths. Moreover, since many high value river organisms such as 'potamodromous' fish species migrate across complementary habitats distributed along stream networks, it is important for river ecosystem conservation to characterize efficiently variations in habitat types over extensive river segments. To this end, the 'riverscape' approach (Fausch et al., 2002) provides a framework which considers degrees of connectivity across various types of river habitats, as well as degrees of habitat heterogeneity (HH) over reaches of various lengths. However, such riverscape characterization requires high resolution (< 5 m), continuous habitat data collected over long river segments, which is difficult and expensive to acquire, especially through field data collection. Bridging this data gap is the objective of this thesis. The research presented here provides low-cost methods, based on satellite and airborne imagery, capable of extracting various metrics of in-stream hydraulic habitat, calculated over scale flexible moving windows, for long river segments (10 – 100 km). The approach is based on combining remote sensing with simplified, "pseudo-2D" hydraulic modeling for the flow conditions at the time of image capture. It can generate 1 m resolution depth and velocity maps from which hydraulic habitat variables can be quantified over various selectable windows (such as pool depth and area statistics, riffle lengths, mean velocity and Froude number profiles, as well as various indices of reach scale hydraulic habitat heterogeneity HH, etc.). We also investigate the use of airborne hyperspectral images to classify and to quantify in-stream habitat features such as the bed substrate composition and submerged aquatic vegetation density, based on the analysis of hyper-spectral signatures.We also apply these methods to extract hydraulic habitat data over ten (10) rivers across five (5) different physiographic regions of Canada (for a total of 163 km of river length) and demonstrate that reach scale hydraulic HH was correlated with local valley scale features (such as tributary junctions and lateral channel constraints) reflecting the regional physiographic context. Further investigation on the HH distributions along the ten (10) studied Canadian river segments revealed three (3) major scales of along river HH variability, governed by dominant geomorphic processes specific to each scale.The methods and concepts presented in this thesis contribute to advancing the field of riverscape science by providing analytical tools available to river scientists and managers, given the decreasing costs and increasing resolution of satellite imagery. The applications of this framework present opportunities to bridge the gap between river ecologists and geomorphologists, and to remedy the lack of uniform methods for characterizing and analyzing lotic ecosystems at the riverscape extent, using scale-flexible metrics.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.