Julien Godfroy

THESIS

Coupling LiDAR and hyperspectral data to characterize fluvial corridors.

This PhD. aims at assessing the ability of emerging fluvial remote sensing techniques (such as terrestrial and topo-bathymetric LiDAR, hyperspectral imaging, and thermal infrared imaging) to characterize and monitor fluvial corridors. The study site is the lower basin of the Ain River which is affected by a range of issues related to channel incision, and is the location of a gravel augmentation project. First, we use hyperspectral imaging to predict channel bathymetry for a reference discharge and depths up to 2.5 meters along a river reach of 20 km. This allows us to confirm the ability of hyperspectral optical models to be extrapolated to a long river reach, and to retrieve with a single campaign bathymetric data for multiple discharge conditions. We study the errors occurring along the 20 km reach and then we use the same methodology to identify morphological changes occurring between 2015 and 2022 and to therefore evaluate its applicability for change monitoring. Second, we combine forestry field surveys with LiDAR and hyperspectral data in order to characterize the riparian forest. We demonstrate the ability of such datasets to describe morphological changes along an age gradient and to predict the degree of hydrological connection of the riparian forest. Changes in the physical site conditions of the riparian forest due to channel incision leads to dryer forest conditions, and suggests the presence of potential water stress at the plot level. Third, we explore this water stress hypothesis by using thermal infrared imaging and eco-physiological field measurements. While mounting thermal infrared sensors on an airborne vector can enable water stress characterization, targeting a specific acquisition window or acquiring multiple campaigns is required to detect stress conditions due to their variability during summer. This work then allows us to make recommendations for both the management and restoration of fluvial systems and the use of remote-sensing data to characterize and monitor such environments.

Ain River, Hyperspectral imaging, LiDAR, Bathymetry, Riparian forest, Water stress, Channel incision, Fluvial remote sensing, Thermal infrared imaging

Hervé Piégay, EVS, ENS de Lyon, CNRS

ED 483 Social sciences

UMR 5600 EVS

6 april 2023

French

• Frédéric Liébault
• Jérôme Lejot
• Candide Lissak
• Hervé Piégay
• Christophe Delacourt