Comparative Evaluation of Hyperspectral Imaging and Bathymetric Lidar for Measuring Channel Morphology Across a Range of River Environments
Abstract
Reliable topographic information is critical to many applications in the riverine sciences. Quantifying morphologic change, modeling flow and sediment transport, and assessing aquatic habitat all require accurate, spatially distributed measurements of bed elevation. Remote sensing has emerged as a powerful tool for acquiring such data, but the capabilities and limitations associated with various remote sensing techniques must be evaluated systematically. In this study, we assessed the potential of hyperspectral imaging and bathymetric LiDAR for measuring channel morphology across a range of conditions in two distinct field sites: the clear-flowing Snake River in Grand Teton National Park and the confluence of the Blue and Colorado Rivers in north-central Colorado, USA. Field measurements of water column optical properties highlighted differences among these streams, including the highly turbid Muddy Creek also entering the Colorado, and enabled theoretical calculations of bathymetric precision (smallest detectable change in depth) and dynamic range (maximum detectable depth). Hyperspectral imaging can yield more precise depth estimates in shallow, clear water but bathymetric LiDAR could provide more consistent performance across a broader range of depths. Spectrally-based depth retrieval was highly accurate on the Snake River but less reliable in the more complex confluence setting. Stratification of the Blue/Colorado site into clear and turbid subsets did not improve depth retrieval performance. To obtain bed elevations, image-derived depth estimates were subtracted from water surface elevations derived from near-infrared LiDAR acquired at the same time as the hyperspectral images. For the water-penetrating green LiDAR, bed elevations were inferred from laser waveforms. On the Snake River, hyperspectral imaging resulted in smaller mean and root mean square errors than bathymetric LiDAR, but at the Blue/Colorado site the optical approach was subject to a shallow bias not evident in the LiDAR. Our results can help prospective users to select appropriate instrumentation for their study objectives and site characteristics.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2014
- Bibcode:
- 2014AGUFM.H52C..01L
- Keywords:
-
- 0496 Water quality;
- 1821 Floods;
- 1855 Remote sensing;
- 1856 River channels