Using remote sensing data to assess salmon habitat status in rivers and floodplains of Puget Sound, USA

Abstract:

Natural processes create dynamic habitat features in large rivers and floodplains, and past land uses that restrict fluvial processes have altered habitat conditions in those environments in Puget Sound, USA. As a result, Chinook salmon and steelhead are listed as threatened species under the US Endangered Species Act (ESA). To help restore these salmon populations, restoration actions often focus on removing constraints on natural processes to restore fluvial dynamics and ultimately restore critical salmon habitats on floodplains. An important aspect of this restoration effort is monitoring whether habitat conditions are improving as anticipated, yet there are currently few protocols available for monitoring trends in large river and floodplain habitats. We identified several remote-sensing metrics that are indicators of salmon habitat condition, and developed repeatable protocols for measuring those metrics. We then tested their sensitivity to land use change by comparing habitat conditions among land cover classes (developed, agriculture, forested, and mixed). As expected, metrics of habitat complexity or condition such as side-channel length, node density, wood jam area, or riparian buffer widths were highest in forested sites and lowest in agriculture and urban sites. By contrast, percent disconnected floodplain and percent armored banks were highest in developed sites and lowest in forested sites. Our results indicate that remote sensing metrics are sensitive enough to detect differences in habitat status among land cover classes, and therefore help us understand the impact of various land uses on habitat conditions. However, detecting trends in habitat condition through time may be difficult because magnitudes of change through time are very small.