Temporal land cover analysis for net ecosystem improvement

Although ecological processes at landscape scales can drive ecosystem structure and function at local sites, the monitoring and assessment of restoration projects and programs rarely incorporates changes in landscapes. Riverscapes naturally have a high level of hydrologic connectivity so when floodplain habitats are reconnected to the main stem river to restore function, it is especially important to consider the potential effects of watershed processes on outcomes. This study demonstrates a method for selecting relevant indicators, identifying appropriate spatial scales, and measuring trends to inform assessment of a large-river restoration program. We identified eight major watersheds contributing to river reaches within the 1468 km2 historical floodplain of the lower Columbia River and estuary. We assessed land-cover change at the watershed, reach, and restoration site scales by reclassifying remote-sensing data from the NOAA Coastal Change Analysis Program's land cover/change product into forest, wetland, and urban categories. The watershed analysis showed a 361.6 km2 total and 198.3 km2 net forest loss during the first six years of the Columbia Estuary Ecosystem Restoration Program, 2001–2006. Total urbanization in the watersheds during 1996–2006 was 48.4 km2. Trends in forest gain/loss and urbanization differed between watersheds. Wetland gains and losses were within the margin of error of the satellite imagery analysis. These findings are important to restoration planning and assessment, e.g., floodplain restoration sites in reaches downstream of watersheds with decreasing forest cover may be subject to increased sediment loads and those downstream of urbanization may experience altered hydrological processes caused by increased impervious surface.