Estuarine Landcover Along The Lower Columbia River Estuary Determined From Compact Airborne Spectrographic Imager Casi Imagery

Developing an understanding of the distribution and changes in estuarine and riparian habitats is critical to the management of biological resources in the lower Columbia River. In a recently completed comprehensive ecosystem protection and enhancement plan for the lower Columbia River Estuary (CRE), Jerrick (1999) identified habitat loss and modification as one of the key threats to the integrity of the CRE ecosystem. This management plan called for an inventory of habitats as key first step in the CRE long-term restoration effort. While previous studies have produced useful data sets depicting habitat cover types along portions of the lower CRE (Thomas, 1980; Thomas, 1983; Graves et al., 1995; NOAA, 1997; Allen, 1999), no single study has produced a description of the habitats for the entire CRE. Moreover, the previous studies differed in data sources and methodologies making it difficult to merge data or to make temporal comparisons. Therefore, the Lower Columbia River Estuary Partnership (Estuary Partnership) initiated a habitat cover mapping project in 2000. The goal of this project was to produce a data set depicting the current habitat cover types along the lower Columbia River, from its mouth to the Bonneville Dam, a distance of (almost equal to)230-km (Fig. 1) using both established and emerging remote sensing techniques. For this project, we acquired two types of imagery, Landsat 7 ETM+ and Compact Airborne Spectrographic Imager (CASI). Landsat and CASI imagery differ in spatial and spectral resolution: the Landsat 7 ETM+ sensor collects reflectance data in seven spectral bands with a spatial resolution of 30-m and the CASI sensor collects reflectance data in 19 bands (in our study) with a spatial resolution of 1.5-m. We classified both sets of imagery and produced a spatially linked, hierarchical habitat data set for the entire CRE and its floodplain. Landsat 7 ETM+ classification results are presented in a separate report (Garono et al., 2003). This report presents classification results from analysis of the CASI imagery. Data sets produced for this project from both types of imagery fill a critical information gap by creating a current description of the condition and extent of estuarine habitat cover types along the lower Columbia River. Results from this study will be used by the Estuary Partnership and its cooperators to: (1) develop indicators of 'habitat health' and biological integrity; (2) develop definitions of 'critical salmonid habitat'; (3) identify and evaluate potential wetland conservation and restoration sites; (4) track exotic and invasive species; and (5) develop an understanding of how estuarine and riverine habitats have changed over the past 200 years. This study focuses on estuarine and riparian habitat cover types important to native species, particularly juvenile salmonids. This study is meant to provide support to the multiple efforts currently underway to recover 12 species of Columbia River salmonids identified as endangered or threatened under the Endangered Species Act.

Developing an understanding of the distribution and changes in estuarine and tidal floodplain ecosystems is critical to the management of biological resources in the lower Columbia River. Columbia River plants, fish, and wildlife require specific physicochemical and ecological conditions to sustain their populations. As habitats are degraded or lost, this capability is altered, often irretrievably; those species that cannot adapt are lost from the ecosystem. The Lower Columbia River Estuary Partnership (Estuary Partnership) completed a comprehensive ecosystem protection and enhancement plan for the lower Columbia River and estuary in 1999 (Jerrick, 1999). The plan identified habitat loss and modification as a critical threat to the integrity of the lower Columbia River ecosystem and called for a habitat inventory as a key first step in its long term restoration efforts. In 2000, the Estuary Partnership initiated a multiphase project to produce a spatial data set describing the current location and distribution of estuarine and tidal freshwater habitat cover types along the lower Columbia River from the river mouth to the Bonneville Dam using a consistent methodology and data sources (Fig. 1). The first phase of the project was the development of a broadbrush description of the estuarine and tidal freshwater habitat cover classes for the entire study area ((almost equal to)146 river miles) using Landsat 7 ETM+ satellite imagery. Phase II of the project entailed analysis of the classified satellite imagery from Phase I. Analysis of change in landcover and a summary of the spatial relationships between cover types are part of Phase II. Phase III of the project included the classification of the high resolution hyperspectral imagery collected in 2000 and 2001 for key focal areas within the larger study area. Finally, Phase IV consists of this final report that presents results from refining the Landsat ETM+ classification and provides recommendations for future actions. Previous studies (Thomas, 1980; Thomas, 1983; Graves et al., 1995; NOAA, 1997; Allen, 1999) produced similar landcover data sets; however, most of these studies used multiple and varied data sources and differed from one another in methodologies. Currently, no single data set has been produced using a consistent methodology and uniform scale data, which describes current estuarine and tidal freshwater floodplain cover types from the Columbia's mouth to the Bonneville Dam (Fig. 1). Results from this study will be used by the Estuary Partnership and its cooperators to: (1) develop indicators of 'habitat health' for target species and populations, and biological integrity at the community and ecosystem scales; (2) develop definitions of 'important salmonid habitat'; (3) identify and evaluate potential wetland conservation and restoration sites; (4) track non-indigenous and invasive species; and (5) develop an understanding of how estuarine and floodplain habitats have changed over the past 200 years. This study focused on estuarine and tidal freshwater floodplain habitat cover types, which are important to native species, particularly juvenile salmonids. Results from this study are meant to provide support for the multiple efforts currently underway to recover 12 species of Columbia River salmonids identified as endangered or threatened under the Endangered Species Act. Spatial scale was an important consideration in this study. Our goal was to create a geographic information system (GIS) coverage depicting habitat cover types for the entire 146 river miles of the study area and the associated floodplain, at a spatial resolution sufficient to resolve important estuarine and floodplain features, wherever possible. Thus, in addition to the small scale (30 m pixel size) satellite imagery covering the study area described in this report, we also acquired high spatial resolution imagery ((almost equal to)1.5 m pixel size) for key portions of the study area using a Compact Airborne Spectrographic Imager (CASI). Compared to the rather coarse, low spectral resolution of the satellite imagery, the CASI imagery provides the spectral resolution (19 CASI bands vs. 7 ETM+ Bands) necessary to discriminate between spectrally similar vegetation types and thus provided a greater ability to resolve habitat features important to species such as salmonids. Unfortunately, image acquisition costs and logistics (time and poor weather) prevented complete coverage of the study area at a 1.5 m spatial scale. However, in 2000/01 CASI imagery was collected for several key focal areas including the Chinook watershed, the area around the Astoria Airport mitigation site, Russian Island, Tenasillahe Island, Sauvie Island, Scappose Bay lowlands, Lord and Walker Islands and much of the shoreline. This report presents the classification methodology and initial results from the satellite image classification.

The Pacific Northwest is a global ecological "hotspot" because of its relatively healthy native ecosystems, a high degree of biodiversity, and the number and scope of restoration initiatives that have been undertaken there. Restoring the Pacific Northwest gathers and presents the best examples of state-of-the-art restoration techniques and projects. It is an encyclopedic overview that will be an invaluable reference not just for restorationists and students working in the Pacific Northwest, but for practitioners across North America and around the world.

The Lower Columbia River Estuary Management Plan (Jerrick, 1991) recognizes the positive relationship between the conservation of fish and wildlife habitat, and sustaining their populations. An important component of fish and wildlife conservation and management is the identification of habitats, trends in habitat change, and delineation of habitat for preservation, restoration or enhancement. Alterations to the environment, such as hydropower generation, dredging, forestry, agriculture, channel alteration, diking, bank stabilization and floodplain development, have dramatically altered both the type and distribution of habitats along the Columbia River Estuary (CRE) and its floodplain. Along the Columbia River, tidally influenced habitats occur from the river mouth to the Bonneville Dam, a distance of 230 km. If we are to effectively manage the natural resources of the Columbia River ecosystem, there is a need to understand how habitats have changed because fish and wildlife populations are known to respond to changes in habitat quality and distribution. The goal of this study was to measure the amount and type of change of CRE land cover from 1992 to 2000. We performed a change analysis on two spatial data sets describing land cover along the lower portion of the estuary (Fig. 1). The 1992 data set was created by the NOAA Coastal Remote Sensing, Coastal Change Analysis Program (C-CAP) in cooperation with Columbia River Estuary Study Task Force (CREST), the National Marine Fisheries Service (NMFS) Point Adams Field Station, and State of Washington Department of Natural Resources (DNR). The 2000 data set was produced by Earth Design Consultants, Inc. (EDC) and the Wetland Ecosystem Team (WET: University of Washington) as part of a larger Lower Columbia River Estuary Partnership (Estuary Partnership) habitat mapping study. Although the image classification methodologies used to create the data sets differed, both data sets were produced by classifying Landsat Thematic Mapper (TM) satellite imagery, making it feasible to assess land cover changes between 1992 and 2000.