Classifying riparian forest cover in the Cedar River Watershed using remote sensing techniques of assessment to identify areas suitable for restoration
Lauren Mollot
Ph.D., Forest Resources
The Cedar River Watershed, located approximately 30 miles east of Seattle, is the primary source of clean drinking water for the 1.3 million Seattle residents. Owned and managed by the City of Seattle, the Watershed occupies over 90,000 acres of streams, wetlands, and forests.
Historically, the watershed was managed for timber and water resources. However, recently the watershed has been taken out of timber production and dedicated as an ecological reserve. In order to promote connectivity of late success ional habitat, the management objectives aim to restore late-successional and old growth forest structures in order to support species dependent on those systems. Due to intensive timber management practices, much of the forested area is in early to mid seral stages of stand development. The overall mission of the Habitat Conservation Plan aims to implement restoration strategies that accelerate stands towards late seral conditions.
This project will identify areas suitable for restoration within the riparian areas of the Watershed using two distinct methods of analyses. The first method will characterize the riparian forest cover in the Cedar River Watershed into four classes: early seral conifer, mid seral conifer, hardwood dominated, and mixed conifer hardwood using a combination of a Geographic Information System (GIS), remote sensing and ground-truthing techniques. Four field plots for each cover class will be set up, and stand measurements will be gathered to determine the general characteristics of each plot. These plots will then be located in remote sensing imagery and software will be programmed to recognize similar stands, based on their spectral characteristics across the landscape. The image analysis will produce a map of similar stands across the Watershed image for each of the four classes. A random selection of the predicted riparian forest cover classes obtained from the image analysis will be ground-truthed by setting up plots in each stand type and taking identical stand measurements to the ones obtained earlier. After this field data has been collected, a statistical analysis will be conducted to determine the similarity between stand measurements in the field versus remote image observations.
The second goal of this study is to determine geomorphological characteristics of the channel and floodplain to discover potential natural disturbance hazards. GIS terrain modeling will be used to derive percent slope from Digital Elevation Models, which can determine areas prone to failure (steeper than 35%). Floodplains that are relatively flat and wide that could be prone to channel migration will also be determined using GIS. This would be a suitability analysis by which areas of potential restoration could be located that are not prone to natural disturbance based on geomorphic conditions. These areas would then be considered most suitable for restoration.
By combining the results from these two analyses, managers will be able
to locate riparian areas of a given cover class not prone to geomorphic
failure. This would provide a decision rule with which to locate areas
for potential restoration.