Willamette Valley Conservation Study Vegetation Types

Spatial elements: A Landsat image from 2010 was segmented using eCognition. The size of the polygons that were generated from eCognition were based upon a scale factor of 25 and a preference for color over shape. This produced about 250,000 different polygons throughout the area. However, the size of the polygons were often too large to depict fine level features of interest to the study such as wetlands. To address this, we used a modified version of the National Wetlands Inventory that was developed by the Institute for Natural Resources in Portland, OR. These modified NWI polygons were essentially "burnt in" to the segmented polygons.

Thematic classes: NWI polygons were attributed with the classes that were part of the NWI dataset. These classes were eventually crosswalked into a "vegetation class" type that relates to an ecological system. Other segments were attributed into various classes based upon previous vegetation and land cover maps such as NLCD, Crop Data Layer, NOAA C-CAP, Northwest Habitat Institute (NWHI), and any other ancillary information. Sometimes LiDAR derived vegetation height data was used to discriminate classes. Each polygon has a relative certainty associated with it based upon the overlap with other data sets. There is also a field called criteria that describes the logic and reasoning used to make this land cover call for the particular class.