Development of a synthetic stream network using binary logistic regression

Site Archive
You are viewing a static copy of the 2013 Sunriver Conference website archived on December 11, 2013. To view current Northwest GIS User Group events and news visit

Stream networks in western Washington are extensive and their associated riparian management zones represent a considerable portion of the land base. Accurate mapping of stream networks is important for effective and efficient forest land management, environmental assessments, and forest land planning. Current statewide stream mapping systems suffer from a variety of well-known errors, such as inconsistent mapping scale, errors of omission, and incorrect stream typing. We present a summary of our efforts to develop a synthetic, typed stream network using field observations, LiDAR-based digital terrain models, and a statistical technique known as binary logistic regression. We describe our methodology for model construction and present preliminary results. Our efforts indicate that a LiDAR-based synthetic stream layer is generally superior to current stream mapping systems, and could be used to build accurate estimates of the location and extent of the stream network and its associated riparian management zone in western Washington.

Chris Snyder and Jeff Ricklefs

Chris Snyder is a GIS Analyst/Programmer at the WA Department of Natural Resources and has worked in the utility and environmental GIS field for 15 years. His area of expertise is the automation of GIS and database processes. Jeff Ricklefs is a GIS Analyst/Programmer at the WA Department of Natural Resources. His work focuses on research and data analysis in support of resource management. His interests and area of expertise include the ecology and management of riparian systems.

Landmark 1
October 17, 2013 - 11:00am - 11:30am
Your rating: None Average: 4.5 (13 votes)