D3: Oral Presentation C Thursday, 11:30 am � 12:00 pm C Lake Superior Ballroom MN

New Technologies in Operational Forest Mapping

D. Samuel Rajasekhar
Minnesota Department of Natural Resources
413 Southeast 13th Street
Grand Rapids, MN  55744
samuel.rajasekhar@dnr.state.mn.us
 

This project is for testing the usefulness of multispectral satellite imagery for practical forest management. It is a reconnaissance-level detection and mapping of upland conifer stands in Itasca County, MN, a heavily forested county 2900 square miles in extent. In further stages of the project, stocking estimates will be developed and assigned to stands.

In keeping with the project's aim of developing robust methods suitable for everyday employment by forest managers, the analysis used existing National Wetlands Inventory data to define upland areas, and two dates of Landsat 7 Enhanced Thematic Mapper imagery � one taken in the growing season of 1999 and another taken before deciduous leafout in the spring of 2000 � to highlight conifers among other forest cover.

Picture elements (pixels) thus defined as representing upland conifers were then assigned to species groups by computer classification of reflectance patterns, based on signatures developed in stands of known composition in DNR's Cooperative Stand Assessment inventory. As the emphasis in this study is detection of manageable stands rather than mapping of scattered trees. This project has potential to expand to other commercial trees like aspen, etc.