2020 ESA Annual Meeting (August 3 - 6)

COS 53 Abstract - The Landscape Change Monitoring System: Leveraging the Landsat archive to map our dynamic lands

Ian Housman1, Leah S. Campbell1, Mark Finco1, Kevin Megown2, Nathan A. Pugh2, Yang Zhiqiang3 and Sean P. Healey3, (1)RedCastle Resources, Onsite contractor to the USDA Forest Service Geospatial Technology and Applications Center, Salt Lake City, UT, (2)USDA Forest Service Geospatial Technology and Applications Center, Salt Lake City, UT, (3)USDA Forest Service Rocky Mountain Research Station, Fort Collins, CO
Background/Question/Methods

The United States Forest Service (USFS) Rocky Mountain Research Station and Geospatial Technology and Applications Center are developing the Landscape Change Monitoring System (LCMS) to leverage freely available remote sensing data and Google Earth Engine to monitor changes across the US from 1984 to present. The goal of this project was to pilot the methods across several National Forests, work with land managers to utilize LCMS products to help inform their work, and develop a web map application to present, explore, and deliver products.

LCMS utilizes annual growing season composites of Landsat TM , ETM+, and OLI. Metrics of change are provided by the LANDTRENDR and VERDET change algorithms. These metrics include the magnitude, duration, and slope of the fitted trajectories from each algorithm. Model calibration/validation data are provided from Timesync, a tool that allows analysts to manually inspect various sources of image time series data to attribute land cover, land use, and change processes for each year.

Timesync data and the LANDTRENDR and VERDET change metrics are used in three Random Forest models to produce annual model confidence of vegetation cover loss and gain, and life-form level land cover classes.

Results/Conclusions

This workflow has been applied across four pilot Forests and the USFS Intermountain Region in preparation for a conterminous United States-wide product in early 2021. Model accuracy has exceeded 85% for all products for all pilot areas. Somewhat surprisingly, the National Forests participating in the pilot project expressed greater interest in the vegetation gain and land cover products, which provide information presently lacking. Example applications include identifying areas with limited vegetation gain after fires in the Bridger-Teton National Forest and monitoring glacial recession and post-glacial vegetation succession over the Chugach National Forest.

The web map application (http://lcms.forestry.oregonstate.edu) presents products from each pilot area. It provides a suite of tools to help the user map, summarize, and download the LCMS products across different areas. We will be applying what we have learned from these pilot areas across the conterminous United States during 2020.