Evaluation of CONUS Tree Regrowth using LCMAP C1.0 Products

Abstract:

The USGS Land Change Monitoring, Assessment, and Projection (LCMAP) project has released a suite of annual land cover land change products (1985-2017) at the 30-meter scale over the conterminous United States (CONUS). The product provides land change information critical to understanding spatial and temporal characteristics related to the type, intensity, and location of land cover changes. This study investigates the tree regrowth rate using LCMAP land cover data. The product has been validated; the tree producer’s and user’s accuracy is 83% and 90%, and the grass/shrub producer’s and user’s accuracy is 80% and 88%, respectively.

Tree regrowth, after disturbances, plays an important role in understanding the global carbon budget and management of the forest resource. In this study, disturbances refer to events (e.g., fire, clear cut, etc.) that cause land cover change from tree to grass/shrub. We calculated the average number of years for trees to grow back. We also analyzed regrowth time variation over different categories of all forest types, land ownership, and disturbance types. We found that Loblolly pine, Slash pine, and White oak/red oak/hickory are the three major disturbed forest types at the national scale. These forest types are mostly distributed in the southeast part of CONUS and typically grow back in 5 years. Forest cut was a major cause of disturbance. Most of the forest regrowth in this region occurred on non-public land. In the west part of CONUS, Douglas fir is the most commonly disturbed forest type, and the regrowth time varies from 10 to 15 years. Common disturbances were fire and cut in the west part of CONUS. Additionally, we found that most of the changes from tree to grass were also captured by the LCMAP spectral change product, while most of the changes from grass to tree were only captured by the land cover map. This is because the tree to grass changes are usually abrupt, but grass to tree changes are usually gradual. This difference suggests that the uncertainty of the regrowth time estimation mainly comes from the change from grass to tree and depends on the classification accuracy. Understanding the spatial-temporal patterns of forest regrowth will provide vital information for effective management in supporting local economies, maintenance biodiversity, and carbon accumulation.