National scale forest cover and change assessment using the Landsat data archive

Peter Potapov, University of Maryland, College Park, peter.potapov@hermes.geog.umd.edu (Presenter)
Svetlana Turubanova, University of Maryland, College Park, paleobase@gmail.com (Presenter)
Matthew Hansen, S. Dakota State University, matthew.hansen@sdstate.edu

Satellite imagery are the only viable data source currently available for the quantification of forest cover and loss within the largest forest countries. Given the vast extent and heterogeneity of forest landscapes, lack of transportation infrastructure, and political instability in some regions, ground-based methods are impractical for many countries. Landsat imagery have become a primary data source for national-scale forest monitoring after the implementation in 2008 of a new Landsat Data Distribution Policy that provides Landsat data free of charge. Recent progress in automated Landsat data processing and mosaicing to produce cloud-free annual or epochal composite images has enabled Landsat-based monitoring for large, typically cloudy forest regions. We present a semi-automatic Landsat-based approach for forest mapping and monitoring at national and regional scales. Our approach is based on an exhaustive mining of the Landsat archive data to select all available observations over the area of analysis with limited cloud cover. The image process relies on USGS EROS Landsat data processed to the L1T level, with subsequent image radiometric calibration, normalization using MODIS atmospherically corrected reflectance data, and per-pixel quality assessment. Individual Landsat images were used to derive multi-temporal metrics and multi-year composites which were integrated with MODIS time series data for regional-scale forest cover and change mapping. The proposed approach has been successfully tested over large forest nations of Democratic Republic of the Congo, Republic of the Congo, Mexico, Indonesia, European Russia, etc. We provided an overview of several national-scale decadal and mid-decadal analyses performed using our automated data processing algorithm.

Presentation Type:  Poster

Session:  Global Change Impact & Vulnerability   (Tue 11:30 AM)

Associated Project(s): 

• Hansen, Matthew: Establishing a global forest monitoring capability using multi-resolution and multi-temporal remotely sensed data sets ...details

Poster Location ID: 277