Continental Mapping of Northern Eurasia Forest Disturbance Using MODIS Products

The capabilities of the MODerate resolution Imaging Spectroradiometer (MODIS) present some exciting possibilities for improved and timely monitoring of forest disturbance at different scales. At a continental scale, several quantitative indices are used in this research to map forest disturbance including fire, clearcuts, and possibly insect damages for Northern Eurasia (between 50N and 70N). The tasseled cap greenness and normalized difference water index are derived from Nadir BRDF-Adjusted Reflectance (NBAR) to map burned- area. The NBAR-based indices have several advantages for this application: (1) standardized reflectance with a unique nadir view and can minimize potential inaccuracy related to variable geometry, and (2) reduced data redundancy and improved vegetation signal. After the QA filtering and temporal smoothing, growing-season vegetation anomalies are calculated to quantify the percentage of the climatological annual grid-point growth either gained or lost during a given year. With the knowledge that major disturbances tend to result in lower vegetation greenness, a threshold-based algorithm can identify burned area for each year. Our preliminary estimates indicate three percent of the 7-million km2 forests were burned from 2000 to 2006, with 70% of this activity occurring in 2003. High-resolution ETM data are used to validate the algorithm using a pixel-based comparison at Sakha. About 80-90% of the ETM burned pixels can be detected by this algorithm. At regional scale, 250-m MODIS NDVI is used to map recent clearcuts based on the difference between growing-season accumulated NDVI for different years. Although MODIS product can identify recent clearcuts at some locations, MODIS does not appear to be good source to map fine-scale clearcuts for all of Northern Eurasia, especially for heterogeneous regions when natural forests are interspersed with human activities. Finally, by determining the post-burned vegetation growth as a function of time, a quantitative index can provide useful information on re- growth rate for different burned regions. Overall, the MODIS product suite is a useful tool in widespread forest disturbance mapping at continental scales. Our work, part of the NELDA (Northern Eurasia Landcover Dynamics Analysis) project, is preparing a continental map of forest disturbance for Northern Eurasia - a preliminary version of this map will be presented here. Such maps will provide a fundamental understanding of the changing patterns of forest disturbance as well as the impacts of climate dynamics and land use changes on forest ecosystems