Geographic Science

SCIENCE TOPICS

• • Alaska Geospatial Data
  • Alaska Volcano Observatory
  • Biology
  • Earth Science Information Center
  • Geography
  • Geology
  • Water

ABOUT THE ALASKA
SCIENCE CENTER

• • Alaska Science Center Fact Sheet (2007)
  • Office Information
  • Employee Directory
  • Employment
  • Invitation to Bid
  • USGS Phonebook

Landscape Effects of Fire Frequency and Severity on Boreal Alaskan Landscapes

Experts believe that catastrophic damage from wildland fire probably will continue to increase across the nation and large fires in recent years have caused concern with land managers as to how to better assess the impacts of fire on landscapes. This is especially true for Alaska’s boreal region, where fire frequency, size, and severity appear to be increasing since the 1980’s, with 2004 being the largest fire season since records were kept beginning in the early 1950's. If this trend continues, research suggests that fire frequency will outpace the ability for the forest to replace itself. In order to provide a more thorough assessment of current fires, and prediction of future fire threats, historical information is often of primary importance to allow for proper scenario modeling. Although extent, location, and frequency are fairly easy to identify using fire history records and historical and recent Landsat imagery, the identification, quantification, and mapping of burn severity has only recently been successfully mapped using Landsat 5 Thematic Mapper (TM; website)and Landsat 7 Enhanced Thematic Mapper (ETM+;website) imagery. These data however, only go back to the early 1980's, being dependent on recent radiometric cross-calibration of the Landsat 7 ETM+ and Landsat 5 TM sensors. No work to date, however, has been made in performing a cross calibration of the early Landsat Multispectral Scanner (MSS) imagery to the later TM/ETM+ imagery that would allow a burn severity record that could potentially span over 30 years, giving land managers and researchers an invaluable burn data record that could assess changes over a number of common fire metrics such as fire frequency, size, location, direction, and severity. With this information, researchers could correlate those fire metrics to changes in climate data over the same period; relate the metrics to Normalized Difference Vegetation Index (NDVI) and fire danger rating systems (such as the Canadian Forest Fire Danger Rating System); and potentially model future metrics based on past and current fire qualities over the landscape and current and projected future climate scenarios. This would ultimately allow better development of strategies for the preparation of fire management plans and the assessment of potential wildland fire risks and strategic decisions on preparedness, suppression, the location and design of potential fuel reduction projects, and other land management activities (such as habitat assessments).

The first object will be to map burn severity using the Differenced Normalized Burn Ratio (DNBR) at 30m spatial resolution for all fires over the Yukon River Basin (including that portion that extends into Canada) meeting a minimum size requirement of 1000 acres or greater. The DNBR approach will be used to map burn severity and analyze fire trends for fires going back to 1973, corresponding to the advent of the first Landsat series of satellites and continuing to 2006 with Landsat 7. Once the burn severity maps are completed, the type and amount of severity will be assessed across the 30 year temporal time frame to assess whether or not fires occurring from the 1970’s to the present are more or less severe than they are currently. This information, along with changes to vegetation (from early succession to present where possible) will be analyzed with climate data to predict potential future burn severity and periodicity of fires in the area along with potential future vegetation types.

The change in spectral qualities between the pre-1980’s Landsat MSS data and the post-1980’s Landsat TM and ETM data, thus, the sensor calibration between the two satellites will have to be conducted. Thus, for this study, one of the primary goals is to conduct research that allow calibration between the early Landsat MSS sensor and later Landsat TM and ETM sensors that allows for sufficient determination and temporal analysis of burn severity; that is radiometric cross calibration between Landsat 1, 2, and 3 MSS data, Landsat 4 MSS and TM data, Landsat 5 TM data, and Landsat ETM+ data.

Principle Investigator: Crystal Kolden (ckolden@usgs.gov, 907 786 7023)