Measuring the Boreal Forest’s Breadth19 May, 2008 By: Kenneth Wong
Remote sensing experts monitor the shrinking boreal forest with GIS.
Is the world's forest exhaling more carbon than before? If so, how much has the volume changed over time? To determine that, some scientists are measuring the changing mass of the boreal forest biome, covering more than 14% of the total land surface. Gregory J. Fiske, a research associate at the Woods Hole Research Center (WHRC), is among them. He uses a method known as remote sensing.
To those outside the scientific community, the term might suggest a paranormal process, perhaps someone with extrasensory perception telepathically receiving messages about places and people far away. At WHRC, it means studying the ecological changes in a region using remote-detection devices.
For his job, Fiske regularly turns to a variety of remote-sensing datasets, such as those of Landsat, produced by the earth-observing satellite missions jointly operated by NASA and the U.S. Geological Survey, and The Advanced Very High Resolution Radiometer (AVHRR), archived at the U.S. Geological Survey's EROS Data Center in Sioux Falls, South Dakota.
"When remote sensing experts come to the WHRC, they often bring along a desire to use their own favorite software — ERDAS Imagine, ENVI/IDL, PCI Geomatics, or open source options like R statistical programming language," observed Fiske. "ESRI's ArcGIS is a connecting link between all of these. At the end of a big project, we still have to make a map to display the results of a spatial analysis."
John P. Holdren, director of the WHRC, noted, "Among a text-weary public, maps are perhaps the most direct route to understanding research results and the status of the earth's natural resources."
Trial by Fire
"Fire disturbance plays a dominant role in boreal ecosystems, altering forest succession, biogeochemical cycling, and carbon sequestration," according to WHRC's research paper "Using satellite time-series data sets to analyze fire disturbance and forest recovery across Canada" (Scott J. Goetz, Gregory J. Fiske, and Andrew G. Bunn, Woods Hole Research Center, January 2006). "We used two time-series data sets of Advanced Very High Resolution Radiometer (AVHRR) Normalized Differenced Vegetation Index (NDVI) imagery for North America to analyze vegetation recovery after fire."
The fire disturbance analysis was the beginning of a series of remote sensing work that led to a key finding in boreal vegetation trends, Fiske pointed out. (For more, read, "Satellite observed photosynthetic trends across boreal North America associated with climate and fire disturbance," Proceedings of the National Academy of Sciences, S.J. Goetz, A. Bunn, G. Fiske, and R.A. Houghton).
The Canadian Forest Service Large Fire Database furnished the researchers with the location of the fires over time. Fiske and his colleagues removed the areas covered by lakes and other water bodies from consideration, using a surface water dataset for Canada.
"We then developed a GIS protocol to run within the ESRI ArcGIS environment to summarize and analyze the NDVI values within the burned area boundaries for all fire years (1981-1997)," wrote Goetz and his coauthors. "This procedure ran an iterative process that selected a single fire year and calculated the mean per-pixel NDVI for all 540 15-day periods in the GIMMS [Global Inventory Modeling and Mapping Studies] 22