REFINING OUR UNDERSTANDING OF CHARCOAL SOURCE AREA TO BETTER RECONSTRUCT FIRE HISTORY

In recent decades, charcoal records from lake-sediment cores have become the primary dataset for reconstructing Holocene fire history. Comparison of charcoal data with pollen and other paleoclimate proxies has provided important insights into fire-climate-vegetation linkages as well as the role of anthropogenic burning in the past. It is well established that analysis of macroscopic charcoal particles (>100 micron in diameter) offers information on local fires because large particles are not carried far from the fire source area. However, our understanding of “local” has been poorly constrained. Macroscopic charcoal records are generally decomposed into a slowly varying trend (the background component) and the positive deviations above that trend (the charcoal peaks component) (Long et al. 1998, Can. J. Forest Res. 28). Variations in background charcoal are attributed to shifts in fuel biomass levels over some area, and charcoal peaks are interpreted as fire episodes (i.e., one or more fires occurring close to the lake) and used to calculate fire frequency. It has been unclear whether these components record fire at the same scale and to what degree the primary fire signal is obscured by secondary non-fire-related processes of transport and deposition. Yellowstone National Park has a rich record of historic fires for undertaking data-model comparisons and more closely examining the source of macroscopic charcoal. Five 500-year-long charcoal records were compared with historical fire databases and the results of a process-based simulation model. It was found that background charcoal registered area burned within a 10-km radius of the study sites; an area about the same size as the pollen source area. Charcoal peaks matched known historic fires within <2 km radius of the study sites and showed little relation to fire episodes at greater distances. That these two charcoal metrics (charcoal background and peaks) reflect fire activity at different spatial scales may explain why (1) trends in charcoal abundance (related to area burned) match shifts in pollen assemblages (reflecting fire-related vegetation change), and (2) peak frequency data often show a poor correspondence with pollen data.