My dissertation on the relationship between the Keetch-Byram Drought Index and wildfire size and frequency in selected natural areas of the US has passed the university's formatting review board and has been published online.
The research indicates that the Keech-Byram Drought Index (KBDI) is not as strongly related to wildfire frequency in the Southeast as it is in the western United States. KBDI predicts wildfire potential by modeling the amount of moisture in the top several inches of the soil and duff layers. It is a conventional and widely-used component of wildfire potential models. Few studies exist that compare actual daily KBDI values to historical wildfire occurrence records. Foresters and ecologists have questioned the efficacy of KBDI alone as a predictor of wildfire in the Southeast. The dissertation examined average KBDI values on days when new fires ignited and on days when no new fires were reported at eight natural areas (four in the western United States, four in the southeast) for the period 1980 – 2010. I found consistent separation between non-fire and fire-start daily KBDI values for the four western locations, indicating that the drought index is a good predictor of wildfire potential at those locations. In the Southeast, one location (Piedmont National Wildlife Refuge) showed no difference between non-fire and fire-start daily KBDI. Two other locations showed a slight difference, and the fourth (Great Dismal Swamp National Wildlife Refuge) showed a difference similar to that found in the western states.
This study suggests that the efficacy of KBDI as a predictor of wildfire occurrence is geographically variable and particularly weak in the southeastern mixed forest ecoregion. Breaking the wildfire record out by ecoregion province, the dissertation shows that wildfire in the Southeast does not follow the single (summer) season of burn pattern observed in the west but is instead characterized by increased frequency in late winter/early spring, followed by a dip in activity, then another peak in the late summer. This seasonal pattern has been noted in the state of Mississippi. My analysis shows it extending into Texas, the Midwest, and Appalachia. I suggest that this pattern leads to a tendency for KBDI to underpredict wildfire potential in the winter and overpredict in the summer in the Southeast.
From the standpoint of wilderness conservation, this dissertation supports the notion that state and county level forestry commissions should not rely on KBDI alone in the issuance of burn bans. Fire-dependent ecosystems are generally more efficiently managed when summer burns are implemented rather than winter burns exclusively. More research and a better understanding of the geographical variability in the relationship between soil moisture, fuel moisture, KBDI, and wildfire size and frequency could lead to changes in how KBDI is used locally and could lead to greater flexibility regarding summer burning in Southeastern forests.