Indices designed to predict fire-atmosphere interactions have traditionally been based on empirical relationships derived from observations of fire behavior and the atmospheric conditions at the time. Although these empirical indices have shown some degree of utility, they are also prone to unpredictable and often undesirable behavior depending on the time of day, the season, the geographic location, and the prevailing weather regime of a given fire scenario. It is therefore desirable to develop indices that are more directly tied to the atmospheric processes responsible for the observed fire-atmosphere interactions, and by doing so produce forecast products that are more generally able to predict fire-atmosphere interactions at all locations, times, and seasons.

Most of the existing fire-atmosphere interaction indices were developed in the 1970s and early 1980s, before regular real-time access to mesoscale simulation data was commonplace. The indices were therefore derived from rawinsonde data, and generally employed observations from 0000UTC. For most of the United States (particularly in the east), 0000UTC is sufficiently close to sunrise that the surface daytime mixed layer is either just forming or still rapidly evolving. However, most strong fire-atmosphere interactions occur in the afternoon, when the surface mixed layer is well-developed and, in many cases, at or close to its maximum depth for the day. Since the depth and strength of the mixed layer are known to impact fire behavior as well as smoke dispersion, it is reasonable to expect that mesoscale model simulations of mixed-layer processes could be used to produce more powerful indices that address the potential for strong and unpredictable fire-atmosphere interactions.

Using research funding from the congressionally established National Fire Plan, the US Forest Service created regional Fire Consortia for the Advanced Modeling of Meteorology and Smoke (FCAMMS) across the country. The Eastern Area Modeling Consortium (EAMC), located in East Lansing, MI, is addressing the need for new fire-atmosphere interaction indices by using mesoscale model simulations (MM5v3) to improve our understanding of the physical processes involved. Many of the indices and concepts being developed by the EAMC take advantage of the model-simulated mixed layer evolution, both in real time and in case studies of past fire events. This paper will present the conceptual framework and some of the index development that is underway at the EAMC. The indices developed by the EAMC are anticipated, after sufficient testing, to be transferred to fire-weather forecast offices for operational use. Additionally, the conceptual models under development are helping to establish a foundation for additional fire-atmosphere interaction research.