Despite the fact that high-intensity crown fires account for an overwhelming proportion of the area burned by forest fires in Canada, fully understanding and subsequently modeling the initiation, propagation, and spread of crown fires remains an elusive goal for fire research scientists in this country and throughout the world. The Canadian approach to the prediction of crown fire phenomena has been largely empirical: an extensive experimental burning program (including many high-intensity crown fires) in major fuel types, accompanied by the monitoring of numerous wildfires, has led to the creation of a large database that is the foundation of the system of fire behavior prediction in Canada. Conversely, fire behavior prediction in the United States has been largely based on a physical surface fire model, which has proven inadequate for predicting high-intensity fire behavior. The International Crown Fire Modeling Experiment (ICFME) was a cooperative international experiment that brought together fire modeling experts from Canada, the United States, and Russia, to address the prediction of high-intensity fire behavior. The goal of the ICFME was to conduct a replicated series of highly instrumented crown fires to quantify parameters essential to modeling the initiation and spread of crowning fires. The study site was located near Fort Providence, NWT, in a dense, approximately 80-year old jack pine stand. Aerial, surface, and forest floor fuels were sampled in ten burn plots. Firelines approximately 50 m wide were established around each plot, which involved cutting and removing standing trees, and bulldozing to mineral soil to facilitate access and control. Some fuel manipulation (pruning trees and/or removing surface fuel) was carried out on portions of some plots, but most of the area remained undisturbed. The ICFME project was carried out between 1995 and 2001.