Modeling moisture content in shrubs to predict fire risk in Catalonia (Spain)

We measured the moisture content in small branches and leaves of Cistus monspeliensis in four Mediterranean locations of Catalonia (Spain), from 1998 to 2000. The moisture of the measured fraction is also known as live fine fuel moisture, and it provides the most significant fuel in wild fires in this region. The aim of the study was to model the live fine fuel moisture based on meteorological variables from automatic weather stations close-by, and several components of the Canadian index of forest fire danger. We wanted to generalize the estimates of live fine fuel moisture from this species to other localities throughout the region from the data obtained by the network of automatic weather stations. We developed two models. The second model, independent of the actual sampling date, was more general, but the addition of the temporal component improved the predictive ability of the first model. We validated the models with data collected at two locations in 2001 and 2002. The resulting adjustment's reached R-2 of around 80%, for the predicted against the observed values for the test period. Therefore, the methodology developed provides a reliable tool to predict forest fire risk throughout the year. (C) 2002 Elsevier Science B.V. All rights reserved.