Print PageFrequent anthropogenic wildfires in tropical landscapes dominated by nonnative invasive grasses such as guinea grass (Megathyrsus maximus) pose a significant threat to surrounding ecosystems and developed areas. To better manage fire in tropical ecosystems, improved fuel load estimates and model predictions of fire potential and behavior are urgently needed. Predicting fire behavior is commonly done with fire models (e.g., BehavePlus), but the accuracy and reliability of these models requires a better understanding of the variability in the parameters that most drive wildfires – fuels and weather. The objectives of this study were to: (i) quantify the spatial and temporal variability in live and dead fuel loads and moisture in tropical grasslands dominated by the nonnative grass M. maximus on the Waianae Coast and North Shore of Oahu, Hawaii, and (ii) use this extensive field dataset to develop a custom fuel model for this species. We measured the spatial variability in live and dead fuel loads and moistures in three subsequent years at four sites dominated by M. maximus that span a wide range of environmental conditions. To quantify temporal variability in M. maximus, we sampled fuel loads and moistures bi-weekly at three sites for one year. Using collected field data, we developed a new fuel model using the BehavePlus fire model and tested it against observed fire behavior in M. maximus grasslands.
Results/Conclusions
Live and dead fine fuel loads ranged from 0.85 to 8.66 Mg ha-1 and 1.50 to 27.74 Mg ha-1, respectively, and varied significantly by site (p=0.03), but not between years (p=0.34). For temporal sampling, total fuel loads did not differ significantly over time. Live fuel moisture, however, ranged from 67% in the summer to >250% in the winter, and dead fuel moisture ranged from <10% in summer to 48% in winter. Using the BehavePlus fire modeling program parameterized with custom inputs from collected field data, predicted wildfire rates of spread for measured summer moisture scenarios, when fire risk is highest, ranged from 0.18-0.32 m s-1. These initial modeling results corresponded well with actual headfire rates of spread observed in M. maximus grasslands in our study site (0.04-0.35 m s-1). These results suggest that a fuel model created from intensive field measurements of fuel loads and moisture can greatly improve the accuracy of fire behavior modeling, thereby increasing capacity for land managers to make real-time predictions of fire behavior in M. maximus ecosystems in Hawaii and throughout the tropics.
