Modelling the Spatial Variability of Fuel Moisture Across a Heterogeneous Forested Landscape

Abstract:

Fuel moisture pays an important role in determining the intensity, behaviour and spatial patterns of forest fires and can vary significantly across a forested landscape. We present a set of models designed to quantify the spatial patterns of near surface drying potential across a landscape with significant variability in above canopy radiation load and canopy coverage. We examined the degree to which these two factors drive drying potential and whether relatively wet and non-burnable areas persist in the landscape during the fire season. Specifically, we measured near-surface temperature and humidity at sites across the landscape, as well as precipitation, solar radiation, and wind speed at a subset of these sites. These observations were then used to evaluate, train, and drive a fuel moisture model that includes sub-components for canopy interception of both precipitation and radiation. The precipitation interception is modelled using a Rutter model approach that accounts for evaporation from the canopy. Radiation interception was modelled using hemispherical photos of the canopy combined with modelled sun locations and ratios of direct to diffuse radiation. A novel model was developed to simulate the moisture of standardized fuel sticks at four different size classes and was evaluated using an independent dataset. Finally, the modelled fuel moisture was used to estimate potential fire-line intensity and ignition probability. Preliminary results indicate that downwelling longwave radiation from the canopy is counteracted by less radiation during the day. This balance leads to smaller spatial variability of daytime minimum fuel moisture which is influenced by both night time and daytime conditions due to moisture storage.