Development and participatory evaluation of fireline intensity and flame property models for managed burns on Calluna-dominated heathlands

Prescribed burning plays an important role in the management of many ecosystems and can also be used to mitigate landscape-scale fire risk. Safe and effective application of prescribed fire requires that managers have a robust understanding of potential fire behavior in order to decide on the appropriate tools and tactics for any burning operation. Shrubland ecosystems, including heaths and moors, are known to exhibit intense fire behavior under marginal burning conditions under which fire would not be expected to spread in other vegetation types. This makes developing fire behavior predictions for such systems important. Traditional managed burning is widely used as a tool in Calluna vulgaris (L.) Hull-dominated heath and moorland landscapes in northwest Europe, but in some regions, especially the United Kingdom, there is significant debate over fire use. Despite the controversy, there is general agreement on the need to (1) understand relationships between fuel structure and potential fire behavior, and (2) improve burning practice to optimize potential trade-offs between different ecosystem services. Our aim was to provide knowledge to improve management practice by developing models of potential fireline intensity and flame length. We conducted 27 burns in three developmental stages of Calluna with different stand structures and estimated fireline intensity, flame length, flame height, and flame angle. Flame properties were assessed using photographs and visual observation. We evaluated our models using a participatory research approach for which conservation and land managers submitted basic observations on fire behavior and fire weather for their burns. Fireline intensity and flame height increased significantly across age-related Calluna phases. Regression modeling revealed that fireline intensity could be adequately estimated by a combination of fuel height and wind speed, with taller fuels and higher wind speeds related to more intense fires. Predictions were, however, improved by accounting for live fuel moisture content. Flame length and height were modeled as a function of fireline intensity using standard approaches, but adequately performing models for flame angle could not be established. Evaluation data provided by land managers was noisy, but their qualitative assessments of fire behavior and estimates of flame length were significantly correlated with predictions from our models. Fire intensities and flame properties seen in northern Calluna heathlands are similar to those encountered in shrublands associated with climates and fuels more commonly perceived as representing high fire danger. The results demonstrated that our models perform tolerably well although there is substantial uncertainty in their predictions. The models were used to develop a fire behavior nomogram that can provide an indication of potential fireline intensity and flame length prior to commencing a burn.