| Popis: |
In this work, a collection of ten wildland fire rate of spread prediction models that take into account the effects of wind are reviewed and tested against 166 individual laboratory fire tests, available in the open literature. The investigated models include the well-known semi-empirical models of Rothermel, Wilson and Catchpole et al., the empirical models of Rossa and Fernandes, developed using laboratory fire tests and the empirical models of Burrows et al., Anderson et al., Fernandes et al. and the Canadian Forest Fire Behavior Prediction System, developed using field measurements. The performance of the ten models is evaluated, both qualitatively and quantitatively, by employing a range of dedicated statistical error metrics. It is shown that the performance of each model is affected by their specific characteristics, in conjunction with the characteristics of the experiments against which the models were evaluated. It is found that the model of Catchpole et. al. yields the lowest statistical error metric values. The empirical models that have been developed using field measurements exhibit significant discrepancies against the experimental data, due to the use of specific parameters regarding fuel type, scale and wind speed. |
