The fire danger (FD) defines the conditions less or more favourable for a fire ignition success and its propagation. FD indexes, that integrates environmental variables related to FD in more or less complex equations and systems, are widely used in wildfire prone countries for both scientific and operational purposes. Assessing the performance of FD indexes is challenging and this issue is quite debated within the fire community, which has been trying to apply several methodologies to evaluate FD indexes. The main aim of this work is to give a contribution to this effort. The analysis was conducted using data from a fire-prone Mediterranean area (Sardinia island, Italy), where 8 FD indexes were evaluated and compared using different statistical approaches. We calculated the daily FD values for the period 2000-2007 over the study area. A set of statistical tools (namely Spearman rank correlation, Index Value Distribution and Percentile Analysis, and Logistic Regression) were applied to evaluate the performance of each FD index by comparing FD values with fire occurrence indicators. The statistical tests revealed a large variability in FD indexes performance, depending also on fire activity conditions. Our results showed that two of the tested FD indexes reached a good overall performance. Findings from this study can help both the scientific community and local fire managers, supporting the evaluation of early warning systems and fire prevention strategies in the Mediterranean Basin
Assessing the performance of fire danger indexes in Mediterranean area
Michele Salis;Bachisio Arca;Pierpaolo Duce;
2018
Abstract
The fire danger (FD) defines the conditions less or more favourable for a fire ignition success and its propagation. FD indexes, that integrates environmental variables related to FD in more or less complex equations and systems, are widely used in wildfire prone countries for both scientific and operational purposes. Assessing the performance of FD indexes is challenging and this issue is quite debated within the fire community, which has been trying to apply several methodologies to evaluate FD indexes. The main aim of this work is to give a contribution to this effort. The analysis was conducted using data from a fire-prone Mediterranean area (Sardinia island, Italy), where 8 FD indexes were evaluated and compared using different statistical approaches. We calculated the daily FD values for the period 2000-2007 over the study area. A set of statistical tools (namely Spearman rank correlation, Index Value Distribution and Percentile Analysis, and Logistic Regression) were applied to evaluate the performance of each FD index by comparing FD values with fire occurrence indicators. The statistical tests revealed a large variability in FD indexes performance, depending also on fire activity conditions. Our results showed that two of the tested FD indexes reached a good overall performance. Findings from this study can help both the scientific community and local fire managers, supporting the evaluation of early warning systems and fire prevention strategies in the Mediterranean BasinI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.