Improving fire season definition by optimized temporal modelling of daily human-caused ignitions. (1st July 2018)

• Record Type: Journal Article
• Authors: Costafreda-Aumedes, S.; Vega-Garcia, C.; Comas, C.
• Is Part Of: Journal of environmental management. Volume 217(2018)
• Page Start: 90
• Page End: 99
• Publication Date: 2018-07-01
• Subjects: Day-of-year modelling -- Human-caused fires -- Logistic regression -- Spatial pattern -- Ignition-based seasons; Environmental policy -- Periodicals; Environmental management -- Periodicals; Environment -- Periodicals; Ecology -- Periodicals; 363.705
• DOI: 10.1016/j.jenvman.2018.03.080
• Languages: English
• ISSNs: 0301-4797

Abstract

Abstract: Wildfire suppression management is usually based on fast control of all ignitions, especially in highly populated countries with pervasive values-at-risk. To minimize values-at-risk loss by improving response time of suppression resources it is necessary to anticipate ignitions, which are mainly caused by people. Previous studies have found that human-ignition patterns change spatially and temporally depending on socio-economic activities, hence, the deployment of suppression resources along the year should consider these patterns. However, full suppression capacity is operational only within legally established fire seasons, driven by past events and budgets, which limits response capacity and increases damages out of them. The aim of this study was to assess the temporal definition of fire seasons from the perspective of human-ignition patterns for the case study of Spain, where people cause over 95% of fires. Humans engage in activities that use fire as a tool in certain periods within a year, and in locations linked to specific spatial factors. Geographic variables (population, infrastructures, physiography and land uses) were used as explanatory variables for human-ignition patterns. The changing influence of these geographic variables on occurrence along the year was analysed with day-by-day logistic regression models. Daily models were built for all the municipal units in the two climatic regions in Spain (Atlantic and Mediterranean Spain) from 2002 to 2014, and similar models were grouped within continuous periods, designated as ignition-based seasons. We found three ignition-based seasons in the Mediterranean region and five in the Atlantic zones, not coincidental with calendar seasons, but with a high degree of agreement with current legally designated operational fire seasons. Our results suggest that an additional late-winter-early-spring fire season in the Mediterranean area and the extension of this same season in the Atlantic zone should be re-considered for operational purposes in the future. Highlights: Fire season definition can be improved considering human-caused ignition patterns. Physiography, human habitation and land uses seasonally influence fire starts. Day-by-day fire occurrence models allow analysing seasonal trends of ignitions. Trends are identified in Spain to delimit ignition-based seasons in two regions. Longer fire seasons may have to be considered in the future.