Abstract: |
The proposed work defines the architecture of an overall system, based on system modelling tools and optimization techniques, which has been developed with the purpose of mitigating forest fires risk over a wide regional or even national area. The goal of the system is twofold. First, a static hazard assessment is carried out, taking into account the distribution of the hazard over the territory, on the basis of topographic information and land use (including vegetational cover), average weather conditions (i.e., climate conditions), obviously considered with reference to the different seasons, and average fuel (i.e., vegetation) conditions, again referred to the various seasons. Static hazard assessment is also based on the analysis of the data available from historical series corresponding to forest fires actually detected in the considered region. The purpose of the static hazard assessment is that of planning the sizing and the location of the different kinds of resources (men, trucks, engines, aircrafts, infrastructures, etc.) necessary to manage forest fire risk over a wide territory. Another objective of such an analysis could be that of obtaining indications about land use and territorial planning, over a small-medium regional area. The system assumes that besides human life and activities, also vegetational, faunal and hydrological equilibrium are targets to shelter from fire. Actually, the environmental and economic costs caused by a fire, considering also the operational cost of extinguishing intervention and the costs of subsequent land-reclamation, are often unsustainable for local administrations, which usually have access to limited economical resources and technical facilities. The second goal, which can be referred to dynamic hazard assessment, is relevant to the assumption that real-time information is available, as well as reliable meteorological forecasts over a time horizon of a certain length (say 2-3 days). Along with such forecasts, the real-time information used for dynamic hazard assessment may come from different sources, for instance: present weather conditions, ground-measured data relevant to vegetational conditions, data coming from satellite or airborne sensors (again, mainly referring to vegetational conditions). The main objective of a dynamic hazard assessment is that of identifying, within the considered territory, the areas affected by the highest hazard, and the time intervals, within the considered time horizon, in which this hazard takes place. On this basis, one can think of taking a number of pre-operational decisions that can reduce the impact of potentially lighted fire over the considered territory, within the considered time horizon. Such actions may include, for instance, re-locating the available resources over the territory, recalling day-off resources to service, alerting local authorities or emergency managers, issuing prohibitions of some dangerous agricultural practices (such as stubble burning), and patrolling the areas affected by the highest hazard. Instead of taking decisions only on the basis of the hazard distribution (both in the static and in the dynamic case), one can think also of introducing specific information regarding the land use, and particularly of the so-called exposed elements, present over the territory. Such a term can be used to refer to linear, point-wise, or areal elements, which can be injured by fire. For any of such elements, the vulnerability (i.e., the kind of the element response to the external stress) has to be evaluate, along with its economic or environmental value. Of course, the vulnerability of an element is a function of the layout of the extinguishing resources over the territory. Eventually, taking into account the hazard distribution, as well as the value and the vulnerability of each exposed element, it is possible to evaluate the risk distribution over the territory. On this basis, different optimization problems can be stated and solved, both for the static and the dynamic case, aiming at minimizing either the maximum or the average risk over the considered territory. An application of the system is described over the whole Italian territory, in order to demonstrate the effectiveness of the proposed approach. |
