ESTIMATING THE RISK OF ESCAPE OF PRESCRIBED FIRES - AN EXPERT-SYSTEM APPROACH

Citation
M. Stock et al., ESTIMATING THE RISK OF ESCAPE OF PRESCRIBED FIRES - AN EXPERT-SYSTEM APPROACH, AI applications, 10(2), 1996, pp. 63-73
Citations number
8
Categorie Soggetti
Environmental Sciences","Computer Science Artificial Intelligence",Forestry,Agriculture
Journal title
ISSN journal
10518266
Volume
10
Issue
2
Year of publication
1996
Pages
63 - 73
Database
ISI
SICI code
1051-8266(1996)10:2<63:ETROEO>2.0.ZU;2-H
Abstract
Prescribed burning expenditures are based on the fire manager's judgme nt about the ''risk'' of the fire escaping and his/her anticipation of the consequences of such an escape. In a high-risk site, more resourc es are needed to prepare the site for a safe burn. If a fire escapes, or if the fire is postponed, additional costs are incurred. Site attri butes, probable consequences, and the individual's experience and inna te risk posture interact to blur the distinction between estimating an d anticipating the escape. Rational allocation of resources requires a clear distinction between the likelihoods of escape and the consequen ces. To evaluate factors affecting the risk of escape separately from consideration of the consequences, and to begin a more diagnostic appr oach to fire risk assessment, we used an expert system approach. Facto rs related to escape potential were evaluated, prioritized, weighted, and encoded into the rules of a prototype expert system which then pro vided recommendations for carrying out a burn safely and effectively. Three test cases supported the logic used in the expert system, and po inted to ways that the expert system could be improved. Experience wit h the system supported previous research on the need to consider the h uman element in estimating fire escape potential.