Planning, Operations, Design and Research Each Need a Different Fire Model
Run time is only half the choice; the other half is what the answer is for. Planning work — where to put a fuel treatment, which neighbourhoods are exposed — runs the same landscape many thousands of times over synthetic seasons and tolerates long runs because nobody is waiting. Operational work asks where this fire will be at six o'clock and must answer inside the incident's decision cycle, which favours a fast empirical model tuned against what this fire has already done. Building and system design asks whether smoke reaches a detector before people finish leaving a room, and is answered once, carefully, with a fine-scale physics model. Research asks whether the physics is right at all and is judged on fidelity alone. The test is three questions: what decision does this feed, when is it due, and is it made once or ten thousand times? The confusion this resolves is that a model described as state of the art is often the wrong tool operationally, and a model dismissed as crude is often the right one. After this Concept you can classify a fire question by its purpose and say which class of model that purpose rules out.
This Concept is waiting for its first lesson!
Run time is only half the choice; the other half is what the answer is for. Planning work — where to put a fuel treatment, which neighbourhoods are exposed — runs the same landscape many thousands of times over synthetic seasons and tolerates long runs because nobody is waiting. Operational work asks where this fire will be at six o'clock and must answer inside the incident's decision cycle, which favours a fast empirical model tuned against what this fire has already done. Building and system design asks whether smoke reaches a detector before people finish leaving a room, and is answered once, carefully, with a fine-scale physics model. Research asks whether the physics is right at all and is judged on fidelity alone. The test is three questions: what decision does this feed, when is it due, and is it made once or ten thousand times? The confusion this resolves is that a model described as state of the art is often the wrong tool operationally, and a model dismissed as crude is often the right one. After this Concept you can classify a fire question by its purpose and say which class of model that purpose rules out.
Are you a teacher? Sign in to start contributing.
Sign In