Empirical fire models, designed for public messaging of fire danger, prescribed-burning decision making, resource management, and fire-spread-rate prediction, are often used in climate studies to investigate potential changes in recent and future wildfires and prescribed burning opportunities. Prior studies have shown that some of these models are more accurate than others, which prompts the question: How sensitive are fire-climate projections to model choice? We consider the potential sensitivity of climate projections to one element common to these models: their functions of atmospheric temperature (T) and relative humidity (RH), termed thermodynamic functions. Thermodynamic functions in 16 models, including operational fire rate-of-spread models and popular fire-weather indices, are presented and compared graphically over broad ranges of T (0—50 °C) and RH (5—75 %). Substantial variations in thermodynamic function between models are found across this parameter space, with one model not quite doubling and others varying by more than 30 times. Substantial variations in the relative importance of T and RH are also found between models, with one model up to seven times more sensitive to T than RH, and other models about six times more sensitive to RH than T. Three hypothetical climate change scenarios were applied to the 16 thermodynamic functions, yielding projections ranging from minimal change to a more than 50 % increase in value. Non-trivial differences in projections were also found between models designed for similar fuels. This projection variability highlights the need for careful model consideration in climate studies, and the importance of acknowledging projection sensitivity to model choice, especially as further developments in models are likely.

中文翻译：

---

火灾天气气候预测对经验火灾模型的敏感性

经验火灾模型专为火灾危险的公共信息、规定燃烧决策、资源管理和火灾蔓延率预测而设计，通常用于气候研究，以调查近期和未来野火和规定燃烧机会的潜在变化。先前的研究表明，其中一些模型比其他模型更准确，这就提出了一个问题：火灾气候预测对模型选择有多敏感？我们考虑气候预测对这些模型共有的一个要素的潜在敏感性：大气温度 (T) 和相对湿度 (RH) 的函数，称为热力学函数。16 个模型的热力学函数，包括运行火灾蔓延速度模型和流行的火灾天气指数，在 T (0—50 °C) 和 RH (5—75 %) 的广泛范围内以图形方式进行了展示和比较。在这个参数空间中，我们发现模型之间的热力学函数存在显着差异，其中一个模型没有完全翻倍，而其他模型则变化超过 30 倍。不同模型之间也发现 T 和 RH 的相对重要性存在显着差异，其中一个模型对 T 的敏感度比 RH 的敏感度高出七倍，而其他模型对 RH 的敏感度比 T 的敏感度高出约六倍。应用了三种假设的气候变化情景16 个热力学函数，产生的预测范围从最小变化到价值增加超过 50%。在为类似燃料设计的模型之间也发现了不小的预测差异。这种预测的可变性凸显了在气候研究中仔细考虑模型的必要性，以及承认预测对模型选择的敏感性的重要性，特别是在模型可能进一步发展的情况下。