Forecasting Resource‐Allocation Decisions Under Climate Uncertainty: Fire Suppression with Assessment of Net Benefits of Research
提出一种单阶段预测方法，用于气候不确定性下的资源分配决策，以美国林务局的火灾季前消防资源合同决策为例，展示了该方法相比两阶段模型和现有决策能大幅降低人员成本，效益成本比达22。
Abstract Making input decisions under climate uncertainty often involves two‐stage methods that use expensive and opaque transfer functions. This article describes an alternative, single‐stage approach to such decisions using forecasting methods. The example shown is for preseason fire suppression resource contracting decisions faced by the United States Forest Service. Two‐stage decision tools have been developed for these decisions, and we compare the expected gains to the agency, in terms of reduced personnel costs, of the single‐stage model over the two‐stage model, existing hiring decisions, and decisions that would have been made given perfect foresight about wildfire activity. Our analysis demonstrates the potential gains to versions of our single‐stage model over existing hiring decisions, equivalent to a benefit‐cost ratio of 22. The research also identified additional gains accruing from imposing biases on the single‐stage model, associated with asymmetric penalties from contracting decisions.