Analyzes of investments that are irreversible and have uncertain benefits should consider the option of delaying a decision. For instance, the benefits of many water resource projects could change if global warming occurs. The magnitude of that warming is uncertain, and delaying projects until more information is available might be optimal. We examine whether this is true for construction of an outflow control structure for Lake Erie. Using Bayesian Monte Carlo (BMC)-based decision analysis, we find that considering climate uncertainty does make a difference. Climate change beliefs, in the form of prior distributions over transient climate scenarios, can affect the optimal strategy: in particular, climate change makes delaying construction more attractive. The option value of deferring the decision to build is as high as $20 million. Ignoring the possibility of climate warming can inflict an expected penalty as large as 20% of the cost of the control structure. We also compare climate risks to uncertainties in stage-damage curves and find that they are approximately of equal importance. ¿ 1999 American Geophysical Union