The physical habitat simulation model (PHABSIM) produces curves showing the relation between flow and habitat that are often used as evidence for decisions on the management of water resources. The curves show values of “weighted usable area” (WUA) over discharge. Minimum confidence intervals for the curves can be developed with the bootstrap method if transects are randomly located within sampling units. This is illustrated with bootstrap simulations that develop minimum confidence intervals for several model rivers, defined in terms of sampling procedures and values of WUA for juvenile chinook salmonOncorhynchus tshawytschaat 15 transects on the Feather River, California. The simulations show that the chance location of transects in PHABSIM studies can introduce such great uncertainty into the resulting curves of WUA over discharge that the curves can be meaningless. Therefore, it is crucial that confidence intervals be developed for curves of WUA over discharge before PHABSIM results are presented to decision makers.