Economically efficient natural resource policy requires the comprehensive management of environmental improvements over a system of sites. Use of multisite demand and benefit models can reduce the number and cost of environmental program contradictions. In this paper we present a multisite demand model that incorporates several quality dimensions for a system of New Mexico fishing sites. Use of the model permits managers to improve total welfare at lower cost by simultaneously distributing opportunity across all elements in resource users' choice sets. Model results show that errors that result from independently adding values of interdependent program elements approach 50%. Copyright 1997, Oxford University Press.

We describe a two‐stage procedure to increase recreational angler use through optimal allocation of hatchery‐produced fish among sites. In stage 1, data on historical fish stocking, angler‐days, and attributes of individual fishing sites are used to formulate a regression equation that shows the relation between angler‐days and the number of fish stocked (stocking elasticity). In stage 2, this stocking elasticity is used in conjunction with data on the hatchery supply of rainbow trout Oncorhynchus mykiss to formulate a reallocation of fish among fishing sites. The algorithm allocates the limited production of hatchery rainbow trout among receiving sites to maximize total angler‐days by equalizing the potential gain in angler‐days from stocking one additional fish at each site. Application of the procedure to selected fishing sites in New Mexico suggests that total angler‐days could be increased by 16% with no increase in budget or other resources. An optimized fish stocking program would produce several benefits: (1) increased angler‐days by the fishing public, (2) a fish stocking schedule that is easily updated any time hatchery supply or on‐site facilities change, and (3) an objective and defensible stocking program that meets the demands of rigorous public scrutiny.