A long-term infrastructure forecast in the Gulf of Mexico is developed in a disaggregated decision- and resource-based environment. Models for the installation and removal rates of structures are performed across five water depth categories for the Western and Central Gulf of Mexico planning areas for structures grouped according to a major and nonmajor classification. Master hydrocarbon production schedules are constructed per water depth and planning area using a two-parameter decision model, where "bundled" resources are recoverable at a given time and at a specific rate. The infrastructure requirements to support the expected production is determined by extrapolating historical data. The analytic forecasting framework allows for subjective judgement, technological change, analogy, and historical trends to be employed in a user-defined manner. Special attention to the aggregation procedures employed and the general methodological framework are highlighted, including a candid discussion of the limitations of analysis and suggestions for further research. (C) 2003 Elsevier Science Ltd. All rights reserved.