Observations from the global observing system are a key input to operational weather prediction. Thus, meteorologists often suggest improving the observing system as a means to improve weather forecasts. However, because resources for new observing platforms are limited, choices must be made among proposed observing system improvements. Consequently, both meteorologists and policy makers are interested in designing an integrated global observing system that cost-effectively meets both atmospheric science and societal needs. This paper combines atmospheric science and public policy perspectives to present a systematic framework for designing such an observing system. Information and research required to implement the framework are also discussed. Although the examples used focus primarily on weather prediction, the concepts are sufficiently general that they can applied to other earth observing system applications (e.g., weather, climate, and other environmental monitoring, prediction, and research), or for designing a multi-purpose integrated observing system.