Shale gas development may involve trade-offs between energy development and benefits provided by natural ecosystems. However, current best management practices (BMPs) focus on mitigating localized ecological degradation. We review evidence for cumulative effects of natural gas development on brook trout (Salvelinus fontinalis) and conclude that BMPs should account for potential watershed-scale effects in addition to localized influences. The challenge is to develop BMPs in the face of uncertainty in the predicted response of brook trout to landscape-scale disturbance caused by gas extraction. We propose a decision-analysis approach to formulating BMPs in the specific case of relatively undisturbed watersheds where there is consensus to maintain brook trout populations during gas development. The decision analysis was informed by existing empirical models that describe brook trout occupancy responses to landscape disturbance and set bounds on the uncertainty in the predicted responses to shale gas development. The decision analysis showed that a high efficiency of gas development (e.g., 1 well pad per square mile and 7 acres per pad) was critical to achieving a win-win solution characterized by maintaining brook trout and maximizing extraction of available gas. This finding was invariant to uncertainty in predicted response of brook trout to watershed-level disturbance. However, as the efficiency of gas development decreased, the optimal BMP depended on the predicted response, and there was considerable potential value in discriminating among predictive models through adaptive management or research. The proposed decision-analysis framework provides an opportunity to anticipate the cumulative effects of shale gas development, account for uncertainty, and inform management decisions at the appropriate spatial scales.