Environmental intervention is often seen as being high risk and high return. Traditional scientific hypothesis testing provides limited guidance to policy makers unless there is a high level of certainty in the supporting scientific evidence. Traditional cost-benefit analysis under uncertainty has shortcomings when considering high-risk investment, largely due to the choice of how to discount uncertainty outcomes. A corollary is that traditional cost-benefit analysis does not place a value on increased certainty, an important outcome of successful scientific research. A fiducial costbenefit methodology is presented in this paper, which integrates hypothesis testing and traditional cost-benefit analysis. The fiducial approach is one way of objectively placing a value on changes in the level of uncertainty that does not depend on an assumption about a decision maker's attitudes towards variability in returns. This has two important implications. First, there is a level of uncertainty at which we would reject an investment with a positive expected net rate of return on the basis that the uncertainty associated with the outcome is too great. Second, it is possible to value a program of research that reduces the uncertainty about a critical decision parameter. An example based on data from a weather modification experiment conducted in South Australia is presented. The approach is the generalised using more traditional statistical methodology.