Best approach to impact assessment is to use empirically based or simulation models to forecast impacts. Environmental Sciences Division Publication No. 1538

This paper advocates the utility of mathematical models, as contrasted with statistical procedures and processional judgment, for assessing environmental impacts. While it would be desirable to use statistical tests to detect and estimate impacts, this is generally difficult or impossible to do, even with existing sources of impact. Empirical modeling, supported by statistical analyses when possible, is proffered as the logical alternative. Next, for purposes of forecasting impacts, the use of models as opposed to professional judgment or experience is considered. The conclusion is reached that, while models cannot answer all of the relevant questions, they can be used effectively and can address problems that are beyond the reach of statistical methods