Meteorological Uncertainty and the Expert Witness

This paper addresses some problems raised due to the fact that the expert witness is often asked to provide a yes-or-no answer in the face of uncertainties in meteorology and air pollution and other environmental variables. In many of these cases, the use of statistical analysis (e.g., 95 % confidence bounds) is appropriate but the results must be expressed in a way that may be understood by a lay-person. Examples of several scenarios are discussed, including: 1) Interpretation of model-predicted pollutant concentration contours; 2) Analysis of whether two meteorological observations are similar or different; 3) Assessing confidence in whether there was freezing rain (or lightning or other meteorological phenomena) at a specific time and place; 4) Estimating and interpreting error bounds (95 % confidence limits) on model predictions or other conclusions regarding meteorological or air pollution variables.