Meteorological Uncertainty Effects in Atmospheric Transport and Dispersion Modeling: A Demonstration

Atmospheric transport and dispersion models require estimates of the wind field as input. Uncertainty in the input wind field contributes directly to uncertainty in the dispersion predictions and therefore to uncertainty in the emergency-response decisions based on them. We present the results of our detailed analysis of one atmospheric transport and dispersion model, SCIPUFF, identifying the modeling components sensitive to meteorological uncertainty and discussing the theoretical justifications for a proposed treatment. Central to our presentation is a conceptual framework that defines and discusses “inner” and “outer” sources of variability in meteorological fields. We show that “inner” variability is already parameterized in SCIPUFF. We suggest that “outer” variability may be parameterized adequately by redefining the model length scale that governs plume spread. We present a possible new definition for this length scale and a supporting data analysis using an ensemble of mesoscale model output. We then apply the new length scale field in SCIPUFF and assess its impact on the uncertainty.