Evaluation of the Concentration Level Uncertainty Ensemble System (CLUES) compared to ensemble numerical weather prediction (NWP) model input using field test data

The Concentration Level Uncertainty Ensemble System (CLUES) is a tool for estimating uncertainty in modeled atmospheric concentrations derived from transport and dispersion (T & D) models. CLUES comprises several models that attempt to characterize the uncertainty of predicted results from the mesoscale model Regional Atmospheric Modeling System (RAMS) and the dispersion models Short-range Layered Atmospheric Model (SLAM) and CALPUFF. CLUES calculates the uncertainty in these model results using a variant of the Monte Carlo method where winds from RAMS are perturbed with simulated errors. The perturbed winds are then used as input to the selected T & D model. The resulting output can be used to determine the possible errors due to input data uncertainties.

For this evaluation, SLAM dispersion model concentrations were calculated using input meteorological data from CLUES and from an NWP ensemble. The NWP ensemble consisted of a total of eighteen RAMS and Weather Research & Forecasting (WRF) model configurations where model physics options were varied. The modeled concentrations were compared to concentrations collected during the Southern California Tracer Study 2007(SOCAT07) experiment conducted in July 2007 in the Mojave Desert region. Various statistical and graphical comparisons were performed and an assessment of the CLUES and NWP methodologies was conducted. The preliminary results indicate some differences between CLUES and NWP ensembles but neither one did significantly better than the other when compared to measured concentrations. This presentation will present a description of CLUES, the methodologies used in this evaluation and the results of the comparisons.