Uncertainty Analysis

CEMPD has researched various methods to quantify uncertainty related to air quality modeling. Our research addresses uncertainty due to emissions, meteorology, and ambient air concentrations of various pollutants. Uncertainty in ambient air concentrations is also propagated to exposure and risk assessment models. Recently CEMPD has concluded an important research contract to address toxics uncertainty in the Houston area. Benzene, Formaldehyde, and Chromium were used for the Houston case study. The goal is to provide a guidance to assist in implementing iterative improvements in subsequent National-Scale Air Toxics Assessment for 1996 (NATA) applications to develop a methodology for quantifying the uncertainty in the NATA results. The objectives of the research program are (i) to develop a methodology for quantifying the uncertainty in the NATA results; (ii) to demonstrate the use of the methodology in a case study for a particular urban area, estimating uncertainties in the NATA results; (iii) to evaluate the extent to which these uncertainties can be extrapolated to a national domain; and (iv) to provide recommendations on how uncertainties in study results can be represented and interpreted.

The following modeling protocol was designed:

Identify the highest sources in each of the four main categories (mobile on-road, mobile nonroad, point, and nonpoint) that contribute to about 90% of the total emissions of benzene, 99% of the total emissions of formaldehyde, and 99% of the total emissions of chromium in the Houston area.
Quantify the emissions rate uncertainties for each of the main categories' emitting subgroups for each of the three subject pollutants. The outcome is fitted parametric distributions for normalized uncertainty factors for total emissions of benzene, formaldehyde, and chromium for each of their source categories.
Use the derived normalized emissions uncertainty factors to formulate a series of Monte Carlo (MC) simulations that are then used to generate 100 emissions datasets for input into ISCST3 in order to quantify the uncertainties in the output simulated annual concentrations of benzene, formaldehyde, and chromium.
Quantify uncertainties in key meteorological parameters in ISCST3 and perform MC simulations to alter the meteorological conditions used by ISCST3 and consequently the output simulated concentrations of the subject pollutants.
Analyze the simulated annual concentrations of benzene, formaldehyde, and chromium and quantify the uncertainties in these pollutants in the Houston area (based on concentrations at a 100 receptors).
Extend the uncertainty analysis to exposure and risk

The study also included a series of sensitivity analyses:

Perform emissions sensitivity studies to assess the sensitivity of the ISCST3 model's simulation of annual benzene concentrations to temporal and spatial variations in benzene emissions.
Perform a sensitivity study for the sake of enhancing the computational efficiency of the modeling exercise, by examining how reducing the number of receptors affects the accuracy of modeled benzene concentrations.

For more information, please contact Adel Hanna.

Relevant Projects

No ongoing projects at this time