Evaluation of three probing techniques in a three-dimensional air quality model

[1] Probing techniques are useful to provide diagnostic evaluations of air quality models and to indicate the responses of model predictions to changes in emissions. Three probing tools are available in a three-dimensional air quality model, CAMx: the Decoupled Direct Method (DDM), the Ozone Source Attribution Technology (OSAT), and Process Analysis (PA). These tools are evaluated in terms of consistency, complementarity, accuracy and robustness, and computational requirements using the 7–15 July 1995 O3 episode over the eastern United States. The NOx- versus VOC-sensitivity of O3 chemistry and relative role of chemistry and transport predicted by the three tools are qualitatively consistent, except that OSAT predicts an NOx-limited O3 chemistry at a few locations where both DDM and PA predict a VOC-sensitive O3 chemistry. DDM and OSAT agree well on the top 10 contributors to O3 formation, but they predict different rankings, with greater importance given to biogenic VOC emissions by DDM and greater importance given to surface anthropogenic NOx emissions by OSAT. The major difference in the DDM and OSAT predictions on the relative impact of sources is that DDM predicts both positive and negative sensitivities whereas OSAT always predicts positive contributions. Compared to the single-perturbation method (brute force method), DDM predicts accurate model responses under the 25% VOC or NOx emission reduction scenarios but inaccurate results under the 75% NOx emission reduction scenario. OSAT predicts accurate model responses under the 25% VOC emission reduction scenario, but inaccurate responses under the 25% and 75% NOx emission reduction scenarios. While these tools provide valuable and complementary information regarding O3 formation, each of them has limitations in terms of its design and application for the design of emission control strategies. DDM is suitable for such an application for small-to-moderate emission reductions of <40%. OSAT leads to incorrect results for VOC-limited areas because it does not account for the titration/inhibition effect of NOx and/or VOC. PA provides an in-depth understanding of the processes involved in controlling O3 formation locally. It can be used to understand some of the differences between the results of OSAT and DDM.