NOx-VOC sensitivity represents a major uncertainty for oxidant photochemistry in urban areas. In previous work we have found that predicted NOx-VOC sensitivity in models can be correlated with predicted values for certain "indicator ratios": O3/NOy, H2O2/HNO3, and others. This finding suggests that measured values for these species can be used to identify biases in model NOx-VOC predictions.

Here, the method is applied to the Texas Air Quality Study. The Community Air Quality Model (CMAQ) has been used to simulate ozone formation in Houston, using different scenarios with different NOx-VOC sensitivity. Model correlations between O3, peroxides and reactive nitrogen (NOy) will be shown for NOx-sensitive and VOC-sensitive locations in each scenario. Model results show a strong correlation between ozone and NOx reaction products (NOz) for both NOx-sensitive and VOC-sensitive locations, but the NOx-sensitive locations have higher O3/NOz and a higher slope. This difference in behavior between NOx-sensitive and VOC-sensitive locations is driven primarily by HNO3. Correlations between O3 and organic nitrates do not appear different for locations with different NOx-VOC sensitivity. Comparisons between model correlations and measured values in Houston will be used to identify whether model-measurement discrepancies provide evidence for errors in model NOx-VOC predictions.