[1] A subset of the hydrocarbon samples collected on the DOE G-1 aircraft during the Texas Air Quality Study (TexAQS) has been found to be contaminated. As quantified below, changes to the 5 city comparison presented in the above paper [Kleinman et al., Geophys. Res. Lett., 29(10), 1467, doi:10.1029/2001GL014569, 2002] when these samples are removed from our analysis are all small. In particular, the main conclusions of the above paper; that 10% of the hydrocarbon samples from Houston have several-fold greater calculated O3 production rates than other cities, and that these high O3 production rates are due to very high hydrocarbon reactivity, remain valid. [2] A comparison of hydrocarbon data sets obtained from different platforms during the TexAQS 2000 field campaign shows that some canister samples collected on the G-1 aircraft have anomalous concentration patterns. The dominant signature of the anomalous samples is an unrealistic concentration of toluene and i-pentane as judged by a comparison with samples taken in Houston from the NCAR Electra and Baylor Twin Otter. In 2 cases concentrations exceeded 200 ppbv. We have removed 23 samples out of 234 collected on the G-1 where the fraction of anthropogenic carbon due to toluene exceeds 10% and/or the fraction due to i-pentane exceeds 20%. This criteria removes all of the anomalous high concentration samples, yielding a data set with maximum toluene and i-pentane concentrations of 3.1 and 8.4 ppbv, respectively; within the range of values observed elsewhere. Of the 23 samples removed, 11 have toluene and/or i-pentane concentrations below 3.1 and 8.4 ppbv. Similar samples, with “realistic” values of toluene and i-pentane but having C atom fractions either exceeding 10% for toluene or 20% for i-pentane are also observed from the Twin Otter and Electra; approximately 3 times less often for toluene and at about the same frequency for i-pentane. [3] We have ruled out analysis problems as the cause of the anomalous samples, as toluene and i-pentane were quantified on 2 independent gas chromatographs with agreement within expected limits. The problem is not endemic to our sampling system as judged by the many samples that have low toluene and/or i-pentane concentrations (i.e. 10–100 pptv) on the same flights that anomalously high concentrations occur. We believe the problem to be discrete instances of contamination, but the source has not been identified. [4] The distinctiveness of Houston is based on a comparison between those hydrocarbon samples in each city which have the highest 10% O3 production rates. For Houston, this 10% subset consists of the 20 samples identified in Figure 1 of Kleinman et al. [2002]. Four of these samples have been found to be contaminated. Three contaminated samples were used to quantify VOC reactivity in Figure 2, the 4th sample being the outlier discussed in the figure caption. Sixteen contaminated samples were used to determine P(O3) for the data subset that excluded the high P(O3) compounds. We have generated statistics for the data set without the contaminated samples to compare with results presented in the figures and text of Kleinman et al. [2002]. The average P(O3) for samples in the lower 90% of the Houston P(O3) distribution is reduced 2% from 11.8 to 11.6 ppb h−1. Changes to the Houston high P(O3) subset are indicated in Table 1. The corrected values in Table 1 were generated by removing the contaminated samples and augmenting the list of high P(O3) samples with 2 additional cases so that the high P(O3) samples still represent 10% of the total. The new values of P(O3), total reactivity, and anthropogenic reactivity are, respectively, 9%, 11%, and 13% lower than those originally reported. More than half of this decrease is due to the additional 2 samples added to the high P(O3) list. [5] Thus, except for the mention that there are some samples with extremely high concentrations of alkanes and aromatics, all of the conclusions of the original paper remain valid.
Read full abstract
Jan 1, 2008
+ 3
Open Access