Response to the Letter on Wormhoudt, J.; Wood, E.; Knighton, W.; Kolb, C.; Herndon, S.; Olaguer, E. 2015. Vehicle emissions of radical precursors and related species observed in the 2009 SHARP campaign; J. Air Waste Manage. Assoc. 65: 699–706

Dear Editor,

We are pleased to respond to B. Rappenglück and G. Lubertino’s comments on our paper (Wormhoudt et al., 2015). However, before responding to their specific comments we want to emphasize that we regard Wormhoudt et al. as an extension of Rappenglück et al. (2013), not a rejection of their work. They present fleet average emission factors for several important on-road vehicle exhaust species. We realized that during the same field campaign we had collected additional data in other Houston sectors that could be compared to their results. Our results also extend the range of exhaust species measured. We think our publication illustrates that fleet emissions are dependent on fleet composition and traffic conditions, which can and do differ in various sectors in a city as large as Houston. We believe the results in each publication should be used to refine and improve on-road emission models for Houston, by providing data that better describe the variability of exhaust emission factors.

We proceed to offer some clarifications, followed by a few comments.

Rappenglück and Lubertino state that their research approach was not intended to determine emission factors for individual vehicles. We did not interpret their results as individual vehicle emission data, nor did we present individual vehicle results in our paper.

It was indeed an oversight on our part to leave the terse notation involving “UH-LOPAP” in the Table 1 caption – it was done at a time in the evolution of the paper when the focus was on distinguishing that only one of the two HONO measurements considered in depth in the paper was used in the table, and was not modified when the table added species, such as HCHO, which were also measured by more than one group.

Regarding the concern that our Moody Tower data selection may have missed most of the rush hour impacts, it is indeed true that the latest time we used in computing a HONO delta was 6:38 am, although only 4 of the 20 cw-TILDAS HONO days had end times earlier than 6 am. Our goal was to estimate HONO correlations with other vehicle exhaust components, correlations which are readily observable at these early hours.

Rappenglück and Lubertino comment “It is mentioned that only for the UH-LOPAP results Wormhoudt et al. used the same method of data selection as in Rappenglück et al. 2013, i.e. a correlation of data streams over the entire morning rush hour period.” They do this to address an issue of data screening, in particular the question of HONO photolysis, but before we respond to this we must raise a delicate philosophical and semantic issue. We are well aware that data screening is critical to obtaining the best results, and therefore the term “method” is often used to include the screening criteria. While Rappenglück et al., through hard work and to their credit, were able to collect a large data set, the data sets considered in Wormhoudt et al. are sparse. There were, therefore, limits to our application of screening criteria, and this led us to refer to “the method” when we only meant “a correlation of data streams over the entire morning rush hour time period”, and were not claiming that our screening criteria were the same as theirs.

Moving on to the question of HONO photolysis, we estimated that the HONO photolysis lifetime at 6 am was about 3 hours, decreasing to 20 minutes by 8 am in full sun. These lifetimes are longer than transit times from highway sources, so HONO photolysis should not be a critical issue.

Rappenglück and Lubertino express concern regarding wind direction screening. We examined plots of the data in our Figures 3 and 4 plotted against wind direction. We saw no useful correlation, consistent with the map in Figure 1 showing major highways within 5 km of Moody Tower in all directions. Therefore, we did not use the same screening criteria for HONO as Rappenglück et al. (2013). For HCHO, we used the stated modifications of the screening criteria of Parrish et al. (2012). The 7 days that passed that screen are listed in the paper. They are different than the days used for HONO, and several had no cw-TILDAS HONO data, so studying the same days for both HONO and HCHO was not an option.

We close with a few comments. Regarding correlating Moody Tower HONO with wind speed, Rappenglück and Lubertino remark, “At certain wind speed thresholds these ratios would approach zero indicating no HONO emission at all, which is unreasonable.” We agree, that would be unreasonable, which is why we don’t agree that the ratios would approach zero. As useful as the linear fit was to guide the eye, at some point it is surely incorrect.

Rappenglück and Lubertino characterize the emission factors obtained by Dallman et al. in 2010 in the Caldecott Tunnel (San Francisco Bay Area) and by Rappenglück et al. (2013) as perhaps “the most representative average values of today’s traffic fleet.” Houston is a heavily industrialized city, and we maintain that the traffic mix at Washburn Tunnel and industrial sector roadways may be just as representative as the single highway site reported in Rappenglueck et al. (2013). In addition, Rappenglück and Lubertino calculate the fleet-average emission factor incorrectly. The diesel and gas fuel-based emission factors should not be weighted by number of vehicles but rather by amount of fuel consumed. Because diesel vehicles (mainly trucks in this country) consume some 5 times more fuel per unit of distance (~5 mpg vs. ~25 mpg), this greatly increases the relative importance of the diesel term compared to the 5% weighting used by Rappenglück and Lubertino. That is, “5% diesel and 95% gasoline” means 5% of the vehicles on the road and 5% of the miles driven are from diesels, but it does not mean that only 5% of the fuel being burned is diesel – it means that 21% of the fuel being burned is diesel. Applying this approximate correction yields

[(0.21*8.0)+(0.79*14.3)] = 13 g/kg(fuel) for CO, even closer to the Rappenglück et al. (2013) value of 10.4, and

[(0.21*28.0)+(0.79*1.90)] = 7.4 g/kg(fuel) for NO_x, which is a factor of 2.6 times higher than the 2.8 g/kg presented in Rappenglück et al. (2013).

In conclusion, given the potential importance of the HONO/NOx emission ratio, the range of results observed by several studies, including Rappenglück et al. (2013), Wormhoudt et al. (2015), and Xu et al. (2015), and the fact that none of these studies was able to derive separate HONO/NOx emission ratios for gasoline and diesel vehicles, it remains an important research endeavor to accurately measure these separate emission ratios. The current value of 0.8% used in MOVES may be inaccurate, but it would be premature to assume that significantly higher values are better.

Sincerely,