Reply to comment by D. A. Hansen et al. on “The 2003 North American electrical blackout: An accidental experiment in atmospheric chemistry”

[1] We thank Hansen et al. [2005] for their interest in our letter, ‘‘The 2003 North American electrical blackout: An accidental experiment in atmospheric chemistry’’ and appreciate the opportunity to expand upon the original, necessarily brief publication. We begin by making corrections to some apparent factual misrepresentations in their comment and then go on to provide a point-by-point response. [2] . Contrary to Hansen et al’s [2005] contention we do not base our inferences and conclusions on a single comparison of measurements performed on August 15, 2003 and August 4, 2002. We also compare measurements inside to measurements outside the blackout area on the same day, August 15, 2003. [3] . Reductions in power plant emissions that we report were not ‘‘estimated’’ as stated in the comment but are based on measurements made by power plants and reported to the USEPA. This was pointed out and referenced in the paper. [4] . We do not conclude, as suggested in the comment, that the observed improvements in air quality went on to benefit much of the ‘‘United States’’ but rather ‘‘much of the eastern United States’’. These are completely different statements. [5] Hansen et al. [2005] state the most important limitation of our study as being ‘‘. . . failure to consider the variability associated with concentrations of atmospheric species’’. They further state that similarity in synoptic weather patterns does not mean similarity in concentrations. However, the authors do not spell out what concentration variability considerations we overlooked nor do they attempt to qualify the statements. Complexities of atmospheric processes notwithstanding, it is not unreasonable to assume that when factors that drive transport and, subsequent physical and chemical transformation of pollutants in the air masses are similar, the concentrations of both primary and secondary atmospheric pollutants should be similar, unless of course, as obtained on August 15, 2003, a major primary pollutant source is perturbed. [6] Hansen et al. [2005] compare observed ozone maps from EPA’s AirNow archives for August 4, 2002 (our control day) and August 14, 2003; a day they contend was synoptically similar to our experimental day (August 15, 2003). They go on to argue that since O3 levels over Maryland, Pennsylvania, and New Jersey on August 4, 2002 were higher than on August 14, 2003 our observed changes in air quality following the blackout could have been a result of phenomena other than reduction in power plant emissions. This argument does not hold because in choosing the control day (August 4, 2002) we did not only consider synoptic-scale air motion but all factors that drive atmospheric chemical and physical processes, including temperature, insolation, and humidity. The average regional surface high temperature for August 14, 2003 ( 29 C) was about 4 C lower than that for August 4, 2002 ( 33 C) or August 15 2003 ( 33 C). Ryan et al. [1999] show that in this temperature range, a 4 C difference can account for as much as 35 ppb in O3, probably enough to account for the unspecified disparity in O3 abundance that Hansen et al. report. Ozone concentrations on August 14, 2003 cannot, therefore, be compared to August 4, 2002 or August 15, 2003. Local meteorology cannot be responsible for the differences in air quality observed in our study because surface O3 maps from EPA’s AirNow archives also show the same differences at the regional level. [7] Regarding the regional representativeness of measurements conducted over a single location, we would like to point out that unlike surface measurements, which Hansen et al. [2005] probably had in mind, aircraft vertical profiles (surface – 3 km) are more regional measurements that capture regional signatures. Moreover, our overall deductions are not based only on the profiles conducted over central PA but on transects and profiles conducted over PA, northern Virginia, and Maryland as well. The PM2.5 comparisons, which Hansen et al. go on to make in support of the representativeness argument, are wrong for the following reasons: [8] 1) As alluded to earlier, the days they compare (August 14 and August 15, 2003) are not synoptically similar. [9] 2) The PM2.5 data presented are 24-hour averages, therefore August 14, 2003 data were also affected (lowered) by the blackout, which started at 4:00 ET on the same day. Thus a comparison of 24-hour average PM2.5 concentrations GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L10813, doi:10.1029/2005GL022385, 2005