Of all NOx sources, improving our understanding of lightning is most critical since it has one of the largest uncertainties and represents what appears to be the dominant source of NOx in the free troposphere. Lightning NO is generated by recombination reactions that occur as this 20,000 K or hotter, 10-MPa pressurized plasma super-sonically expands and cools. Quantifying NOx production from such events is still quite problematic.

Much of the current debate stems directly from estimating the amount of NOx produced by a "typical" lightning flash. This process has involved combining estimates of the average energy deposited per lightning flash with evaluations of the NOx produced per joule of energy released and the average global frequency of lightning. Using this approach, the global estimate of 2TgN/yr by Kumar et al. (1995) lies at the low end of a clustering of similarly derived estimates based on very low yields of NO^ per lightning flash (~3.6x 1025 NOx molecules/flash; e.g., Lawrence et al. (1995) and references therein). At the high end, Liaw et al. (1990) have argued for a source strength as large as 200TgN/yr based on a correspondingly larger value for NOx yield per lightning flash, but this estimate appears unreasonably high based on nitrate deposition records. Many of these treatments have relied on the application of scaling or normalization factors to other investigators' results that may not be valid, and frequently the various forms of lightning have been treated as if they were one type only.

Combining more refined values for the production of NOx per joule of energy (Goldenbaum and Dickerson, 1993) with updated lightning flash energy values results in a per-flash NOx yield for negative CG (cloud to ground) lightning of 1 to 2 x 1026 NO/flash. By contrast, for positive CG and IC (intracloud) lightning, the yield is approximately 5 x 1026 and 0.5 x 1026 NO/flash. Furthermore, current evaluations of the global distribution of lightning (Goodman et al, 1988; Christian et al, 1992) in combination with estimated spatial distributions for different types of lightning (Orville, 1994) results in a nominal 100 global flashes/s being proportioned 75% to IC lightning and 25% to CG lightning. Seventy percent of CG lightning is associated with the tropics/subtropics, with 5% positive strokes, and 30% is associated with higher latitudes having 30% positive strokes. Combining these estimates with the midrange value for NOx production per flash for each lightning type results in a conservative estimate for total NOx production of 2.5 TgN/yr by IC lightning, 3 TgN/yr by negative CG lightning, and 1 TgN/yr by positive CG lightning.

These global lightning estimates are found to be in generally good agreement with other independent NOx assessments not dependent on the mechanistic details of lightning. For example, Albritton et al. (1984) constrained the global lightning source strength by examining nitrate deposition records from remote global areas that were expected to be free of impacts from anthropogenic sources. They estimated a lightning source of approximately 8 TgN/yr (range 2 to 20). More recently, Levy et al. (1996) have used a global chemical transport model in conjunction with remote, upper tropospheric NO measurements to constrain all lightning sources. Their results, which critically depend on their choice of parameterizations for deep convection, indicate a range of values for NOx production from lightning of 2 to 6TgN/yr.

A final issue of importance concerning lightning emissions relates to the altitude distribution of emissions. This is influenced not only by the initial production from lightning but also by subsequent vertical mixing in convective storms. Pickering et al. (1998) have recently produced estimates for the vertical distribution of lightning emissions. Their results show that most lightning NOx is delivered to the upper troposphere; however, the more vigorous mixing of midlatitude continental storms leads to more downward transport of lightning NOt than for maritime storms or tropical continental storms. They also showed that the peak NOt in continental storms occurs at higher altitudes than for maritime storms.

0 0

Post a comment