the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Historical (1960–2014) lightning and LNOx trends and their controlling factors in a chemistry–climate model
Kengo Sudo
Abstract. Lightning can cause natural disasters that engender human and animal injuries or fatalities, infrastructure destruction, and wildfire ignition. Lightning-produced NOx (LNOx), a major NOx (NOx = NO + NO2) source, plays a vital role in atmospheric chemistry and global climate. The Earth has experienced marked global warming and changes in aerosol and aerosol precursor emissions (AeroPEs) since the 1960s. Investigating long-term historical (1960–2014) lightning and LNOx trends can provide important indicators for all lightning-related phenomena and for LNOx effects on atmospheric chemistry and global climate. Understanding how global warming and changes in AeroPEs influence historical lightning–LNOx trends is also helpful because it can provide a scientific basis for assessing future lightning–LNOx trends. Moreover, global lightning activities’ responses to large volcanic eruptions (such as the 1991 Pinatubo eruption) are not well elucidated, and are worth exploring. This study used the widely used cloud top height lightning scheme (CTH scheme) and the newly developed ice-based ECMWF-McCAUL lightning scheme to investigate historical (1960–2014) lightning–LNOx trends and variations and their controlling factors (global warming, increases in AeroPEs, and Pinatubo eruption) in the framework of the CHASER (MIROC) chemistry–climate model. Results of sensitive experiments indicate that both lightning schemes simulated almost flat global mean lightning flash rate trends during 1960–2014 in CHASER. Moreover, both lightning schemes suggest that past global warming enhances historical trends of global mean lightning density and global LNOx emissions in a positive direction (around 0.03 % yr−1 or 3 % K−1). However, past increases in AeroPEs exert an opposite effect to the lightning–LNOx trends (−0.07 % yr−1 – −0.04 % yr−1 for lightning and −0.08 % yr–1 – −0.03 % yr–1 for LNOx). Additionally, effects of past global warming and increases in AeroPEs on lightning trends were found to be heterogeneous across different regions when analyzing lightning trends on the global map. Lastly, this study is the first to suggest that global lightning activities were suppressed markedly during the first year after the Pinatubo eruption shown in both lightning schemes (global lightning activities decreased by as much as 17.02 % simulated by the ECMWF-McCAUL scheme). Based on the simulated suppressed lightning activities after the Pinatubo eruption, our study also indicates that global LNOx emissions decreased after the Pinatubo eruption (2.41 % – 8.72 % for the annual percentage reduction), which lasted 2–3 years. Model intercomparisons of lightning flash rate trends and variations between our study (CHASER) and other Coupled Model Intercomparison Project Phase 6 (CMIP6) models indicate significant uncertainties in historical (1960–2014) global lightning trend simulations. Such uncertainties must be investigated further.
- Preprint
(20800 KB) -
Supplement
(17098 KB) - BibTeX
- EndNote
Yanfeng He and Kengo Sudo
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-301', Anonymous Referee #1, 28 Apr 2023
This manuscript presents the results of model simulations of lightning flash rate and resulting lightning NOx emissions over the period 1960 to 2014 using the CHASER chemistry and climate model. The authors attempt to isolate the effects of global warming, changes in aerosol abundance, the Pinatubo volcanic eruption on flash rates. They are very successful in doing that with regard to warming and Pinatubo. However, the CHASER model only includes the radiative effect of aerosols, and the microphysical effects of aerosols in convective clouds are not included. On one hand, I could see that this could be a reason to reject the manuscript. However, on the other hand, it is interesting to see the effect on flash rates of the aerosol changes to atmospheric radiation. If the paper is allowed to proceed to publication, the authors must put more emphasis throughout the paper on the fact that the microphysical effects of aerosol on flash rates are neglected in this analysis. Yes, a caution about the aerosol results is written into Section 4 (Discussion and Conclusions), but it also needs to be prominent in the abstract and in the Results section. Inclusion of the microphysical effect could cause changes in flash rate of opposite sign compared to the results reported in this paper.
Other Comments:
line 54: add Romps et al. (2014) reference; change "minor" to "other"
lines 90-92: over what years?
lines 151 - 152: I understand that this formulation takes into account convective clouds of various heights within a model grid cell. However, Hi < freezing level should not produce any lightning. Is this taken into account?
line 166: If QRa includes cloud ice, it should not be described as "precipitating ice"
line 191: emissions. For example, Allen et al. (2019) and Bucsela et al. (2019) found that lower LNOx production per flash was associated with larger flash rates.
lines 196 - 200: these should all be written in the past tense since OTD and TRMM/LIS are no longer operational
line 203: is LRTS the mean for each day over the 20-year period?
line 235: This equation needs more explanation. I'm not familiar with this formulation.
Figure 1: Both schemes put too many flashes over oceans. The colors cover the factor of 5 range from 10 to >50 are too similar. It is difficult to differentiate them. Need to use a different color scheme.
Figure 2: Earlier in the paper it says that LIS data only extends to +/- 38 degrees latitude. Why is this figure using +/- 41.25 degrees?
Figure 5f: Why are contributions to lightning coming from the shallow clouds (below the 600 hPa level)? Ice is needed for lightning.
line 369-370: Why would that necessarily reduce QRa? Clouds could be taller and have similar QRa as before global warming.
lines 374 - 378: Here is one of the locations where the lack of consideration of the microphysical effect of aerosols on hydrometeors and flash rates needs to be mentioned. This effect could increase hydrometeors, QRa, and flash rate.
line 386: need to compare 1960 and 2014 in Figure 5f
line 390: again, indicate that the decrease in cumulus reaching 200 hPa with increases in AeroPEs is only considering radiative effects.
line 417: Here again, the aerosol results on flashes from the model may be opposite of what has been observed. There have been numerous papers looking at observations over China that show increasing flashes as aerosol increased and decreasing flashes as aerosol has declined.
Figure 10: Is there greater density of negative relative differences? If so, that should be mentioned.
Citation: https://doi.org/10.5194/egusphere-2023-301-RC1 -
RC2: 'Comment on egusphere-2023-301', Anonymous Referee #2, 02 May 2023
I have read this paper on historical modelled trends in lightning activity from 1960-2014. Unfortunately, I have many problems with the paper that are described below. The paper looks at three factors that could have impacted lightning activity (and LNOx) over this period. The first is the increasing temperatures, the second is the increasing aerosol loading of the atmosphere, and the third is the Pinatubo volcanic eruption in 1991. The authors use 2 different parameterizations of lightning in their model.
The model results show no real change in the lightning activity for changes in temperature and aerosols (different to other simulations published by others and CMIP6), irrelevant of the parameterization used. The trends in lightning over this period are not significant, and given the small increases in temperature we may not expect any significant changes. But when looking at the model lightning response spatially (Figure 6) the anomalies are in the polar regions mainly, and not where we normally see lightning. Other studies predict increases in lightning in the tropics. More worrying is that the increase in lightning due to temperature increases (greenhouse gases) and aerosol increases show the same spatial response, while the aerosol forcing is very different in location to the greenhouse gas forcing. Both simulations show strong positive anomalies in the Arctic and Antarctica, and strong negative anomalies around Antarctica, regardless of whether the forcing is from aerosol changes or temperature changes (greenhouse changes). This raises a red flag since the distribution of the changes should be spatially different. And in both scenarios little change occurs in the tropics. Not realistic. How are the global trends calculated? From the global absolute lightning changes? Or the changes in each pixel (%) averaged over the globe. The results imply that the lightning in this model is NOT sensitive to changes in temperature and aerosols.
However, the aerosol effect is only the radiative effect. The authors state they do not include the microphysical effect of increasing CCN since they simply cannot in this model. So simulating the aerosol impact with no microphysics is misleading, and not usedful for other researchers. It is clear that having more aerosols will cool the surface and hence should reduce convection. But this is not a new result. Others have already shown this. So why publish these results if not realistic without including all the process linking aerosols and lightning? Furthermore, unlike what is expected, the aerosol simulations show an INCREASE in lightning over time. How do you explain this?
More worrying is the temporal change in the simulated lightning from the two parameterizations, especially F1 (Figure 9 e and f). The F1 parameterization shows two maxima in global lightning every year, while the F2 parameterization shows only one. From my knowledge the LIS/OTD data show one maximum per year in the northern hemisphere summer. If so, how can you use the F1 parameterization in this study when it cannot reproduce the seasonal variability of global lightning? Without explaining the observed temporal variability of global lightning we cannot trust the results of this parameterization. And if the F1 parameterization agrees with LIS/OTD, then F2 does not.
The Pinatubo experiment is the most interesting part of this paper, and I would focus ONLY on this in the revised paper. While this effect is clearly seen in the CMIP6 simulations of lightning, I do not think there has been a specific paper on this topic. Hence, I would encourage the authors to publish a paper only on this, and remove the sections on temperature and aerosols. However, here too, if the model cannot duplicate the climatology of lightning and the annual cycle, then how can we believe the results related to the Pinatubo eruption.
Minor comments:
Line 7: "that result in human....."
Line 18: "Results of sensitivity...."
Line 36: Lightning is not a disaster, it is a hazard
Line 193: I think Pickering et al.,1998 was the first to publish this
Pickering, K.E., Y. Wang, W.K. Tao and C. Price, 1998: Vertical distribution of lightning NOx for use in regional and global chemical transport models, J. Geophys. Res., 103, 31203-31216
Line 203: Does the satellite data provide daily data or monthly data? Since the satellite is in LEO orbit, how does it get global data every day?
Line 313: Why did you stop at 2014? You now have the LIS-ISS data you could also add to the time series.
Line 319: (NCEI) (Figure 3c)
Line 348: Even if this is statistically significant, the r-value is so low that this implies that the temperature and aerosols explain basically nothing of the monthly variability of the lightning activity in the model.
Figure 3: If the aerosol experiment is only a radiative cooling effect, why do we see the largest increase in lightning here? You claim the radiative cooling of aerosols should stabalize the atmosphere. Please explain.
Figure 4: The temporal variability of the aerosol forcing and greenhouse forcing should be different, while the response of the lightning is basically the same!
Line 368: If the CAPE is larger, and the clouds are deeper, why is the volume of precipitating ice decreasing over time. This does not make physical sense. If the mixed phase region shifts to higher altitudes why would that impact the volume?
Figure 6: See comments above. Spatial anomalies similar between parameterizations, and similar between forcing of aerosols and greenhouse warming. Most of the changes are in the polar regions, which globally make up a tiny amount of global lightning. Maybe plots with absolute changes would be more informative. I do not see the harched lines in the top plots.
Figure 8: Why are the trends for LNOx largest for the aerosols experiment?
Line 434: Figs. 9a-f
Figure 9: why do the simulations not return to agree with each other a few years after the Pinatubo eruption when the aerosols have been removed from the stratosphere? The two maxima in global lightning in 9e is very worrysome, and a cause to reject the paper.
Given the reduction of lightning after the Pinatubo eruption, why not calculate the % change per degree C for this event to show the sensitivity of the model lightnign to global cooling from a volcano. This may also be of interest to the reader to compare with global warming studies, ENSO studies, solar constant studies, etc.
Figure 13: Why does the blue curve peak in 1991 if this is for the run of Volcano-off??
Line 565: Wouldn't a linear regression be better than simply using the first and last data point?
Line 587: You do not model the microphysical effects on lightning, so this is misleading. Your study does not simulate the real world effects of aerosols on lightning
Line 605: You claim it plausible to see a suppression of lightning due to aerosols, but your simulations show an increase in lightning!!
In conclusion, I would normally reject such a paper, but allow the authors to make major revisions addressing all my points before consideration of whether to reject or not.
Citation: https://doi.org/10.5194/egusphere-2023-301-RC2
Yanfeng He and Kengo Sudo
Yanfeng He and Kengo Sudo
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 216 | 59 | 10 | 285 | 23 | 2 | 3 |
- HTML: 216
- PDF: 59
- XML: 10
- Total: 285
- Supplement: 23
- BibTeX: 2
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1