May 22, 2020, ©. Leeham News: In our Corner series about flying during the COVID-19 pandemic, we look closer at the available research around passengers that fly with virus infections and if these spread to other passengers during a flight.
How much do we know and what are guesses?
What statistics are available around COVID-19 infections in airliners?
There were several studies available on virus infections (especially SARS but also seasonal flu) before IATA (the airline association) started searching the market for data around COVID-19 infections during airliner flights.
Most studies were around the spreading of SARS, a Coronavirus illness similar to COVID-19 (The COVID-19 disease is caused by the virus named SARS-CoV-2), with others analyzing the spreading of normal flu.
The studies, except for a couple, had the same result. Infection during flight passenger to passenger did not happen.
One SARS study had infections of co-passengers, as had an earlier flu survey. The infected passenger had in both cases a developed illness and coughed/sneezed repeatedly during the flight without wearing a mask.
From Figure 1, we can understand what happened. The strong jet of unprotected sneezing/coughing can overpower the orderly downdraught of the cabin air.
But these studies are exceptions (the SARS study was from 2003 with the flu study made earlier).
IATA has collected additional statistics around the present COVID-19 pandemic from its member airlines.
The results are informal, as only two scientific studies have been published regarding COVID-19 so far. IATA initially surveyed 18 larger airlines and then 70 of its airlines, representing half of the global passenger traffic.
The result from the two published studies and the 18+70 informal IATA surveys are the same. No passenger to passenger transmission of COVID-19 was found.
Three suspected passengers to cabin crew infections were reported and four pilot to pilot infections. In the latter case, the risk period included ground layovers for the pilots.
The passenger to cabin crew infections show the increased risk for a cabin crew when it leans into an unprotected passenger’s breathing when serving meals in a seat-aisle.
The published studies are one from Canada, where a long-haul flight with a confirmed COVID infected passenger resulted in no further infected passengers. Another report analyzes a flight on 31 March from the USA to Taipei with had 12 infected passengers on the flight. It generated no new cases from the 328 other passengers and crew members. All have subsequently tested negative.
Why is the risk of a passenger to passenger infection low?
We have discussed this briefly in the first article. We will dig deeper into this aspect in the next Corner.
Check out the theory of Prof Sunetra Gupta that Covid19 has basically already passed. She is a proper scientist and epidemiologist as well as unfortunately being a prolific novelist.
I don’t really know what to make of it, but it demonstrates just how little we know.
If she is right, then we are going to have to confront the horrific possibility of Trump being able to claim to have been right.
Gupta’s comment was around the spread in the UK. Let’s look wider.
This is John Hopkins CSSE data for the top 25 countries of confirmed COVID-19 cases:
It shows the US, Russia, and Brazil racing to the top followed by the UK, all countries managed by leaders that played down the pandemic and now would love the theory to be true.
With the US consistently at 25k new cases per day since mid-April and Russia and Brazil accelerating, who can think COVID-19 has passed?
Then compare the US to China, which has 4.4 times the population. The US increases with 25k cases per day, China 0. But this is of course “because the US has better testing”.
There is a lot of wishful thinking going around re. COVID-19. The naked facts say something else.
Amen to that! Many of the thinking US population are cringing at some of the statements being made by our leaders. The truth is that it got away from us because we saw this outcome as being extremely unlikely, and thus reacted too slowly. As did much of the world. There is some evidence that the same thing went on in China initially. But it would be far better to admit this and learn from it, than to deny that it happened, or point the finger of blame elsewhere, or make ridiculous claims.
“The truth is that it got away from us because we saw this outcome as being extremely unlikely.”
Why did we see it as “unlikely” and who is “we”? The lavishly funded peak UN body, the WHO, an organisation that is tasked to process and distribute information communicable disease got everything absolutely wrong (always do opposite of what WHI says would have worked). and It not only reputedly gave the wrong advice but actually argued against measures such as mask wearing and partial border closures despite it having been given information.
So can we blame politicians for initially taking this advice even if it was absurd?
This body was even supported as an exclusivity authority by much of one ideological polar of the main stream media to the point of censorship by social media corporations of those disagreeing. The WHO consistently maintained absurd and illogical positions (such as the claim that there was no person to person transmission for much of January while a pandemic was spreading and Wuhan had emptied of millions of people). I remember thinking how impossible this was and thinking that I have to disregard these people. WHO underplayed it through the remainder of January even claiming mask wearing was unnecessary. That’s not political that is a matter of record.
The Organisations voting is political not scientific. It must be reformed, bypassed or replaced.
Can billions (trillions) of dollar of aviation investment be safe with these people in charge?
One guy wants to use it to promote his vaccine business/charity. The Chinese governments wants it to recognised Chinese medicine so it can create a 50 billion nutraceutical business.
The people that got this right was Taiwan who tested, wore masks and shut down their border with China in December. Taiwan is excluded from the WHO by sycophantic governments bowing ignominiously to Chinese Government economic pressure. (Am I lying?) Singapore and Korea also took their own advice.
Lots of folks now saying we shouldn’t ‘blame’ and move forward in an impartial manner as if this is only a matter of sticking to the ‘science’.
I understand management and quality control. Blame causes folks and we go from problem solving/cooperation to defensive/aggressive and thereby to hide problems and also solutions but if there already was malice, selfishness and pernicious strategizing then blame is required. This is beyond mistake. We had people silenced with threats so this is not a mistake.
Moreover Ive noticed that same media who now wish to use this as an opportunity to condemn say Donald Trump or Borris Johnson for not locking down sooner are the same as those who were condemning border closures , advocating attending and hugging people at the open air celebrations for the 25 January Chinese new years festival that should have been banned as part of a general temporary lockdown. The liberal New York mayor told people to go to the cinema. Now these same folks are the ones most insistent on not lifting harsh lockdowns.
So neither side of politics is entirely clean but one side is clearly more hypocritical than the other.
William, you’ve given a political perspective, so I won’t argue, but will only point out that numerous timelines have been constructed of public statements by the various parties. They show a rapidly developing understanding of the seriousness of events in Wuhan, beginning in early January. The understanding moved from less to more serious, and that trend was observed almost universally among involved parties.
The key difference is whether the parties acknowledge, in retrospect, that initial mistakes and underestimations were made. The WHO has done this, China has done this. In the US, some of the government scientists have done this as well, but the political leadership has not. Instead they have sought to blame others.
This goes back to my point about the need to admit mistakes and learn from them. The discouraging thing about Trump’s letter to the WHO is the amount of misinformation it contains, as well as failure to accept any responsibility of his own. The true purpose is to deflect criticism and gain support within his political base, who will not question the accuracy of his statements, as has been done correctly and vigorously around the world.
It’s not a document that speaks to the learning, or leadership, that we expect and need right now. Or to building up the world’s capacity to deal with COVID-19, which in some places is only getting started.
In some of those areas, the WHO may be the only organization with the ability to help, as the local governments may not have that capacity. Why would we seek to withdraw resources from that help? You mentioned selfishness, how selfish is it to put at risk a capable response within those countries, in order to defend yourself politically?
These are the things that are so disturbing. If the WHO has problems, why would we not seek to address and/or remedy them? Why is the solution always to claim unfairness and then tear down the perceived opponent? Why does one seek / advocate / encourage a negative outcome rather than positive?
These questions have no simple or just answers. But I think it’s important to ask them.
‘Who is we’ is the right question to ask – in these days everyone is an expert, or thinks he is, and slips smoothly into the ‘we’ mode, as if now ‘we’ exist, as in ‘we the people’, as if a nation, as a community, and so on, and a nation of the expertly informed and capable of making decisions
This is the opposite of the truth
One of the The main objections to ‘lockdown’ ‘martial law’ and so on is these measures were not passed by ‘we’, on decisions not made by ‘we’, from information largely not available to ‘we’
‘We’ never thought much about previous and similar bugs, did ‘we’, but then neither much did ‘they’ – WHO and other national CDC’s and so on, at least not enough to do any adequate preparation or prevention or management or anything at all – apart from ‘they’, and perhaps their ‘we’ too, in Asia
Instead of squabbling about this % that % fatality indexes, as if ‘we’ were putative experts and well informed about the ‘science’, why not address the concerns that are legitimate to the ‘we’ – The politics, not Trump is bad pseudo politics, but the nature and structure of
(Than Asia) Why is and was the ‘West’ so much less well prepared less efficient at decision making, less able to inform and congregate any form of consensus with their ‘we’ the people?
In the West both ‘we’ and ‘they’ behaved with extreme stupidity, many more lives will be cut shorter by the results of these stupid actions and measures than any bug could have dreamed of killing (and more jobs lost more poverty more of everything and anything negative you care to name)
Why is this and what can be done?
KellyAnne Conway is your sister?
So much “alternate Truth” and finely sifting the landscape for snippets marginally fitting your narrative.
WHO judgment never was “just a flue” and it was not static ( how surprising.)
While setting the path for killing DEM voters Trump had enough excess energy for the imbecility of increasing sanctions against Iran to hinder their coping with the epidemic.
While every fourth worker in the US is now jobless the “HaveLots” have gained ~15% in loot in the last 90 days.
The ‘lavishly funded’ WHO gets less per year than the average single major metropolitan hospital in the US. All that money that is ‘given to it’ (or taken away in the case of Trump) is for medical causes that are to be administered through the WHO. It is forbidden from carrying out its own research or from making its own investigations. All this because the ‘great powers’ could not tolerate an UN body from having separate authority.
The “naked facts” are open to much interpretation and confirmation bias Bjorn and won’t become genuine, reasonably impartial facts until there has been a thorough, granular, scientfic, standardised post mortem. I read the reports of multiple different analysts because I’ve found that almost all, at some point, have shown themselves politically biased. So I read all opinions and favour none.
Anyway, looking at the graph you use here it is clear that it is meaningless as support for your assertion as it is not normalised to a per capita basis. It doesn’t mean your assertion is wrong, simply that the supportng evidence is unsuited.
But, FWIW and I can’t provide links, normalised charts I’ve seen online (taking everything with a pinch of salt when only one source providing) show NY state as the significant outlier, with highly controlling and compliant Singapore, and early lockdown (I think) Ireland also having more cases than the USA (in toto). Neither Singapore nor Ireland fits your “managed by leaders that played down the pandemic” assertion. But also does Singapore really have more cases per capita than the USA? Despite the apparent significant spread in foreign worker dormitories it feels unlikely that it does. So the testing and dissemination regimen does matter.
That said, China’s extremely strict quarantine does appear to have squashed transmission at this stage.
More generally, I see the point in following number of cases as the way to manage the exit from lockdown (and so the economic “success” of the way leaders have handled the crisis) but I think “excess deaths” is the more meaningful measure for most people at this stage. After all, this has so far been a non life threatening, even mild, illness for almost everyone under 50 and most of working age.
On “excess deaths”, the FT publish charts for several countries. The FT has in the past had a decent reputation for journalistic rigour. It has been damaged during the pandemic, but nevertheless it is a worthwhile resource. Linking from a tweet as I’m not sure if the URL linked to is personal, follow https://twitter.com/jburnmurdoch/status/1263035490200158209 to see “excess death” plots for several countries (not including Brazilor Russia unfortunately). The timeframe they choose is arbitrary (would be helpful to see the “excess deaths” of each country over the preceding winter) and they rely on existing statsitics which I suspect significantly enhance differences due to data gathering, processing speed etc. (eg death rates by age seems resonably consistent across countries, but the U.K. apparently shows many more in younger people and so many more “excess deaths” overall. Why? Some analysts immediatley claim a poor U.K. approach. It may be, but could also be a reporting/statistical artefact) but they do plot the USA and UK among others.
So, what do they show? That the USA is at c.half (or less) the excess deaths of Spain, Netherlands, Italy, Belgium and the U.K. among “first world” countries. Now, some death rates (eg murder in the specific US context) may have declined due to lockdown, and there is no context of eg lcoal weather, relative internationalisation of cities etc.. But, on face value, on these data sets, for actual deaths, it is simply not justifiable to claim the USA overall has got things wrong.
Woody, the point is not to establish that this or that country got things right or wrong. The issue is far too complex to be that reductive. As we have seen with Sweden, there are costs and benefits to every approach, and also elements of data & presentation can be used to support every side.
Example, some countries in early days, did not count nursing home deaths, although we now know that is where clusters have developed. Currently there is pressure on the CDC to separate the anti-body test results from viral swab results, as mingling them skews the created perception. It’s a learning process and the drive always should be toward the truth.
The point is that right or wrong, you stand up and say this was/is our reasoning, and let it be subject to both favorable and unfavorable review. That is what Bjorn is doing here, but is not what we have always seen from the US, which has tried to deflect responsibility to others, rather than admit possible early mistakes, and learn from them.
Here is the website that Bjorn used to create the plot he posted. It has provisions to look at the data on a per capita basis.
Thanks for posting the graph. It led me to the site you used to produce the graph, which is excellent.
While I realize your post was to answer Grubbie by using data to refute the notion that the pandemic has already passed, I was disappointed to see that your response included additional opinions that are needlessly political and perhaps only partially supported by the data.
If you look at the data, USA daily new cases have been declining for the last 40 days. The decline is gradual (not nearly as rapid as the previous increase) but unmistakably there. Russian daily new cases have also been on a slight decline for the last 10 days. The “accelerating” countries seem to be India and countries in South America and the middle east.
The prevalence and extent of testing does indeed have a significant effect on the number of confirmed cases. Just look at the data when normalized by population. The US is third (Spain is first, Belgium is second but only slightly higher than the US) in the normalized total confirmed cases. However, the US is eighth in normalized total deaths. To date, Belgium has the most deaths per 100,000 people, followed by Spain, the UK, Italy, France, Sweden, and Netherlands. While there are a bunch of factors that contribute to this apparent discrepancy such as, demographics, overall health level, health care quality and abundance, stress on healthcare system, etc. , and even the weather, I maintain that a significant factor is the number of tests performed relative to the population.
Finally, the leadership of many countries and world organizations played down the pandemic. China and the WHO come to mind as well as the US.
Mike, it’s good to see you back, you are missed!
At the Johns Hopkins link for confirmed cases per 100,000 population, the US is shown to be currently 4th in the world:
The interesting thing is the slope of the curves. The first three countries have made the turn and are leveling out, whereas the US slope is not leveling yet. I think this was the basis of Bjorn’s comments.
Unless those countries are actively restricting or reducing their testing, you would not see the leveling that has occurred, assuming the percentage of new cases would remain similar. Conversely, if their testing rates have remained similar, the leveling would indicate a strong reduction in new cases. In the US, the slope is declining which indicates progress, but not yet leveled.
To me, the most interesting thing is to compare the results of the US to Sweden. Despite having opposing strategies (lockdown vs open), the slopes of their curves are similar, with Sweden actually being slightly less.
So the US is approaching herd immunity faster than Sweden. I think that indicates the breath of factors involved in this pandemic. So many things affect it, that even a broad government policy may not have the intended effect, or may not be the dominant effect.
This also may indicate that Sweden’s voluntary compliance with social distancing is working better than the US mandatory compliance.
It’s a good lesson that there is still a major learning curve ahead. That curve is not something to be feared, but rather understood.
read my post again. It was an answer to the two statements by Grubbie:
“…Covid19 has basically already passed…”
“….Trump being able to claim to have been right…”
With a look at the base curve from John Hopkins you can see COVID has not passed. The cruves in key countries, like the US, Brazil, Russia… are not flattening out. In Brazil and Russia they are accelerating, in the US the slope is about the same with new cases now leveled at 25k per day.
Re Trump being right: To claim the difference in cases to countries like China is because the US has more thorough testing than China is not plausible. Taken per capita we have 4787 cases per million in the US and 60 in China. To my knowledge China tested more thoroughly than the US, yet still has 80 times lower infections per capita.
I agree with Woody that when comparing regions with vastly different populations one should use per capita statistics. I also agree with Mike Bohnet that “The prevalence and extent of testing does indeed have a significant effect on the number of confirmed cases.”
I prefer to use per capita COVID-19 deaths rather than per capita confirmed COVID-19 cases as a metric to compare prevalence and growth of COVID-19 in different regions, since I believe that the efficiency with which deaths are detected and reported varies much less across regions than does the efficiency with which cases are detected and reported.
Below are statistics for COVID-19 deaths per million population, as of 5-20-20 and 4-21-20 according to Worldometer, for some of the countries that have been mentioned in this thread.
Cumulative COVID-19 deaths per million population as of 5-20-20.
Cumulative COVID-19 deaths per million population as of 4-21-20.
Percent increase in COVID -19 deaths per million population 4-21-20 to 5-20-20 (a period of approximately one month).
Belgium: 790/518=1.525, i.e.: 52.5%
Spain: 596/455=1.310, i.e.: 31.0%
Italy: 535/408=1.311, i.e.: 31.1%
UK: 526/255=2.063, i.e.: 106.3%
France: 429/319=1.345, i.e.: 34.5%
Sweden: 380/175=2.171, i.e.: 117.1%
Netherlands: 336/229=1.467, i.e.: 46.7%
USA: 284/137=2.073, i.e.: 107.3%
Brazil: 85/13=6.538, i.e.: 553.8%
Russia: 20/3=6.667, i.e.: 566.7%
Singapore: 4/2=2.00, i.e.: 100.0%
China: 3/3=1.00, i.e. o.oo% (china reported 4,634 cumulative death as of 5-20-20 and 4,632 as of 4-21-20.)
I think it is important to keep in mind that COVID-19 prevalence may vary greatly across large counties, and between densely populated and rural areas of the same country.
The US states with the 3 highest cumulative deaths per million population on 5-20-20 according to Worldometer.
New York: 1,473
The US states with the 3 lowest cumulative deaths per million population on 5-20-20 according to Worldometer.
Montana: 15 deaths per million population( 16 total deaths)
Alaska: 14 deaths per million population( 10 total deaths)
Hawaii: 12 deaths per million population( 17 total deaths)
Do lock down measures appropriate for New York makes sense in rural states where there have been less than 20 deaths, and many more rural counties have had no deaths? Can residents of rural areas, where there have been very few COVID-19 cases, be blamed for being skeptical of media reports that show hospitals overflowing with patients when their local doctor and ambulance crews say that they have yet to see a case?
We know that for COVID-19, most cases do not result in death, and that death occurs preferentially in vulnerable populations. So using deaths as the only metric will skew perceptions. For example in Italy, there is both a statistically older population and also multi-generational households where the exposure was much greater, even in lockdown.
Your last paragraph questions are reasonable but raise the issue of cause & effect. Is the lack of rural cases an expression of immunity in those areas, or a reflection of lack of exposure? If the latter, is that lack of exposure aided by lockdown?
I think there’s room for local communities to make their own decisions, but in that there needs to be recognition that absence of infection is not an inherent property, it’s a desirable property that should be maintained by their decisions.
Below is an excerpt from the 5-22-20 Washington Post story at the link after the excerpt.
“The Imperial College researchers estimated the virus’s reproduction number, known as R0, or R naught. This is the average number of infections generated by each infected person in a vulnerable population. The researchers found the reproduction number has dropped below 1 in the District and 26 states. In those places, as of May 17, the epidemic was waning.
In 24 states, however, the model shows a reproduction number over 1. Texas tops the list, followed by Arizona, Illinois, Colorado, Ohio, Minnesota, Indiana, Iowa, Alabama, Wisconsin, Mississippi, Tennessee, Florida, Virginia, New Mexico, Missouri, Delaware, South Carolina, Massachusetts, North Carolina, California, Pennsylvania, Louisiana and Maryland.”
This web site gives real-time R values for all 50 states, based on a modified version of a widely accepted model. Most states are now below 1. with many having started above 1.
This reflects success in driving down the effective transmission rate, but as the authors note, does not necessarily mean the virus is waning. For that we need to see that it stays down as controls are removed.
Also the confidence intervals (error bars) are of interest. For some states like Illinois, Georgia, Colorado, Florida, the interval is small as the uncertainty is small, so the estimate is solid. For other states with more uncertainty, the interval is still large.
In Michigan there has been an increase along with resistance to lockdown, but the value has still remained below 1.
This is what I meant by local decisions retaining the progress that has been achieved. Tools like this should allow us to do that.
Look at the US as set of countries.
You can see that the gravity of infections is now shifting. Early New York was very much up front.
( and at 1500 death/million world leader.)
Now populous states : Illinois, California, Texas, .. are gaining/overtaking.
You won’t see a drop off until all the “infections per state waves” have build up and receded.
No idea yet if this will peter out in the sparsely populated states or just spread slower.
“.. since I believe that the efficiency with which deaths are detected .. ”
death numbers are deformed by health system competence and how much available capacity is (over)tasked.
Look at numbers for Italy, UK, Germany, US.
Italy and UK show intensive care being overrun.
Strangely the US has comparable/better IC facilities to Germany but the death rate goes out the roof. Strange.
IC resources only for the rich?
China, being the first region to be hit had some initial issues ( aided by local administration failures ) corrected by higher up government and then they really bent to the task doing the right things:
Result :: success, epidemic subdued with very reasonable human loss.
Failure of the US;UK,… is aggravated because they would have had coped with just copying China.
( and up front swallowing their tongue before “just a flue” could be uttered.)
Bjorn, after reading pt 1 I thought I remembered seeing a simulation video of droplet spread inside an airliner. Found an artcile about it recently at https://www.popsci.com/article/science/how-sneeze-particles-travel-inside-airplane/. It is for a sneeze, no face mask. So not directly applicable to C19 (coughs would have fewer droplets, ejected at lower speed, and presumably droplet size mix and ejection vector could be different) and maybe not relevant in return to service for many airlines, but of use in seeing what could have happened back in early days of spread.
Of course, this is just one study and should be treated as such.
Also in pt1 you claim the smaller droplets are more dangerous. This is not a valid claim yet as there has not been sufficient scientific analysis (for instance on the relevance of viral load in determing illness and, stepping back, on the ability of C19 to be transmitted in suffcicient load via aerosol). All that can reasonably be said is that the smaller droplets are able to spread further, presumably in free space and at “normal” room temperature (20-22/23C) and a specified RH.
As an aside, the latter point seems to be almost totally ignored across media in referring to how long C19 (or any virus) may last on surfaces. A little digging will reveal studies indicating eg survival time increasing an order of magnitude if temperature drops to 4C. An airliner cabin is much closer to “room tmperature” but also has a much different RH to factor in when determining surface survival times of C19.
The study you reference is made by ANSYS, a CFD tools supplier. It’s not a company that designs airliner cabin ECS systems.
It shows the droplet cloud traveling upwards, then outwards and down, 180 degrees to what the OEMs say is the design of ECS cabin flow. In all, this is an article of questionable value. It takes the aim of using an unproven source for a CFD simulation where you have no knowledge about how its made and of constant movements in the cabin etc. to make the point, “viruses go everywhere”.
This is 100% against the design of the ECS flow by the OEMs and the real studies in the matter we reference here. I will go in detail regarding all this in the next Corner.
Regarding droplet size and their dangerousness, I stated the smaller droplets stay airborne for longer. The first scientific study on COVID-19 droplets was published recently. It says the finer droplets stay in the air for several hours in spaces with no ventilation (no air draught). This is why airliner cabins are safer than most other indoor public places. The cabin has a strong uniform airflow designed to immediately take these droplets away from the breathing area of a co-passenger and then to pass the droplets through a HEPA filter. Only special rooms in hospitals do this and then at 25% of the airflow speed.
I’m aware of some of the brackground of Ansys. It seems strange that a business selling into aerospace clients would risk their credibility with the same by not checking basic details. Anyway, pt4/ECS sounds interesting.
Another study (I believe sponsored by Boeing) run by Emory University and Georgia Institute Of Technology and published in 2018 (https://www.pnas.org/content/115/14/3623) concluded transmission possibility was limited to a maximum 2 seats lateraly and the rows immediately in front and behind. But the study was in situ observational rather than being eg CFD based.
I find the thinness of research, using different approaches, rather surprising.
I’m not sure when you write “study” whether you mean published in a peer reviewed journal, or at pre-press or other. But there was a scientific “study” of C19 “aerosol and surface stability” published more than 2 months ago (https://www.nejm.org/doi/10.1056/NEJMc2004973, 17 March) that I read back then. I think it is the source of “C19 last x amount of time on surface 1, y on surface 2” reports that were reported in newspapers etc shortly after.
What is the study published this week that you are refrring to please? Would be interesting to see how it compares with the one in the NEJM.
Specifically re the “special rooms in hospitals”, are you referring to the use of negative pressure (containment) and positive pressure (exclusion) rooms?
Dear Björn, unfortunately I find the statement that ‘infection during flight did not happen’ in the case of SARS in 2003 a bit optimistic. Air China flight 112 is a well documented case where transmission of SARS took place: https://en.wikipedia.org/wiki/Air_China_Flight_112; https://www.nejm.org/doi/pdf/10.1056/NEJMoa031349. The index case infected 22 of the about 120 on board. Strangely, not his neighbors but people up to 7 rows away. Several conclude from this that aerosol transmission is a viable transmission route, and hence that you don’t only infect by coughing out droplets on which the whole 1.5 M social distancing policy is based. Consensus emerges that confined spaces with little ventilation are the likely cause of superspread events – the church in Korea, the choir in Washington state, the apres-ski locations in Austria, and some sad church gatherings in elderly care houses in my own country leading to devastating implications. Singing, loud talking, inhaling and exhaling deeply seem additional risk factors, plus the time that you spent in such confined places. One may argue that in AC 112 the index passenger showed clear symptoms before boarding already and if we can prevent such people from flying, such cases cannot repeat themselves. Unfortunately another arising consensus is that COVID-19 is spread for 40-50% by people that are infected, but do not display symptoms or are in the 5 days + of incubation in which they do not show symptoms yet. So, unless we have an on the spot test (or an immunity passport) we cannot be sure we allow only people on planes that cannot spread the virus. And then AC 112 unfortunately shows an example that doesn’t look good. The only thing that is an advantage for planes is the high rate that air is refreshed, but the fact you can be up to 12 hours in the same space as a potential index patient when flying long-haul, is an obvious issue. See for some quite well-written summaries on the COVID-19 problem by Tomas Puyeo at https://medium.com/@tomaspueyo/coronavirus-the-basic-dance-steps-everybody-can-follow-b3d216daa343. He, nor me, is a virologist, but as a scientist I am pretty well trained in critically evaluating scientific literature and his analyses seem pretty consistent and logical. I come to this assessment with a heavy heart, being part of a very international community and unable to see many of my international friends – like probably everyone on this forum I’d like to see some way of being able to travel again in the near future.
Air China 112 is one of the cases I reference. The index person had symptoms and was ill during the flight, which I interpret as having fever and coughing (the symptoms of SARS). Also, it’s a 737-300 which is an airplane produced anywhere between 1984 and the late 1990s. Did it have HEPA filters of modern standards?
The widespread of the infections put this in question, as these filters, installed on all modern airliners, filters out 99.9% of germs in the recirculated air including viruses.
AC 122 is the outlier in all the studies together with the flu case, also in an old regional jet where you can question HEPA filter existence in the Environmental Control System (ECS = advanced Air Conditioning in an airliner).
Later studies where HEPA filters are definitely installed, latest the COVID ones I reference in this Corner, all point to the same pattern. With modern airliners with HEPA filters your risk of infection is not during the flight when the ECS is active and the flow has settled down to an orderly top-down flow and recirculation through a HEPA filter. I prefer to rely on studies where we know for sure the ECS has HEPA filters in the recirculation path (without such filters the infection pattern on AC112 is plausible).
Regarding the reference for when you are infectious and the research available, thanks for the links. We went through this subject in the last Corner where we referenced German Charite’s Professor Drosten and his 41 podcasts on the subject.
For the most part most everything points to people passing the virus in close confined spaces “like an airplane.” It just might be too earlier to document this, like science has done for tuberculosis transmission in a plane, and other disease.
It’s not too early, read the three Corners or study the many reports on the subject before giving such blanket statements.
The “confined airliner space” has air ventilation which in flow control and exchange ratio of air surpasses the intensive care hospital standards by four times. We will look into this deeper in the next Corner, there is a reason you don’t get infected on a modern airliner during flight.
I generally agree with what Bjorn has documented, although I think it is still possible given the density of some of the planes. I think the filtration systems are pretty good but this has really never been my major concern. When the air is flowing I think likelihood is low, but when the plane is sitting on the ground and there is little or no airflow is a concern. I know I have been in plenty of planes that sat on the ground with little to no air circulation before pushing back or were delayed. Now maybe the airlines change how they do things but in the past I have been in situations where the air is not circulating in the cabin. Maybe the risk is low but I think thats a hard thing to determine, what does the risk profile look like when the circulation system is not running?
My other concerns are not being on the plane but all the other things one must do to get to and on the plane. The passenger boarding bridge, airline terminal space, airport checkpoints (TSA), the shuttle ride to and from the airport terminal. I think its a much bigger issue than just sitting on the plane during a flight, its all the other stuff that goes along with it.
I honestly do not see air travel getting back to normal measures until there is either an effective treatment or vaccine, just my opinion. Good articles as usual, thanks Bjorn. 🙂
I agree with all you say. Observe how I always state “during flight”.
The key risk areas are passing the airport and boarding and de-boarding. Airlines will be told to have the ECS running at top capacity during boarding and de-boarding for this reason (they have the APU running to supply power and air to the ECS for this), to press down the exhaled air from people as fast as possible from your breathing area. We go into this in detail in the next Corner. Compulsory mask-wearing also changes this for the better.
Might have to change liability so airlines are legally on the hook for passengers’ illness or death should they skimp on cost and not run the air at full speed. Example: flight with SAS out of LHR in August last year. Sitting for a very long time at the gate with no air, LHR doesn’t allow to run APU for environmental reasons and SAS is too cheap to pay for ground power. In case of conflicts between airports and airlines, both go to jail. Since it’s now a real emergency if the airline suddenly kills the AC, would it now in the age of Covid19 be legal for passengers to pull the Emergency Exit and leave the aircraft under such circumstances?
P. I. A Airbus 321 crashed on 4th attempt with loss of power on both engines.
An A320 and it was the first approach but was approved for a go around
avherald indicates that one go around was with retracted gear and “touching” the engine nacelles with resultant damage onto the tarmac. emergency power generator extended visible on video footage.
Hello Dukeofurl and Uwe,
Below are some excerpts from a 5-23-20 Times of India article at the link after the excerpts regarding the Pakistan A320 crash. It appears that the designers of the A320’s idiot proof flight control computers may have forgotten to include a push button for fixing engine damage after landing with the gear up during an severely unstabilized approach, that comes closer to being a dive bombing attack than it deos to being a competently executed airline transport approach.
“The aural warning in the cockpit appears to be of landing gear not down and locked. Preoccupation due to an unstablised approach could lead to pilots missing out some warnings,’’ said a senior B747 examiner.
The ATC recordings indicate that the aircraft was initially too high on approach to land. Capt Amit said, “Flightradar24 granular data shows that the aircraft was about 2500ft higher than the required profile. Prior to that the ATC had instructed the aircraft twice, to discontinue the approach but to no avail. The crew apparently persisted with the approach descending at a very high rate of 2000ft/min or more till reaching 1500ft. The rate of descent was then lowered to 1000 ft/min.”
“It’s not confirmed yet whether the engines did make contact with the runway surface during the first attempt at landing. But going by the photographs of the PIA aircraft, taken when it was air borne during the go-around, it does seem to be a possibility. Marks indicating damage to the lower portion of both the engines is distinctly visible in the photographs, a senior A320 commander said.
The pictures also show the Ram Air Turbine (RAT) extended. A RAT is a small two-blade wind turbine that is deployed automatically in an A320 aircraft-it protrudes out from the aircraft underbelly to take advantage of the wind sweeping the lower portion of aircraft-to generate an alternate source of power when the aircraft’s conventional sources of power have failed, which is what would happen in a dual engine failure. “Since RAT was deployed, it’s likely then that both the engines had failed. The engines probably sustained severe damage during the first landing and it was exacerbated during the go-around and the engines failed,” said the commander.”
I believe electrical generators are usually installed near the bottom of engine nacelles. Perhaps both the right and left generators were both damaged when the engines scraped the runway during the first landing attempt and that is what initially triggered deployment of the RAT during the go-around from the first landing attempt? The aircraft would not have been able to execute the go-around if the combined thrust from both engines was very much less than 50% of normal takeoff thrust.
Another possibility is that lubrication was damaged, such that the engines would continue to function for a short time during and after TOGA, but eventually seize or fail, leading to loss of thrust (as reported) and RAT deployment.
We don’t know without data, but it may be the crew realized the gear were up after the flare, then had the choice of a belly-landing or go-around. They chose the latter but the engines could not spool up in time to avoid contact with the runway.
The approach was high and hot and the CRC was heard in some of the ATC traffic. But we have to wait for the recorder data to learn what exactly happened.
Re: “Another possibility is that lubrication was damaged, …”
That is a definite possibility.
According to the videos at the links below, on a CFM-56 the following components are all installed near the bottom of the fan casing, i.e. the widest part of the engine, which would likely be the first to contact the ground in a gear up landing. The engine oil tank is located on the side of the fan casing, at about the 3 or 9 O’Clock position. While textbook diagrams of jet engines are simple, what a complex piece of machinery an actual jet engine is!
Fuel pump,filter, and flowmeters
Oil pump. filter, and temperature sensors
N2 speed sensor
Engine fluid drain plugs
Being dragged along the ground at 150MPH is not likely to positively affect the operation of any of the above systems or components.
Below are some excerpts from the 5-24-20 Times of India article at the link below, reporting on the information in a preliminary investigation report.
“According to the report, prepared by the country’s Civil Aviation Authority (CAA), the Airbus A-320’s engines had scraped the
runway thrice on the pilot’s first attempt to land, causing friction and sparks recorded by the experts.
After the third impact, the pilot took the aircraft off into the air again, which officials found very strange as the crew in the
cockpit did not inform the Air Traffic Control (ATC) at the Jinnah International Airport of any problem with the landing gear,
The News International quoted CAA sources as saying.”
“When the aircraft scraped the ground on the first failed attempt (https://timesofindia.indiatimes.com/topic/first-failed-attempt)
at landing, the engine’s oil tank and fuel pump may have been damaged and started to leak, preventing the pilot from
achieving the required thrust and speed to raise the aircraft to safety, the report said.
The pilot made a decision “on his own” to undertake a “go-around” after he failed to land the first time. It was only during the
go-around that the ATC was informed that landing gear was not deploying, it said.
“The pilot was directed by the air traffic controller to take the aircraft to 3,000 feet, but he managed only 1,800. When the
cockpit was reminded to go for the 3,000 feet level, the first officer said ‘we are trying’,” the report said.”
FlightRadar24 has published the altitude and airspeed data. The aircraft was flying fast and high on an unstabilized approach with at least a 6 degree glideslope. The pilot said he had ILS lock, but either this was just the localizer and not the glideslope, or he had the 6 degree glideslope lobe instead of standard 3 degree.
There is speculation that the gear did not extend on command due to excessive airspeed, and without the drag of the gear, it would be very difficult to both descend and lower speed. At the flare, airspeed was above 250 knots, which is about 100 knots over normal landing.
The gear warning may have been lost in other alarms for sink rate and speed. If it does not extend initially, it must be recycled after airspeed is reduced. So possibly they intended to reset after slowing down, but never really slowed, or they became rushed & task saturated in the very fast descent and approach.
One pilot has simulated the flight and showed that the gear could refuse to extend at the start of the descent. He also showed two options for recovery after the touchdown and engine failure, but they both require an immediate direct return, with knowledge that engine life would be about 4 minutes.
Re: “At the flare, airspeed was above 250 knots, which is about 100 knots over normal landing.”
I found this very hard to believe, this would likely be much faster then the maximum speed for extension of landing flaps, so I was motivated to take a look at the FlightRadar 24 data. The extended mode S graph at the link below shows IAS around 240 knots at 2000 feet AMSL, around 220 knots at 1000 and 500 feet AMSL, at around 190 knots at the minimum in the altitude trace during the first landing attempt, and dropping to a minimum during the first landing attempt of about 170 knots during climbout from the landing attempt. According to Wikipedia, the elevation of the Karachi airport is about 100 feet, in which case altitude AGL would be approximately (altitude AMSL -100 feet).
This is a certainly a very fast approach, but (from the Department of Nitpicking ), I would describe the flare speed as having been around 220 knots rather than “above 250 knots”.
I believe typical landing speed for an A320 would be 130 to 140 knots, which if true and the landing speed of PK8303 was 190 knots, as I suggest above, means that the landing speed of PK8303 was 60 to 50 knots too fast.
A landing speed of 190 knots (218.5 MPH) would have been about right for the Space Shuttle Orbiter, but a little fast for the X-15 rocket plane.
“The orbiter’s main landing gear touches down on the runway at 214 to 226 miles per hour, …”
“Because of the large fuel consumption of its rocket engine, the X-15 was air launched from a B-52 aircraft at about 45,000 feet and speeds upward of 500 mph. Depending on the mission, the rocket engine provided thrust for the first 80 to 120 seconds of flight. The remainder of the normal 8- to 12-minute flight was without power and ended in a 200-mph glide landing. Because the nose landing wheel lacked steering and the main landing gear employed skids, the X-15 had to land on a dry lakebed. ”
From the data I referenced (the link you posted), there was an initial flare at about 1500 feet, probably to adjust glideslope, during which speed went above 250 knots. That indicates the aircraft energy was far too high on approach to landing, which was my main point.
At the second flare just before touchdown, the speed was still 225 knots, still far above the normal landing speed.
Airspeed only decreased to 170 knots after dragging along the runway for thousands of feet, and then lifting off again. The aircraft traded altitude for airspeed, but even at that point, it had just reached normal landing speed.
So I stand by my original reasoning and comments. The basic point was the aircraft could not dissipate enough energy to achieve a normal landing.
A320 maximum flap extension speeds according to the link below.
Slats 18, Flaps 0 – Holding: 230 knots
Slats 18, Flaps 10 – Takeoff: 215 knots
Slats 22, Flaps 15 – Takeoff/Approach: 200 knots
Slats 22, Flaps 20 – Takeoff / Approach /Landing: 185 knots
Slats 27, Flaps 35 – Landing: 177 knots
If the Flight Radar 24 extended mode S graph that I referenced above and my reading of it are correct, and on the first landing attempt PK8303 was at or above 220 knots until around 500 feet AMSL, at or above 200 knots until around 350 feet AMSL, and landed at about 190 knots, then the aircraft was travelling above the speed limits for all landing flap settings during the entire first landing attempt, and above the speed limits for all approach flap settings until around 350 feet AMSL.
Yes, in the pilot simulation that recreated the flight profile, it was not possible to select more than flaps 1 during the approach, due to excessive airspeed.
The extended data file shows that the airspeed at touchdown was 215 knots, decreasing to 191 knots during the runway contact, then lifting off again with a gain in altitude but a corresponding drop in airspeed to 173 knots. That was the minimum airspeed recorded for the landing attempt. As I mentioned, the aircraft traded altitude for airspeed at that point.
From the Department of Nitpicking:
Re: “The COVID-19 virus full name is SARS-CoV-2”
COVID-19 is the name of a disease that is caused by a virus named SARS-CoV-2. Below is an excerpt from the WHO website at the link below.
“Official names have been announced for the virus responsible for COVID-19 (previously known as “2019 novel coronavirus”) and the disease it causes.
The official names are:
severe acute respiratory syndrome coronavirus 2
Thanks, text corrected. Good we have such a department :-).
Thank you for all your informative and always well researched posts.
I think as additional studies and data become available, they will support the notion that the risks of air travel can be managed, which is Bjorn’s main point. It will require adjustments and compromises in some procedures, but there’s not reason to believe it can’t be done.
In the meantime, the first travelers will be “early adopters” who find the personal risk acceptable. They will help generate the data which will confirm or refute the risks, and the effectiveness of procedures. For those who are not willing to accept the risk, they will avoid flying at first, and that’s perfectly fine.
If you are afraid of covid-19 DO NOT travel by air.
Let us fly without you.
I wish there were more detailed studies published showing the actual transmission history in a small area. It might answer some questions on how to act in public safety wise. Is 6 feet enough distance? How well do masks work? Is staying up wind (via Air Conditioning etc) a good idea?
How long does it take for the virus to hop from one person to another in different situations?
How many passengers were tested by the airlines, or test result communicated to airlines before the lock downs? Desperate airlines and the IATA paid by them. Suspect sources for objective, terribly incomplete data.
We should be honest, objective. Not positive, siting incomplete figures & create perceptions.
Aircraft are hundreds of people densily packed together for hours.
Passengers touch everthing, seats, bins, tables, lavatories, catering, themselves, IFE, shared armrests. Cabin air filters, even carbon, can’t deal with the virusses quick enough.
A new study showing the effectiveness of a simple mask. If airflow in an aircraft, or some sort of hooded mask system can be utilized, and strict protocols put in place, maybe it could be safe to fly again?
There is a non-zero probability that no vaccine will ever be assembled against COVID-19.
Same as AIDS, but of course COVID can spread much more easily. That disease is 40 years old (at least).
What is the acceptable passenger experience ? Wearing a mask on public transports (metro, bus) for 20-30 minutes a day is unpleasant. Wearing a mask on a 10-hour long-haul flight is a no-go: many/most potential travellers will prefer not to travel and use other forms of communication (business travellers), or entertainment (leisure traffic), or transportation (cars).
The very wealthy will revert to bizjets.
The mere mortals will not accept (or very reluctantly) being squeezed into seats with unknown strangers around them.
This is not an issue this month when aircraft are being reactivated but the load factor is around 25-40% (instead of 85-90%). Not viable for any airline.
Should be aircraft cabins be re-designed to offer more privacy and less infection risk ?
Any re-design must be safety acceptable, ie. vertical plastic shields around seats are probably not an option.
So any re-design will probably mean less passengers per cabin. Which means higher prices. Is the low-cost model doomed ?
What will be the new economic model ? Will economy class look like the layout of most business class cabins, but with simple seats and parsimonious food, and significantly higher prices ?
New lavatories have to be invented. Today most flights have one lavatory exclusively reserved for cabin and cockpit crew members. Not viable on a 180-seat airplane. One can imagine a flow of UV radiation in the lavatories between two occupants.
At the end, the flight experience can be improved, compared to pre-COVID service levels which were becoming almost non-existant.
Let’s use our imagination.