To Lhasa – By Train or Plane

The short version : adapting to high altitude takes time. You can travel to Lhasa fast , and spend time in hangover country before coming closer to full capacity again. The other option is to spend some time enroute , and get more out of each day. This should be an easy choice for anyone who has more time than money , since the time  in Tibet will be the most expensive days  on any trip to China.

Train , as in the nonstop Beijing -Lhasa run , or flying in , also from Beijing , are the options most will chose between The benefits or risks are the source of intense speculations , but with the massive numbers of travellers to Lhasa  there is also some hard data now.

Reality check : one in three will get altitude sickness , coming in by train  from Beijing .
And this is good news : twice as many get it on flyins.

The high figure for AMS is a shocker for many , but is in the same order for Cusco travellers , and a 2011 study from Lhasa, which also found that AMS was more common on flyins. Early stages are of AMS are undramatic , and many won’t even think of the hangover like situation as sickness. On the other hand a few , depending on bad luck in the genetic lottery in combination with really fast ascent do get severe AMS. The rare and worst case scenario is pulmonary edema , which only appeared in the flyin group – see the notes and reading suggestions below.

There is a rich flora on speculations regarding the train vs plane issue , the main reasons given for not taking the train being that it a) runs too low for benefit , or b) it runs too high . A conservative plotting of altitude as the available amount of oxygen , factoring in the extra oxygen on the train , gives this elevation profile between Xining and Lhasa :



Fly in to Xining at 2283 meters , sleep minimum one night there . Every extra night at 2000+ is a major positive factor. Next leg of the trip to Lhasa by train or plane . The reason you probably haven’t heard of the flyin option from Xining is that you couldn’t , until this year. Train is still better than the plane , from two reasons : you get more time between Xinings and Lhasas altitude , and you spend your first day at 3000+ meters at rest. Think of it as being locked up in a Lhasa hotel room , with a oxygen tank by your bed . And with a lot better views.


Fly in from one of the international airports Beijing , Kunming or Chengdu . Obviously not my recommendation for firstcomers to high altitude – the chances of getting really sick are low , but the chances of the first day or two being spent in hangover country are excellent.


The really good news here is that there are a number of interesting options that are even better than the two day options , if you are willing to spend more time .

Take the train from Lanzhou , near the Beijing-Lhasa halfway point at Xining, and add acclimatization time at Labrang. Lanzhou is at 1600-ish meters , which is too low for effective acclimatization. Labrang monastery , an important pilgrimage site for Tibetans , is only a few hours away with bus , at 2945 meters.

Kunming ( altitude 1850 ) : International airport , and railway point.

Still too low for effective acclimatization , but a convenient gateway to the large Kham Tibetan region in Yunnan. Take the train to Lijiang @ 2400 meters ( and a day tour on the worlds highest cable car ) , onwards to Shangri La/Gyalthang .


Shangri La old town was very interesting until a few months after my visit , when a massive fire destroyed most of it. Lots of other good reasons to visit the region – see  Losangs blog for a wealth of information on Kham and Amdo regions.

Chengdu : International airport (CTU) , and railway.

Chengdu is abysmally low , but gives  easy access to higher ground in the Kham Tibetan region. Kangding/Dartsedo is reached same day with bus. Despite it’s modest altitude at 2600-ish meters Kangding is serviced by the third highest airport in the world at 4280 meters , so you can boost your acclimatization by flying back to Chengdu and take a flight to Lhasa . A cheaper , more environmentally friendly day trip  to 4000+ is taking the cable car from near the town center.


The medical issues with travelling fast to 3000+ meters altitude are well known since decades back – see the guidelines from International Society for Mountain Medicine or CDC. With litterally millions coming in to Lhasa every year , two million by train , it’s now also  possible to make large scale comparisons between different travel modes , with comparable  groups.

Incidence of High Altitude Illness Among Unacclimatized Persons Who Acutely Ascended To Tibet ( Yeshung Ren et al in High Altitude Medicine & Biology )

This describes the results  of  monitoring flyins to Lhasa in a 3268 strong group of military personell. 57 % had various degrees of AMS , 12 %  vomited and received medical attention  , 2 % started to develop pulmonary edema.

A Survey of Acute Mountain Sickness And Vital Signs in Subjects Ascending To Lhasa Via The Qinghai-Tibet Train ( Younjun Luo et al in Scientific Research and Essays )

A similar group of military personell , forty nine persons , was closely monitored during the train ride and for a long period in Lhasa . AMS peaked in 14 % of subjects during the first three days in Lhasa. Worst outcome , in four out of fortynine , was a AMS score of 4 at the Tangu La pass high point and first and second day in Lhasa. One person vomited at the Tangu La. None developed pulmonary edema.

Altitude Illness in Qinghai-Tibet Railroad Passengers ( Tian Yu Wu et al in High Altitude Medicine & Biology )

Tian Yu Wu was a key figure in planning the health program for the workers constructing the railway to Lhasa , the worlds largest construction project at extreme , up to 5000+ meters , altitude. The key data here comes from 160 random Chinese lowlanders travelling straight to Lhasa , without any acclimatization stop at Xining. 31 % developed AMS , 4 % vomited at the Tangu La. Worst outcome : one person developed a balance disorder , which improved after receiving extra oxygen. He was given intermittent oxygen at Lhasa hospital , a clean bill of of health after a CT scan of the brain , and resumed his Lhasa tour after a single day of observation in the hospital.

Acute Mountain Sickness among Tourists in Lhasa, Tibet  – A prevalence study ( Labazangzhu , from Oslo University )

This is the result of survey among around two thousand tourists in Lhasa , nearly half Chinese , which gives a narrow advantage for the train and shows a AMS incidence of 51 %.  It has a number of problems making valid comparisons . It includes many different travel modes , persons normally living at 2000+ meters , and for example people travelling from Kathmandu in Nepal , which will include both unacclimatized persons coming directly from Kathmandu , and well acclimatized persons after  trekking at 3000+ meters.


AMS is obviously common enough on both the nonstop train run and flyins to Lhasa to consider medicating with Diamox. Talk it over with a travel doc , not your GP. In China you will invariably be offered Rhodiola/Hong Jian Tin instead , which has no documented effect ( including the few who used it on the the train in Wu’s study ) . Rhodiola is often presented as a unique Tibetan herb and tradition : it’s not . The plant is found over large parts of Asia and Europe , including my home mountains near the Arctic Circle and the Pyrenees . It has  been known to western tradition from the first century greek physician Dioscorides , was renamed by Linnaeus in the 18th century , and became a part of the Stalin era medical research in the twentieth .

The adjusted elevation profile comes from recalculating the effective altitude from the inspired  amount of oxygen on the train , factoring in the higher oxygen concentration after Golmud . John B West , editor of High Altitude  Medicine & Biology , measured the O2 concentrations on one of the first runs of the train , and calculated the effect to a 900-1200 meter lowering of the effective altitude. I plotted it as one thousand meters lower  , with a build up factor during the first 100+ klicks after Golmud – which still are lower than Lhasa on the map.

Bad Altitude Info Award , June 2011

Just when you thought you had heard it all , from the Loveland Ski FAQ : make a solid foundation for your cerebral edema with low sodium diet , and high water intake – starting one week (!) before ascent.

There is no full proof method for avoiding altitude sickness, but drinking extra water and avoiding salty foods the week before you arrive seems to be the most helpful. For more info click :

( No , the American Heart Association has never , ever said anything like this. )

Runner up : the New York Times travel blog , for high lighting the dangers of the train ride to Lhasa – and failing to notice two separate oxygen delivery systems in every car , at every seat.

A Trans-Himalayan Train , Hurrah (?)

The gleam in the Himachali politicians eyes have have started grow a little more intense : a definite proposal for a Manali-Leh train has been kicked to national level by the Railway Ministry.

Details are still scarce , but the cited distance for Bilaspur-Manali-Leh , 489 kilometers , is nearly identical to the present Manali-Leh road. This should mean that they are aiming for westerly shortcut from Darcha , same as being underway for the Manali-Leh highway.

Some question marks : the Rothang tunnel work that has started is a dual lane road only , no data on if it has been planned with a third railway track in mind. The sale pitch is that this will be an all year round open route once the Rothang has been bypassed : some Ladakhi politicians have been less convinced that this will work out , and have pushed for a a route over Tso Moriri ( ouch ! ) to Spiti and the NH 22/ Hindustan-Tibet road as a more practical solution.

Indias border policy has involved a certain conscious under-development of roads and railways , with the idea that poor roads slowing down the enemy would be a strategic asset : Spiti at least should perhaps be seen as a resounding success in this strategy…

The current road and railway work in Tibet and adjoining areas ( railway to Lhasa , construction/planning of tracks to Shigatse and Nyingtri , highways to Yadong and work on re-opening the Stillwell /Ledo / road thru Burma) has been used a lot over the last years as arguments for abandoning this policy.

Meanwhile , in practical terms : expect very little change in the coming years. We are still waiting for the completion of the railway to Srinagar , soon a decade over the schedule , and there is still no formal decision or any money allocated in this years railway budget.

High mix : oxygen on the Tibet train

Oxygen is supplied in two ways on the Beijing-Lhasa train . The visible hardware is the oxygen outlets placed under seats , and on the walls in compartments and corridors. Double pronged soft nasal cannulas (not the most efficient delivery , but somewhat more comfortable ) are used with



The invisible part is the extra oxygen that is added directly in to the air of the compartments, to a concentation between 24-25 % .


Oxygen pressure at different elevations at 21 % and 24 % (measured by John B West , 24.1 -25.3 %, see the Revisited .. post)

The only figure that relates directly to how how much oxygen there is in what we inhale is the oxygen (partial ) pressure . One fifth of air is oxygen , this means 21 kilopascal at sea level. At the highest point along the train ride , Tangula pass at 5076 meters the ambient oxygen pressure has been reduced to 11 kPa but inhaling 24 % oxygen(upper curve)  instead means that you inhale the same amount of oxygen as on  a  lower elevation. At Nakchu , 4500 meters , you will actually be marginally better  off than in Lhasa.

Oxygen is added from Golmud (2800 meters) and onwards. The oxygen is produced by oxygen generators that separate the nitrogen and bleed the oxygen into  the compartments. This is the same technology that has been used for a number of years in for example Chajnantor Observatory , one of the worlds highest civilian workplaces at 5050 meters. 

Using the supplementary oxygen lines : 

The main advantages of taking the train lies in the time spent between Xining and Golmud (2400-3000-2800) , after that conditions are very close to being at Lhasa´s altitude. Most , but not all , handle coming up to Lhasa´s altitude without  any major problems after a intermediate stop but I have for example met one person with early stages of altitude sickness in Tabo , some four hundred meters lower than Lhasa.

Use the oxygen lines if you have any symptoms of hypoxia , i.e. always if  you have headaches. Oxygen saturation in the blood will always drop as you lie down, if you had any symtoms during the day you will at least sleep better with oxygen.

Tibet Train , Revisited.

The March issue of High Altitude Medicine & Biology holds an interesting trip report from the Qinghai railway to Lhasa. It brings up the same reservations I held in my earlier post about the onboard oxygen system – and refutes them in some detail.

First of all it leaves speculation aside : John B. West brought along a oxygen meter , which actually showed somewhat better result than what has been said earlier. 24.1 to 25.3 per cent , comparable to what´s being served in recovery wards after routine anesthesia .

My practical reservations about this system could be summed in two simple questions : what happens when you pop off to the restaurant wagon , and the train stops to let off passengers . Opening the doors should mean that the atmosphere more or less instantly equalises with the outside.

The answer seems to be that they don´t open the doors , at least not for the first six hours :

In fact , we only stopped three times between Golmud and Lhasa , and only at one of these did all the doors open so we could stretch our legs.At the other two apparently only one or two doors were opened for passengers who had arranged to leave or enter the train at that stop. “

Another interesting aspect is what it shows on the changes in modern China , which is illustrated by what is considered acceptable human cost : building the first highways in to Tibet came at a staggering cost in human lives , the Sichuan-Tibet road took 3000 lives , and the road between Xining and Jyekundo meant more than ten dead per kilometer. (See Losang´s article on the Tibet Highways. )

Today the Qinghai railway company claim, with support from the doctors responsible , that not a single worker has died from altitude sickness in what was the highest construction site in the world.

True or not , it is obvious that huge resources were put in place to achieve this goal. Workers had a minimum transport-only time of four days to the highest work sites , oxygen bars were rigged for recuperation … and large field oxygen generators were actually pumping oxygen to the drill face in the highest tunnels, to achieve a few percent increase in air breathed by the tunnel workers.

When all of this failed there was a willingness to pay another kind of price : doctors had and exercised a right to evacuate workers by helicopter , often based principally on a failed balance test.

(West´s editorial is at the moment available  on line , a description of the of the health/safety organisation during the construction can be glimpsed in this article ).

High Tea : Railway to Lhasa

In 2004 theTibet train world started to become aware that the the ralway to Lhasa indded was becoming a fact. The near-completion started a chain of reactions : in some Chinese netizens it became an object of national pride (some would say chauvinism) , and among others it was seen as one more nail in the coffin containing Tibetan culture. The Wikipedia  article soon became a  battle ground along predictable lines , starting with What is Tibet , anyway ?

In the media , a lot of attention was centered on the engineering challenges of worlds highest railway , running over 5000 meters high , and the brand new , shiny thing : the worlds first train with a pressure cabin . Everyone – the BBC, Guardian etc went aah .. – so did I .

The one question that would have blown this story to bits was never asked : so what happens when we pop off for a nice cup of tea in the restaurant car that looks so nice ? No air locks ..

The pressure cabin myth has acquired a life on it´s own : you still hear of people boarding the train with that thought , being awakened by the sound of hard drives going clunk! . Bye , bye , iPod.

What eventually materialised was a train with oxygen generators , spiking the atmosphere up to 23-24 percent (according to the latest issue of High Altitude Medicine), and nasal cannulas with low (undefined ) flow of pure oxygen.

Oxygen enrichment was not a new concept , it has been used in high altitude observatories up to 5000+ meters for a number of years , but the train solution is obviously a compromise between comfort/safety and economy : the observatories run by Caltech University use for example 24 % in their observatory at 3800 meters , and 27 % in their highest facility the Cosmic Background Imager at (corrected for difference in air pressure) around 4800 meters , very close to the long high run of the Tibet train.

So how does these numbers work out ? Well , in the context of travelling to Lhasa : surprisingly well . Assuming these figures are correct , you will breathe the same effective oxygen concentration at 4500 meters as in Lhasa – with the added benefit of having access to the supplemental oxygen by nasal cannulas.

This is if everything works as stated – I would expect the whole car dropping down to 21 % more or less immediately every time people get off the train, for one thing .

Not worse than flying in , with a very limited advantage in the time spent before Golmud . Basically the first twenty or so hours on the train is a dead loss , acclimatisation wise , you need to reach something like 2000 meters to start effective acclimatisation. If your motivation to take the train is to be gentle on your system , you will profit a lot more by flying in to Xining , and take train the next day . Golmud would be even better , at 2800 meters , but by all reports it´s harder to arrange your travel onwards from there.

Minimising your time in Beijing , if you have flown in , will also a give a subtle advantage , tying together the time at cabin pressure (2100 meters, at best  ) with the second half of the track.