In reply to my email, Søren wrote:
I was the head of the forecast office in 1996 and remember us setting up a service for a Danish Everest Expedition.
We chose to supply wind and temperatures derived from atmospheric models used for weather prediction. Exactly which data we used and supplied has escaped my memory, but I may be able to find out.
We even gave the forecasters information on mountain climbing and routes on Everest based on expedition accounts by Reinhold Messner.
The weather forecast for Everest was, if I remember correctly, done by one of the duty forecasters. Hence it will not have been the same forecaster every day but a group of forecasters.
This, along with Audrey Salkeld’s posting of 5 May, confirmed beyond doubt that accurate weather forecasts were being provided directly into Everest Base Camp prior to 10 May.
I contacted Søren again to ask if he knew what source of data the DMI would have used to provide these forecasts. Whether they might have any of that data still stored in their archives.
The reply came back:
I am afraid I cannot help you much on this level of detail. I do not have the instructions to the forecasters any more and the forecasts themselves are no longer in our archive. What I am sure of, though, is that the forecasts were based on data from ECMWF (European Centre for Medium-Range Weather Forecasts) who may be willing to help you with the raw prognostic data. Additionally, the forecasters will have compared the model data with observational data from the area and with positions of jet streams and wind fields from World Area Forecast Centre London (UK-MET OFFICE).
Normally the quality of forecasts for the Himalayan area should be expected to be quite high due to the fact that it is downstream from the data-rich European area. Although satellite data is a very important source feeding into atmospheric forecast models it would be wrong to say that forecast data is derived from satellite data. A huge amount of data from many sources feed the models ranging from data from weather stations on the ground, weather balloons, data collected by commercial aircraft to various kinds of satellite data.
It will have been very difficult to sample meaningful data on the ground in the Everest region. The data would simply depend too much on wind/lee or shadow/sun to be used as a forecasting tool. We carefully pointed this out to the climbers. We said what we probably would be able to predict with some skill was the temperature and strength of the flow at different heights, but not the temperature and wind on a specific location on a face or ridge of Everest. Technically it could be done to some extent, but would have taken considerable amounts of time, effort and all the expedition’s money.
Regards
Søren Olufsen
The extra information Søren gave me in this reply would turn out to be invaluable, not least due to his reference to the ECMWF data and the fact that his forecasters would have compared the model data with information from the UK Met Office.
While looking up the names of Danish climbers on Mal Duff’s team in 1996, I came across that of Dr Henrik Jessen Hansen, leader of the Danish expedition and, at the time I contacted him, a consultant in the Department of Cardiothoracic Surgery, Gentofte University Hospital, Copenhagen. In fact, it was Henrik Hansen who set up the triage hospital at Camp 2 immediately after the disaster, ready to receive and treat the injured as they were brought down from higher on the mountain. While Henrik was organising this, Mal Duff and one of his clients, Euan Duncan, made a rapid descent to Base Camp to collect emergency medical supplies.
I wrote to Henrik and asked if it was himself or another member of the Danish team who obtained the forecasts from Denmark, and if they compared these with one the Imax team were receiving.
Henrik’s reply:
‘All the Danes on the Mal Duff team were together and we had an agreement with the Danish Meteorological Institute and we exchanged information with the Imax team.’
Although Henrik’s reply hinted that the Imax team had their own forecast, it did not actually go as far as to say this. Audrey Salkeld had made no mention of a second forecast in her report of 5 May, so maybe there wasn’t one.
I relished going to see Geoff that Friday evening. The opportunity to land him unexpectedly with something this important was a rare pleasure indeed. This time I’d given him no warning of what was coming.
Leaning across the table, I presented Geoff with the paperwork, consisting of the 5 May report from the PBS archives and the replies I’d received from both Søren and Henrik. I sat back, quietly sipping my beer. Although slightly perplexed by the fixed smile on my face, he carefully made his way through the information before him, nodding to himself as he did so. Finally, after several minutes, he gradually raised his head.
‘Good lad!’ were his first words.
Praise indeed, I thought, from this retired Yorkshire lawyer – a friend whose opinion I valued so much. Only for him to follow by saying, ‘But you know you’re going to need more.’
In the knowledge that Henrik provided my best hope, I contacted him again. I asked if the Imax team had also received a forecast from the DMI or if theirs had been from a different source.
Henrik’s reply:
Imax weather forecast was from London. I don’t remember the institute. I don’t remember the details but ours were pretty good but no way perfect. There was a general feeling of immortality and belief in the 10th as a good summit day because Rob Hall had succeeded on this date three times before. Actually several expeditions including Hall and Fischer had a meeting around the 1st of May and there decided the 10th should be the day – and that without any forecast going that long.
Although this confirmed a separate Imax forecast, and suggested these were being received before 10 May, it did not say that unambiguously. David Breashears had admitted he contacted the London-based Met Office, but it was not clear from the information I had read, or from what he had told me directly, if he meant before or after 10 May.
Something else Henrik had said intrigued me. Neal Beidleman had told me that: ‘In ’96 not much credence was put into the forecasts.’ Yet here was Henrik saying that although the forecasts were in no way perfect they were actually pretty good.
I asked Henrik if the Danes had received five-day forecasts and whether both his team and the Imax team had received and compared these two separate weather forecasts, from Denmark and London, at the beginning of May.
Henrik’s reply was the answer I had been searching for:
The London-based institute were quite famous and had a specific name that I do not remember. We did have five-day forecasts, maybe Imax were a bit longer.
We did receive and compare the forecasts throughout May and did not find the 10th good [the Danes therefore scheduled their planned summit bid for 12 May].
Henrik saying that he did not find the 10th good suggested trying to summit even on this day would be ill-advised, let alone on 11 May. First Rob and then Scott had been motivated by their belief in the 10th as a good summit day; this was based on the success that Rob had achieved in previous years.
I asked Henrik if the name of the London-based institute had been referred to as Bracknell.
His reply:
‘It might be, it sounds familiar.’
As I had with Ed Viesturs and Neal Beidleman, I put to Henrik my theory about how the summit attempts had been planned. I also let him know that I had been in touch with Neal. I enquired whether Rob and Scott had got the forecast through the Danes or the Imax team, or if everyone had sat down together to consider both forecasts. I went on to suggest to Henrik that Rob and Scott had seen a forecast before heading up for their summit attempt.
Henrik started his reply by saying he remembered Neal very well and could I say hello to him. He continued:
Everyone did share forecasts and it was correct that they got a forecast as they moved up the mountain but another very important point was that both teams’ clients [Rob and Scott] were so weak that there will only be one go and afterwards they will be worn out and go home, summit or not. And the competition between the two and their journalists were very present although there was some cooperation.
I wanted to make sure that there wasn’t a point at which these teams might have dropped out of touch with the five-day forecasts. To put this beyond reasonable doubt, I sent Henrik two further questions:
Did you get the five-day forecast you were receiving from the Danish Institute at the very beginning of May 1996 once every five days or did you have the five-day forecast updated each day?
Was it the Danes or the Imax Base Camp that radioed up the weather to Rob, Scott and the Imax team as they moved up for their summit bids?
Henrik’s reply indicated the actual volume of information they were receiving:
We did receive a forecast every day in May. I do not remember who radioed the forecast, but most likely it was the Base Camp managers from each expedition. [The Danes, Imax team, Rob and Scott.]
There was one if not two forecasts being received by the Danes, the Imax team, Rob Hall and Scott Fischer every day in May. Why had it taken me three years to finally get this information?
From the Imax team, I had Audrey Salkeld on record, but I thought it might be wise to see if I could find a second source from within this team confirming that forecasts were being received prior to 10 May. With nothing concrete coming from my direct enquiries to members of that team, I realised my best chance lay on the fringes. I began a search of more obscure articles on the Internet relating to Everest 1996. Amongst these I came across the summit journal of Ed Viesturs, posted on
Outside
magazine’s Internet counterpart called
Outside Online
. In its archives, I discovered a fascinating report sent back on 5 May 1996 by Paula Viesturs, the Imax Base Camp manager (and Ed’s wife):
But it’s still nice and clear at the summit. The only thing, really, seems to be the winds. We just got an updated weather forecast and supposedly the weather is going to hold. And these forecasts have been accurate so far. The only issue is the wind.
I would later discover that David Breashears had taken part in
Frontline
live online on Wednesday, 14 May 2008; this was reported in the
Washington Post
. I looked at the transcript. During this, David was asked a direct question by someone from Chicago about the weather and turnaround times for the climbers on 10 May. The following is the exact part of the question and David’s reply:
Chicago: How could a cyclone from the Bay of Bengal ‘sneak up’ on these teams with their support?
David Breashears: There was no accurate weather forecast back then.
I also recalled a statement made to me by Michael Groom, about his previous experience of a Met Office forecast. This had been supplied to Harry Taylor by the Met Office after a request from the British Army in 1993, some three years prior to the 1996 disaster:
I first used the information from the British Met Office in 1993 when the British Everest Expedition for that year passed on the forecast to me. It proved to be very accurate, with a predicted weather window of 24 hrs with wind speeds of around 10 kph for 10 May 93. Harry Taylor and I summited as a result.
These two statements relating to the accuracy, indeed to the existence, of weather forecasts for Everest conflicted with David’s answer. It also didn’t make any sense to me that there wouldn’t be accurate forecasts in 1996 if there were such reports in 1993.
Catherine had taken her mother away for a short holiday during the week in which Henrik’s last message came through. The person who’d been closest, the one who’d witnessed the highs and lows, who’d endured both our lives and home being taken over for three years, was denied the opportunity to share with me the sense of relief and vindication that this brought. I had confirmation that Rob and Scott had received a constant stream of information from two weather forecasts and that this had even been updated as they moved up the mountain. By the time Catherine returned on the Friday afternoon of that week, I was already pondering my next problem. I realised, in the final analysis, that my understanding of the events that had led to the tragedy of 10 May now hinged on what the forecast data in 1996 had predicted. Most importantly, I needed to obtain this from the same source that would have been accessed to supply the Danes and the Imax team with their forecasts, and through these two teams to Rob and Scott. I needed to compare this to the actual weather that had struck Everest.
I emailed Mike Harrison, the meteorologist who’d been trying to track down information in the UK Met Office, to let him know that the Danish Meteorological Institute had supplied the forecasts based on data they had received from the European Centre for Medium-Range Weather Forecasts.
Little could have I known what would be in Mike’s response:
‘I have a friend and colleague in the ECMWF. Would you like us to pull all the relevant data from the archives at ECMWF and give you our professional opinion?’
Not one, but two professional opinions. What an offer! ‘Absolutely,’ was my instant reply.
I waited nervously for several weeks. Whatever the data said, I would have to face the truth one way or another. Either our team had been put on a collision course with an incoming storm that Rob and Scott knew was on its way, or I was wrong.
Eventually an email came back from Mike:
‘I’ve spent the last few days at ECMWF and my colleague there had very kindly obtained the relevant data and we’ve gone through it together today.’
I sat motionless as I absorbed every single word of Mike’s report:
We had three sets of information to examine:
1. The ECMWF analysis. This is ECMWF’s best estimate of the observed atmospheric circulation patterns around the world every six hours – this should be an excellent estimate of reality (which is difficult to measure precisely every six hours across the globe!). In this the development of a jet over a rapidly deepening depression over southern Russia and northern India was clear. Winds in the Everest region increased steadily from about the 8th onwards, to reach maximum intensity on the 11th. By the 12th, wind strengths were reducing quite quickly. It is reasonable to assume that cloud and snow accompanied the strong winds at least on the 11th. Temperatures appear to have remained fairly constant throughout this period, so we are unable to say that the storm was accompanied by any major drop in temperatures – my guess is that the snow plus the winds were the main hazards, although a temperature drop cannot be discounted entirely within the blizzards.
2. The ECMWF forecasts. These were available to the Met Office at the time. As I’ve said in earlier emails, I cannot tell whether any forecast that the Met Office might have produced took these into consideration. Forecasts went out to ten days and we examined those starting on the 3rd, 5th and 7th. These forecasts were encouragingly consistent, all indicating the strength of the winds on the 11th quite clearly. All also indicated the steady strengthening of the winds from the 8th onwards.
3. The Met Office forecasts. Unlike the Met Office, ECMWF does have an easily accessible archive of these forecasts, and it is on these that I am 99 per cent certain any service provided by the IFU in 1996 would have been based, either uniquely or primarily. Unfortunately the ECMWF archive includes only a rather limited number of details, and those we were examining with the ECMWF data above were not included. Hence we had to interpret from less direct information. Also, for reasons we do not know, it seems that only a reduced number of values were archived for these forecasts, such that the plots we were examining were rather poor quality as compared to those that would have been available in 1996. Thus we had two problems – unavailability of the data we required ideally plus further interpretation difficulties given the poor plots. The forecasts only went out to six days, so relevant ones were not available on the 3rd, and we examined those starting on the 5th and 7th. In both, despite the difficulties, the storm of the 11th was apparent, although we cannot say how the winds predicted by Met Office compared with those of ECMWF.
Hope that helps you.
Best regards
Mike