Previous: Prologue
Chapter 1: Models of circulation
In which a request for help is posted, clarifications are asked and a promise is made
On an otherwise unremarkable, ordinary dull day in the late winter of 2027, kagetsuko posted a message on her fediverse account that particularly intrigued me.
She was one of those people I followed to have a nice balanced timeline, not one of my close circle but we were mutuals, regularly boosted and liked one another’s posts and occasionally interacted.
I had long forgotten why I’d started to follow her. I liked her handle because it referenced Aggretsuko, the irritable red panda from the eponymous anime; but it also could be read as ka-getsu-ko, meaning “full moon child”; and as kage tsūkō, which would mean “shadow sailing”. I suspected this was entirely intentional. Her posts were often funny, sometimes flippant, frequently absurd and only occasionally serious. Her profile said “science/chocolate/coffee; languages: Japanese (native), English (fluent)”. No pronouns. Avatar a minor character from the Yuyushiki manga, quite obscure outside of Japan. Ran her own single-user server so she had to be somewhat geeky.
I gathered she was an atmospheric scientist working on climate models. She’d done her PhD in the UK and then returned to Japan, but I had noticed a while ago she kept European hours. That was nothing remarkable, a lot of fedi people had non-standard wake/sleep cycles. And yet it was a little odd, one of those small mysteries that are the spice of life.
I was checking my timeline over coffee and I immediately noticed the post:
@kagetsuko:
Hi fedi, I am looking for some old FORTRAN codes that were developed in 1975 for the Soviet BESM-6 computer by a collaboration of scientists from the US and the USSR. They are simulation codes for a global circulation model. The scientists involved were Nathan Lightman and Bernard Miller at the University of Hawaii/NOAA Joint Climate Research Effort in Honolulu and Alexey Chinchuluun and Gury Marchuk and maybe also Valentin Dymnikov at the Computer Center in Novosibirsk. I’ve gone through all the declassified documents from the CIA Electronic Reading Room, but I haven’t been able to find the codes or any scientific papers or technical reports about this research. I have emailed both institutions but I don’t have any contacts there so I don’t think I’ll get anywhere. Any pointers welcome; please boost!
Now that was very interesting. The BESM-6! The near-mythical, highly influential Soviet super computer of the 1970s. It was what you got when you let an electronic engineering wizard design a computer with inputs from scientists, but cut out programmers and compiler experts from the development. It could solve differential equations at an amazing speed, but simple data movement was embarrassingly slow, because you had to do the pointer arithmetic in floating-point format.
For all the antagonism between the Eastern and Western blocs in those days, the Soviets had quickly adopted the capitalists’ FORTRAN programming language (developed by IBM, no less!) for their super computers. FORTRAN was a marvel, the closest thing to immortality in the computing world: designed before there even were proper computers, and still going strong today; much maligned by “real programmers” but beloved by scientists, and still healthy and under active development. (And for the last 40 years it had been called Fortran.)
I wanted to know more. I replied:
@lores:
@kagetsuko I might be able to help a little, I’m quite familiar with the BESM-6 and ye olde FORTRAN and there are a few code repositories that I’m aware of. What is the context for your question?
Her answer came almost instantaneously:
@kagetsuko:
@lores Thank you so much! I’m doing a postdoc at ECMWF on novel general circulation models to improve prediction of extreme weather events. The Soviets were the first to develop such simulation models. There is a seminal book by Marchuk and Valentin Dymnikov from 1984 but the work had already started in 1974. The problem is that to predict extreme events we need very high resolution, but we also need a global scale. On those old computers, they had to be very efficient, in particular with memory, so I thought I might get inspiration from those old codes. From the letters exchanged between the two groups, they were getting very promising results, but somehow they never published.
The European Centre for Medium-range Weather Forecasts; in Reading, barely a stone’s throw from here. I knew it well. That explained her strange hours. I smiled to myself: a small mystery solved. I reflected that it was a good time to do this kind of historical research: the war between Ukraine and Russia was finally over, and international relations had been normalised quickly. A few years ago, it would have been impossible. But 1975 was a long time ago and pre-internet, so it wasn’t going to be easy.
I stared out of the living room window towards the rain clouds over the slate grey surface of the North Sea, less than a kilometre down the gentle slope on which the house stood. A year ago, we were hit by a hurricane-level storm, and unfortunately the MetOffice’s predictions had underestimated it. It had ripped the roof off and scattered the contents, and what remained had been drenched by the relentless torrential downpour. We’d been homeless for months, and we were amongst the lucky ones: we hadn’t been flooded, and we could rent temporary accommodation while the house was being repaired. With a better forecast, people would have had more time to prepare. It would have made a huge difference.
I was retired, I was intrigued, and it was right up my street, so I rose to the challenge:
@lores:
@kagetsuko I’ll have a look at the CIA docs and the BESM-6 repos
Next: Chapter 2: Détente