It’s always interesting to note the absolute faith people place in the accuracy of computer programs or applications.
If anyone can find a way to prove a non-trivial computer program is error free, the IT world will fall at their feet. It’s even worse when amateur programmers who claim to be climatologists are involved.
But it’s even worse than that. Digital computers, by their very architecture, have finite precision. Unfortunately, chaotic systems cannot be relied on to produce approximately the same results given inputs which only vary by a value beyond the limits of the computer architecture. Generally, this will be a binary architecture, and there is always a value between the last zero/one pair at the digital computer’s limit of precision, which cannot be represented.
A practical example is shown by the efforts to predict the stock market sufficiently well to make money on each transaction, which would be theoretically possible, if the the stockmarket movement are truly chaotic and deterministic.
Alas, the best and the brightest, backed by the finest computer equipment, are unable to do better than throwing darts at a board, assisted by human luck. The stock market movements appear childishly trivial compared with the infinitely complicated movements of the various components of the Earth system.
There’s probably a chance of an afternoon storm here. The BOM, with its greatly superior pool of talent and equipment, agrees. A couple of days ago, they forecast a 10% chance of rain. It didn’t, but then again they didn’t say it would. I figure there’s a fair chance of rain. The Wet season started over a month ago, and we haven’t seen much rain yet, but I’m sure it will come. So is the Met Bureau, apparently, super computer models and all.
Cheers.
