by Jay Johansen | Jan 1, 2017
There's one news story that I haven't seen in the global warming debate. That's the news story that begins, "Predictions made by climate scientists 20 years ago closely match what we are seeing today."
The reason why we haven't seen that story, of course, is because the predictions of climate scientists have proven to be wildly inaccurate.
In a speech to the UN Conference on Climate Change in Copenhagen in 2009, Al Gore said that, based on the latest research, "there is 75 percent chance that the entire polar ice cap during some of the summer months could be completely ice free within the next five to seven years." It's been seven years and we still have a polar ice cap.
In 2004, the Climate Research Unit of the University of East Anglia predicted that snow would soon become a rare event in Britain. To highlight the point, they said that within 20 years, the skiing industry in Scotland would be dead because of lack of snow. And indeed ski resorts across Scotland were forced to close in 2013 -- because, according to the BBC, they were "buried under unprecedented amounts of snow", the heaviest snowfall in 70 years.
In 2005, the United Nations Environment Programme (UNEP) produced a report warning that by 2010 there would be 50 million "climate refugees", people fleeing from rising sea levels, hurricanes, and desertification caused by global warming. They even made a map showing where most of these refugees would come from: primarily the Caribbean, Pacific islands, and some coastal areas. The UN has since removed the map from its web site. They explained the map was "causing confusion".
In 1988, NASA scientist John Hansen gave a presentation to Congress in which he said that computer models predicted that by 1997 global temperatures would be 0.45 C warmer. When 1997 rolled around the actual increase was 0.11 C, or less than a quarter of the predicted increase.
Many of these predictions come from the climate scientists' computer models. Whenever a doubter asks for evidence for the apocalyptic predictions, the true believers routinely point to the computer models as proof. Apparently we're supposed to hear the word "computer" and be instantly awed into silence. I can understand how this might have worked in the 1960s, when computers were something new and mysterious and awesome. But this is the 21st century. Many people have computers in their homes. Do people seriously believe that anything you see on a computer screen is unquestionable fact?
I develop computer software for a living. Any computer programmer will tell you that all but the most trivial programs do not work correctly the first time. They always have bugs. Programmers routinely expect to go through a cycle of write program, test, find errors, fix errors, test again, find more errors, fix more errors, test again, find yet more errors, etc. To test a program, we figure out what the right answer is. We then we run the program and compare the program's results to the known right answer. When it doesn't match, we study the program to find what's wrong and fix it. I've had plenty of times that I've spent many hours trying to find the cause of a seemingly simple error, like a heading not being properly centered or the last letter being chopped off a name.
Of course what the "right answer" is depends on the computer system. In an order entry system it might be "customer is billed for $23.72 and a message is sent to the warehouse to ship item #4328920". In a video game it might be "the soldier's pistol runs out of ammo after firing 50 shots". Etc. Sometimes the right answer is simple and obvious. Like if you type your phone number into your profile and then go back to the profile page, the phone number should match what you typed in. Sometimes the right answer is very complicated. Like what will the speed of a rocket be at a specified time in its flight, taking into account the thrust of the engines, the initial weight of the rocket, the change in weight as fuel is burned, gravity, air resistance, and dozens of other factors.
Do you see the problem with a climate model that attempts to predict the weather 20 years in the future? How does the person who wrote the program know what the right answer is, in order to test his program and see if it is working correctly?
If I was writing such a program, I'd say, Let's enter data only through 20 years ago and see if it correctly predicts today. If it doesn't, then you know it's wrong. If it does that doesn't prove it's right: You could tweak the formulas to make the known right answers come out. But at least it would be something. (Have climate scientists tested their models this way? I wonder.)
The programs could be tested and adjusted until they give the answers that the authors want them to give. This may or may not have anything to do with real life. I can easily write a computer program that displays "2+2=5" on the screen. Just yesterday I was playing a video game where the aliens defeated me in a space battle because their anti-matter torpedoes had a range of 10 while my meson blasters had a range of only 5. Do you suppose that the fact that this computer game has these numbers proves that if and when anti-matter torpedoes and meson blasters are invented, that anti-matter torpedoes will have twice the range? Of course not. They're just numbers that the game-maker invented. When they tested the game, they presumably tested that indeed the computer allowed you to hit other spaceships within this range but not beyond. I doubt they invented meson blasters and tested them in a physics lab to compare the results with their computer game.
A climate model is very much like a video game. At best, scientists invent some formulas that describe how they think the climate might work, and they write a program based on these formulas. They have no way to know if those formulas describe the future climate until they make predictions and see if they come true. The reality is that most of their predictions have not come true. Their formulas are wrong. They have been conclusively proven to be wrong. If they seriously want to know the truth, they will modify the formulas until they can accurately predict the past, and then try again to predict the future. If they don't really care about the truth and just want to make apocalyptic warnings for political reasons or so they can get continued funding or whatever, then they'll continue to use the same proven-wrong formulas, and they'll continue to say, "See, the computers models say this is what will happen. The computer said so, it must be true!"
© 2017 by Jay Johansen