Farewell To Long-term Climate Change Predictions
Trump administration orders government agency to stop predicting long-term climate change impacts, The Independent
“The Trump administration has told a major US government department to end predicting what the long-term effects of climate change will be on the country. Director of the US Geological Survey (USGS) James Reilly – a White House-appointed former oil geologist – ordered that scientific assessments only use computer-generated models that track the possible impact of climate change until 2040, according to The New York Times. Previously the USGS modelled effects until the end of the century, the second half of which is likely to see the most dramatic impacts of global warming. The order is likely to impact the US government’s National Climate Assessment, an interagency report produced every four years which outlines the projected impact of climate change in every corner of US society.”
Keith’s note: NASA is one of the agencies participating in The National Climate Assessment. It is only a matter of time before these edicts affect NASA. Alas, NASA will continue to study long term climate trends on other planets – just not the one we live on.
We all knew this was going to happen.
What is the accuracy?
OK, I see what you’re doing there…
Here’s the thing about predicting weather and climate: both are subsets of Chaos Theory.
Don’t worry about wrapping your head around this idea, which is kinda simple, actually. A chaotic system is hugely determined by input values. As time progresses, very small changes in initial values over time result in dramatically different outcomes. Some point to the Butterfly Effect as an illustration.
Here’s the kernel that matters: the more one can characterize input values in a chaotic system, the more accurately one can predict future values of that system. We already know the direction or trend of the climate system. What we need is more accurate data on input values.
Next question?
This data is not what I would call “chaotic” by any means (including the current deviation). And surface temperatures follow suit. Solar cycles, obliquity cycles and many others are predictable and periodic across millennia. Yes, there are many variables such as continental position and drift, etc., but those are on scales are so far beyond average human consciousness/experience that they cant even be considered. You imply an undamped oscillatory system that does not regularly cycle (i.e. self-correct) over time; and yes, the only way to make models predictive and pragmatic is to collect more and more data over time. Models attempt to explain current behavior and evolution over time and are based on direct observations, and that is all we have, that is what science really is.
https://uploads.disquscdn.c…
The climate system is not chaotic. Weather is chaotic on sub-seasonal time scales. The warming due to Greenhouse gases can be accurately modeled, and largely explains (among other things) the variation in surface temperatures among Venus, Earth, and Mars. The original calculation was done in the late 19th century by Arrhenius and hasn’t changed much since then.
Except that the effects of global warming _are_ those smaller scale things. Will hurricanes or tornados become more common? Droughts and floods? Melting of ice sheets falls in that category as well. And, when you talk about a triggering effect if warming exceeds two or three or however many degrees, there is feedback from those smaller scale processes. Which does, actually, drive up the uncertainties in the long term models.
And even ignoring that, those governing equations for climate (not weather) have large, nonlinear terms in them. That can mean chaotic regimes exist, although we aren’t in one. It does mean that, if you really want to estimate the uncertainties in predictions, it’s much more complicated that the usual statistics 101 approach most people use. And the actual uncertainty is higher than that approach suggests.
I suggest you do your own research and stop depending on copying others homework; but here you go to start: https://www.sciencealert.co…
Exactly! How accurate are these long term forecasts? With weather being an inherently chaotic system, the predictions would vary wildly using even slightly different initial conditions. Kind of the definition of chaotic systems.
As a statistician, I know such systems are not what you’d want to make any long term predictions about. Or even mid term predictions.
It’s ridiculous. Climate modeling is still in its infancy.
As was stated above, the macro changes we see due to climate change are not at all chaotic. And as stated above, those macro changes due to climate can impact weather, which is chaotic. Climate change changes the inputs to weather. For example, hotter oceans (climate) will provide more energy to hurricanes (weather). So long term changes to climate can be used to predict some (long term average) changes in weather.
We’re already starting to see some of the effects of climate change. We’re predicting the long term trends here. How accurate they are isn’t like the accuracy of a weather forecast predicting the percentage chance of rain in Cincinnati tomorrow.
I’m actually curious if anyone has done any sort of Monte Carlo error analysis on these models. It’s hugely expensive in terms of computer time, but I’ve found it to be the best way to propagate measurement errors through the whole process. Basically, you take the input measurements and make random changes to them. Changes at the level of measurement the errors. That gives you a simulated input data set, which is what the measurements would have been if the random measurement uncertainties had gone a differently. Then you run that simulated input data through the whole model, and see how different the results are. Then you do that again, many times (I’ve found 100 times to be about right) and use the statistics on the results to determine the uncertainty. That does mean running the whole model a hundred times, so it’s inefficient. But it does turn up subtle things which more conventional techniques miss. It also makes it very easy to see where there are correlated uncertainties (which is most of the time and a significant factor with a nonlinear system.)
James Reilly is a former astronaut, He should have resigned.
Yet another data point that being a former astronaut isn’t necessarily a qualification for much else.
Might as well pass a law that says the earth is the center of the universe and everything revolves around it.
No, Trump is the center of his universe and everything revolves around his ego. This is such a dumb decision and path to follow. Complete hubris, ego and narcissism all bound up by greed. This country has earned and deserves exactly what it gets.
I wonder how our Administrator will respond to such a order by Trump? Bridenstine supposably is one of few Republicans who does not call climate change a hoax. Besides political ramifications, climate study of earth also provides experience in sensor data and comparing trends to predictions. Compare what was observed in space to what is measured here on the surface. Then able to apply that to other planets.
Nobody is telling NASA to cease long term climate models. The order is don’t use long term models to for impact studies. The problem is these models are extremely unreliable.
“The problem is these models are extremely unreliable.” No they are not. Any small inaccuracies are because the models are too conservative and underestimate measured and real climate change. The documentation for this is so overwhelming, I will not try to summarize. Repeating a lie over and over again – even when well financed by the fossil fuel industries – does not make it true.
One of the main consumers of the impact studies is DOD. For good reason. It would seem that they have determined that climate change is likely to have significant national security implications, and I tend to listen to experts that tell me something is a national security concern. It doesn’t matter to them whether coastal flooding in XYZ sets up a food crisis in 2050 or 2070 – what matters is that such a food crisis COULD happen and what that entails.
Bridenstine is definitely not inclined to push back against Trump, otherwise he would have rejected the Moon 2024 initiative. But I think he only said he believed in climate change to get past the nomination process.
Reilly is a veteran of three space shuttle missions – STS 89 , 104, and 117. He’s been up there , above the atmosphere , when any fool looking out a spacecraft window can see the white shadow and fluid substance of Climate Change, right before his eyes. Pollution and smog trains, wildfires, unusually large cyclonic storm systems, receded glaciers, algal blooms, dust storms measured in hundreds of miles… a hundred visual waypoints. He knows the science, he knows the scientists, he knows the planet.
Reilly therefore is judged to be disingenuous and a political puppet to a false dogma perped by the Trump cabal. He needs to step down, and hopefully have an epiphany and speak out once removed from government service which he allowed to compromise hisself. Otherwise , geophysical redemption will not be his, but intellectual gangrene might.