How we can predict the weather decades ahead when we can't predict it ten days ahead? Well it is remarkable that we can predict it even one day ahead, and by looking at how the forecasters do that we can begin to understand how the models can work over longer timescales.. When I was a child in England in the 1960s, with our unpredictable weather, a cautious person would take rain gear with them almost no matter what the forecasters said. Even as late as 1987 we had Michael Fish's famous weather blooper. This broadcast is so famous here that it starred in the Olympics 2012 opening ceremony.

"Earlier on today apparently a woman rang the BBC and said she'd heard there was a hurricane on the way. Well if you are watching, don't worry, there isn't. But having said that, actually, the weather will become very windy, but most of the strong winds incidentally will be down over Spain and across into France."

His full weather forecast broadcast is here.

For more on this see this interview with Michael Fish: "Great Storm of 1987: Michael Fish's 'white lie'". The white lie there is that actually nobody rang in to the studio, to say that they had heard there was a hurricane on its way. He just said it as an embellishment to enliven his forecast.

He was on the early shift when he said that. Later on the same day, we had one of the largest storms we've had in the UK in recent times, the UK's "Great Storm of 1987". There were wind speeds of up to 115 mph, many trees blown over and 18 people died. They'd got the strong winds right in the forecast but they thought the worst of the weather would hit Spain, and instead it hit England.

And here he is correcting it 25 years later :)

So, how can that happen, and can it be fixed? By answering this we begin to get an insight into how the longer range predictions work over years and decades as well.

I will look at this first, how they predict the weather one day ahead. Then I'll take a historical perspective on how they came to be confident in the longer range forecasts decades into the future, indeed they make predictions tens of thousands of years into the future too. Many of the questions raised by climate skeptics were major issues for climate modelers in the 1950s through to the 1980s. It may be interesting to read how they eventually found a way through them. Before one can begin to criticize what they say in an informed way, one needs to understand why they are now far more confident in their models for long term climate change prediction than they could be even as recently as the 1980s.

HOW THEY PREDICT THE WEATHER ONE DAY AHEAD

First you might think the way to do it is to constantly improve the resolution of the models and accuracy of your data. Eventually maybe you can simulate individual clouds right down to this level of detail:

Here is an example of a much more complex high resolution simulation running for the weather over Germany with many different cloud types at once.

If you could do that, and it was accurate, then you'd be able to predict every single cloud and you could do predictions as accurate as this scene from "Back to the future II"

Doc: First you've got to get out and change clothes

Marty: Right now, it's pouring rain!

Doc: Wait five more seconds

Right on the Tick. Amazing, absolutely amazing. Too bad the post office isn't as efficient as the weather service.

It's not clear how he knows this (in that fictional world). It could be from previous visits to that time, but the newspaper cover later in the movie shows that they are able to predict weather to the minute at least

There's discussion amongst the fans about how it works in that fictional alternative present - whether it is through accurate weather prediction, or perhaps most likely, it could be through weather modification. See also fan discussion here.

At any rate it's clear we aren't in that future yet. Could that ever happen in reality? Could we predict the weather accurately to the minute? And if not, how can they predict the weather at all?

To answer this let's look at the history of climate models. Many of the issues the climate skeptics raise today were asked by the original modelers in the 50s through to the 80s. By looking at the many discoveries they made along the way we can begin to understand how they gradually developed confidence in their models.

SURPRISING DISCOVERIES ABOUT WEATHER PREDICTION

Originally in the 50s and 60s, weather forecasters thought we would be able to predict the weather as accurately as we liked, just by getting more data, more accurate data, and using very precise models. But then when they tried to follow through that program, they found that it didn't work. The detailed history of what happened is quite complex, see Chaos in the Atmosphere for details. But in summary they found that the weather would often change quite unpredictably no matter how carefully they modeled it. One of the early surprises came in the 1950s when Fermi used an early computer MANIAC at Los Alamos to study what would happen to a system of imperfect springs coupled together (not perfectly elastic) a bit like a sprung mattress. The physicists were sure that the system would eventually stabilize and that is what happened.

“Some of the 2500 tubes used in MANIAC, the Laboratory’s first electronic computer (1952–1957)." photo from Los Alamos shared on Flickr by Marcin Wichary - larger view here

This was the first computer to turn up some of the unpredictability you get from running computer models for a long time.

But one time they decided to run the program for a bit longer than usual and they were surprised to find that they system moved out of its steady state and started fluctuating wildly again. It would go steady for a fair while, then start fluctuating, move to another state, stay there for a while and then fluctuate again.

That was just the first hint that something strange was going on. Later on they got much more of a wake up call with Lorenz's discovery.

Lorenz wrote a program consisting of just three equations to simulate a simple form of "weather". It just modeled convection cells in a uniform layer with different temperatures above and below - a bit like heating a thin layer of water in a saucepan.

The same phenomenon happens in many different situations - in a saucepan, on the sun, in the Earth's mantel. Here is a demonstration of it in a saucepan (used as an example for solar convection).

Rayleigh–Bénard convection cells - these can be modeled by Lorenz's equations

You could hardly imagine a simpler system to model to make a start on weather prediction. Lorenz wrote his program to output just three numbers to describe the state of the complete system at each moment of time.

X = Intensity of convection motion
Y = Difference in temperature between ascending and descending currents
Z = how much temperature varies vertically

He then plotted those three points on a diagram. This video shows his model in action and shows how the fluid flow (at the top) relates to the position of the point in the Lorenz diagram (at the bottom).

To find out about it, see The Lorenz attractor and how it describes flow patterns in a layer of water with videos. And for more maths details see the Wikipedia page and the Wolfram maths page.

The next step was a matter of serendipity - random good fortune. His computer printed out only the first three digits of the six digit numbers to save paper. So he tried putting the numbers for the start of the simulation back in and running the program again, expecting to get the same results as before, pretty much. He was astonished to find that this new run produced completely different results. It turned out that changing just the fourth or fifth or even the sixth digit of the data would lead to completely different outcomes a short while later. Although he could hardly have guessed the far reaching implications of his discovery, this was the first ever inkling anyone had of what is now known as "Mathematical chaos" - the sensitive dependence of a system on initial conditions.
This video shows the result of running the simulation forward with two almost identical points shown in red.

Animation from Wikimedia commons by MicoFilós

The two red dots start off with the same values almost exactly, yet quickly diverge to rapidly different paths. This is known as a "bifurcation". If you start with almost identical points like that, sometimes they will depart from each other almost straight away and sometimes they may stay together for a long time, but eventually they will go their different directions.

The average time the system can run before it bifurcates like that is known as the Lyapanov time. For weather systems then that time period is about 30 days. The weather could bifurcate right away or maybe not for a long time, but on average then it will become impossible to predict exactly after around a month or two.

For more about chaos theory, this is a good intro

And this explains how you can do weather forecasts and also do climate change predictions even though it is a chaotic system.

Dr. Gavin A. Schmidt, Director of the NASA Goddard Institute for Space Studies

COPING WITH CHAOS

The result is that you can't predict the weather exactly over long time periods. But we know how to work with chaos theory. Example, there is a lot of chaos in our solar system.

Saturn's moon Hyperion rotates chaotically. If you have a spacecraft doing a flyby of Hyperion, you can't predict which face you will see. It becomes unpredictable over a timescale of 30 days or so. Yet its orbit is easily predictable, only its rotation is not. It would be easy to do a close flyby to within a few tens of kilometers. But you just wouldn't know which face you will see.

In the same way - you can predict many things about our climate on long time scales. You can predict trivially that the winters will be colder, the summers hotter. Here on the Isle of Mull you can predict that the end of the autumn, through to winter, especially late November onwards, is the season of storms. You can also predict that particular months have most sunshine and also be the driest month of the year and so on.

So, chaos in this mathematical sense is something we can deal with so long as our computers are fast enough. Early computers were so slow you barely had chance to do one run before you had to make a forecast. But nowadays we can do lots of runs, each with slightly different data. Then if there is a bifurcation, then some of the runs will go one way and some the other way. If you do a hundred runs and get rain for thirty of them and dry weather for seventy, then you can say that there's a 30% chance of rain tomorrow.

If that was the only issue, then there wouldn't be much anyone could say to challenge the climate models, not if they understood how they worked. However that's not the only problem

PROBLEMS OF DATA AND RESOLUTION

The main reason they missed the storm in 1987 was because of lack of data. Due to financial cutbacks they no longer had the weather ship Romeo in the Bay of Biscay which is where the storm started. They didn't have the modern weather buoys and their satellite data was also insufficient. No matter how many runs you do, your model can't be accurate if there are significant gaps in the data.

Even with enough data, the models then were coarse in spatial resolution. This became a major issue with the Lothar and Marin storms in 1999. This time the storm was predicted, but the severity of the winds was not. Even two hours before it hit Paris the inland speed were still predicted to be 90 - 130 km/h (56 - 81 mph). The winds actually experienced were 125 - 175 km/h (80 - 110 mph) and brought down 300 high voltage transmission pylons, and left 3.4 million without electricity. The highest speed gusts reached 75 meters per second, or 270 km / h (168 mph), the highest wind speeds ever recorded in Europe (see page 20 here). That was recorded on the Hohentwiel mountain in Southern Germany.

Ruins of Hohentwiel castle photographed from a small plane, credit Peter Stein. This mountain recorded wind speeds in 1999 of 270 km / h (168 mph). The models at the time didn't have sufficient resolution to predict such high speeds, and were not able to take account of details of the topography.

The problem this time was the low resolution of the model. In order to predict the winds correctly they needed to take account of the detailed topography. A later analysis in 2011 came to the conclusion that even a model with a resolution of 16 km, which they didn't have back then, was not sufficient to predict the high winds. Only the highest resolution model they had then, at 2.5 km resolution predicted them correctly. Neither model was available in 1999.

The map on the left shows the forecast for the 16 km resolution model. The highest winds are over 40 m/sec or 144 km/h. The map on the right shows the forecast of the 2.7 km resolution model. There are many spots with speeds greater than 50 m/s (180 km/h). Neither model was available in 1999. This shows that they simply didn't have the models available back in 1999 to predict the intensity of the winds.

So we have several types of sources of error in climate forecasts. We have the chaos theory bifurcations. This can be dealt with by doing many runs of the same model with different starting conditions. Then we have the systemic errors due for instance to insufficient data, or insufficient resolution. You can also get systemic errors because you leave out important factors in the process of making the model. It's impossible to model all the factors that can influence our climate. So which factors are the important ones? What if you have left something out.

CLIMATE CHANGE RANDOM WALKS - 1970s

This is one of the first ideas the climate modelers came up with. What if the climate just randomly wobbles a bit warmer then a bit colder? And maybe if it shifts a bit warmer by chance, then it has more of a tendency to keep going that way? Then the entire process of moving in and out of the ice ages could be a result of random processes. This is one such paper from 1976. This is a later paper from 1990 though by then most climatoligists had moved on to new ideas . Some climate skeptics today say they think climate is just following some kind of a random walk. Well back in the 1970s they would have had good company from many of the climate modelers of their time

One of the best examples of a random walk is "Brownian motion". In this example then the fat globules in milk are being continually buffeted by the molecules of the milk around them. This causes them to jiggle around. Sometimes you may see some of the fat globules moving consistently in one direction for a while, but it is just random and they might well come back again.

So the idea was that the climate of the Earth itself was just moving randomly like that. We can predict it short term they thought, but it gets chaotic beyond a few days - and then beyond that, it is just not possible to predict the climate very exactly. Sometimes it gets a bit warmer, sometimes a bit colder, and if it keeps randomly getting colder for long enough we hit an ice age.

The last three interglacials all lasted for about ten thousand years. So, though we couldn't predict when it could happen and our weather could head towards an ice age at any time, they thought that it would happen at the same time again. Our current Holocene interglacial started 11,700 years ago. So based on that simple argument, it would seem that we are due for another ice age just about now.

If you ask what the contribution is to the Earth's current temperature then only a tiny amount of that is due to human CO2. If there were no CO2 in the atmosphere the average temperature would be -25 C. But it doesn't take much to nudge it a few degrees up or down. If we have business as usual it means around 4 C by 2100. Bear in mind that that is just an average and that the poles are affected far more than the tropics. And that the difference in average temperature since the ice ages to today is only 5 C. So a 4 C rise from now to 2100 is a huge amount really. Not going to make the world uninhabitable. But a lot to adjust to. Nearly as much as adjusting from the ice age to the present in the amount of temperature difference.

ICE SHEETS of up to 3 -4 km thick covered large areas of western Europe and North America, and also the Alps and the Himalayas during the last ice age. The world was only 5 C colder on average than it is today. But the higher latitudes are more strongly affected. Think what a difference it will make if the world becomes 4 C hotter than it is today, nearly as much difference as there was between the last ice age and today!

WHAT DOES IT MEAN TO SAY THE WORLD IS WARMING UP?

The climate is very variable on a year to year level. But if you average it over five years then you can see how it is changing. This plot shows the five year averages from 1880 to 2015. Notice that the Earth changes temperature most in the Arctic region, and that there are a few spots in the middle of the oceans which have changed only by a fraction of a degree. These are actual measured temperatures, not a simulation.

Five-Year Global Temperature Anomalies from 1880 to 2014

CORRELATION WITH THE MILANKOVITCH CYCLES

The first hint of the idea that we might be able to do long range weather forecasts came with the theory of the Milankovitch cycles - the idea that changes in the direction the Earth's axis tilts, over very long periods of time, coupled with other effects, can cause ice ages.

The Earth is in a mildly elliptical orbit so is sometimes closer to the Sun and sometimes further away. At the moment the Earth is closest to the Sun in the first week of January. So the southern hemisphere is tilted towards the sun when Earth is closest to the Sun and the northern hemisphere is tilted away.

That makes the southern summers warmer, and winters colder than they would otherwise be. While in the northern hemisphere then summers are cooler and winters warmer than average. It's been like that for thousands of years. But eventually it will change, and the northern hemisphere will get the more extreme seasons again.

Shows how the position of the pole star varies over a long cycle. Image credit Tau?olunga

Thuban, shown on illustration of Draco Constellation by Torsten Bronger. This was the Egyptian pole star around 5,000 years ago.

So anyway - this obviously changes the seasonal temperatures in the two hemispheres. But when the southern hemisphere is warmer in summer, the northern hemisphere gets a colder summer, so averaged over the year the average temperature doesn't change. How can this cause ice ages? Also we don't get ice ages every 26,000 years. The pattern is far more complex than that.

Well first, there are many other things that change, though that's the most easily obvious of them.

This image from Windows to the Universe shows the three main ways the Earth changes

Orbit gets more circular and then more eccentric, in two cycles which combine together, every 96,000 and 413,000 years

Earth's axis precesses, cycle 26,000 years - this is why the southern hemisphere is pointed towards the sun when we are closest to it. This is a very slow change. But noticeable. At the time of the Ancient Greeks, 1000 BC. the Southern Cross was visible from Athens. And as I said above, the Egyptian pole star was Thuban.
The orbit also slowly precesses, cycle of 112,000 years, not shown here. This combined with the axial precession means that it takes between 20,800 and 29,000 years for the closest point to the sun to return to the same date in the year, with an average period of 23,000 years, not 26,000 years as you might expect.
Earth's axial tilt varies slightly, cycle of 41,000 years.

After the axial precession, the main effect there is the way the orbit gets more circular or more eccentric. At present the Earth's orbit is quite close to circular. That makes this a time of stability according to his theory, because when the orbit is close to circular the other effects on the climate are much less. There's not much that can affect it at all as the Earth is at much the same distance from the Sun in the summer, winter, and all the year round, in both hemispheres. For instance it's no longer likely that we are going to head for an ice age, or have any other dramatic change in climate, in the ordinary course of events.

Also when our orbit is more eccentric then the Earth spends a little less time closer to the Sun because it travels faster when closer to the Sun. These eccentricity changes are the only ones to actually change the total amount of sunlight the Earth gets per year, but they do so only by tiny amounts.

Anyway Milankovitch worked out the total amount of sunlight received at various latitudes. This can vary. For instance at present we get rather less sunlight over the year in the northern latitudes - our summer sun is not quite so bright as it would be because it is further from the sun in summer, and in winter when it gets less sunlight, it is closer to the sun but the higher latitudes have much of the day in darkness. Above the Arctic circle this effect is even more pronounced of course as there you get a dimmer sun in summer, and no sun at all in winter.

So anyway he then found out that the ice ages end at times when we get particularly high levels of sunlight at northerly latitudes.

This is a modern diagram from the NOAA. The warmer interglacials are shown in yellow. The graph at the top in red shows how warm it is in July at 65 degrees North, the modern version of Milankovitch's curve.

There's a strong cycle there every 23,000 years due to the precession of the Earth's axis. Interglacials like the one we are in now happen when the weather is warmer in the northern hemisphere, though not every time. As you see, it has happened every five cycles for the last fifteen cycles.

The bottom line in blue shows the CO2 concentration in the atmosphere which is nowadays thought to play an important role in the warming, though Milankovitch didn't realize that. (Calculated from bubbles of atmosphere in the Dome Fuji ice core)

The middle line shows the temperature in Antarctica (calculated from hydrogen isotope ratios in the Dome Fuji ice core). The temperature in Antarctica is higher at times when the northern hemisphere gets most sunlight, even though at those times Antarctica gets least sunlight, because of global warming.

This theory was first put forward by Milankovitch. Here he is as a student with a pocket watch.

Milankovitch as a student, Vienna, late nineteenth century

This though was not very convincing right away. It is easy to prove the craziest things using graphs and correlations.

Correlations of railway train collisions with US crude oil imports from Norway - from spurious correlations

At best a correlation like that might lead you to consider a hypothesis to investigate further.

Is it a real effect or is it just a coincidence like that? The scientists needed to find out more to check this out.

So, brief history first, his paper on the subject is from 1920. Mathematical Theory of Heat Phenomena Produced by Solar Radiation included modern climate of Mars, Venus. The climatologist Koppen noticed the paper and noticed that some of the significant heat anomalies there matched in timing with the ends of ice ages. The ideas got developed through to his book in 1941. But after that it fell into disrepute. Didn't seem to explain things so well, and his curves didn't match in detail.

But later going on into the 1970s then after more research, more accurate measurements of past temperatures and so on, his theory started to pass all the tests instead of failing them. Then scientists began to sit up and pay attention to it again. Brief history here as powerpoint slides. They still had the puzzle, that the temperature changes seemed far too small to affect the climate so strongly, especially since the temperature averaged out over the entire year would hardly change.

There are two main ideas about how that can happen, or can be a combination of the two. Milankovitch's original idea was that when the upper latitudes get warmer, then ice melts in the summer, this exposes rock. Ice reflects away more heat than rock, and so cools down the Earth. So with the higher latitudes warmer, there is less ice, warms up the Earth, and that leads to less ice in a feedback cycle.

Left shows the extent of permanent ice sheets during an ice age, and right shows them today - just covering Greenland. When the ice sheets are smaller, then less heat gets reflected away so the Earth is a bit warmer. Left shows the Earth 18,000 years ago, and right shows it at present, image credit: Mark McCaffrey NGDC/NOAA

The other main effect is that when the Earth is warmer then the oceans can't absorb so much carbon dioxide, which then increases the amount in the atmosphere, which warms the Earth up further and so on.

Both of these would be very minor effects on their own. Though carbon dioxide is a greenhouse gas, it's not a very powerful one, not compared to methane. But any slight warming of the Earth gets amplified by the effects of water vapour. Where carbon dioxide scores is that when you add a pulse of carbon dioxide, it's a rather one way process, far easier to add than to remove. It stays a long time in the atmosphere. While with water vapour, you can only add so much to the climate at any particular temperature. It will just rain out, within days, if you add more water vapour than the atmosphere of the world as a whole can hold at that temperature and will do that very quickly. So you can't warm the planet up by adding more and more water vapour to the atmosphere. But you can do it by adding more and more carbon dioxide.

This study from 2010 showed in a simple way that water vapour is an important part of the global greenhouse effect. This is an interesting study where the authors ran a climate model, first without the water vapour, then without the CO2.

They showed that the carbon dioxide accounts for about 20% of the greenhouse effect water vapour and clouds, 75% and minor gases and aerosols, 5%.

However the water vapour can't warm the Earth by itself. To test this, they took the model, in steady state representing todays atmosphere, and removed all the non condensing greenhouse gases and aersols, and then ran it forward in time, the greenhouse effect collapsed as the water vapour precipitated out. Earth became a snowball Earth.

It ended up with an average surface temperature of -25 C, after around 20 years of removing all the CO2 from the atmosphere from a starting point of a pre-industrial world with average surface temperature of around 4 C.

More details here: CO2: The Thermostat that Controls Earth's Temperature, by Andrew Lacis

So the CO2 by itself would not be enough to keep Earth as warm as it does. It needs the water vapour as well. The water vapour amplifies the effect of the CO2.

SMALL TEMPERATURE CHANGES MAKE HUGE DIFFERENCES IN THE CLIMATE OF EARTH

Also bear in mind that though we are making only small differences to the concentrations of CO2 in the atmosphere, it does't have to be a huge effect. The difference in warming between the depth of the ice ages when glaciers covered New York and the warmest interglacial periods, is only 5 C (or 9 F). In that context, you can see that an extra warming of 2 C is quite significant even though it seems a small temperature change. A small change in the average temperature of the whole Earth can lead to large local differences in temperature, same also for the oceans, for instance New York will experience much larger changes in sea level than the planet as a whole averaged out.

CO2 LAGS

Also, I can answer your question about CO2 lags. I think you must refer to the studies of ice cores that show that the Earth starts to warm up first, before the CO2 levels increase.

Yes Earth's weather is influenced by many other things, not just CO2. Mainly by its orbital cycles. For instance 20,000 years ago, these cycles triggered warming in the Arctic. This flooded the oceans with fresh water, which disrupts the exchange of heat between hemispheres. This then lead to the southern oceans getting warmer. As a result the solubility of the CO2 in the oceans fell, and that lead to more CO2 getting released into the atmosphere. As a result, the warming effect was amplified.

So there was a small increase in temperature first, then that lead to CO2 release which then lead to more warming, which then lead to more CO2 and so on. In the most recent glaciation at least, the global temperature lags behind the CO2, only the Antarctic temperature precedes / correlates with it.

Graph from CO2 lags temperature - what does it mean?

Average global temperature (blue), Antarctic temperature (red), and atmospheric CO2 concentration (yellow dots).

For more scientific detail on the occasional CO2 lag after temperature rises, see the New Scientist article Climate myths: Ice cores show CO2 increases lag behind temperature rises, disproving the link to global warming And for the paper in detail, Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation

CO2 LAG

I've answered your 2) already, but perhaps a recap will help. The CO2 levels lag behind the temperature of Anatarctica (where the ice core record is from) slightly for the interglacial warming up because it is both an effect and a cause. As the planet warms up due to changes in Earth's orbital cycles, then this melts the Antarctic ice. As a result of the melting ice in the oceans, more CO2 is released which in turn causes more warming which melts more ice which leads to more CO2 released. Something can be both an effect and a cause, the two are not mutually exclusive.

The global warming however lags behind both the CO2 and the Antarctic warming, so the CO2 increases first before the Earth as a whole warms up. (I've just edited my previous comment to make this clearer).

At present the increasing levels of CO2 are the result of humans burning fossil fuels and so on.

There is an effect element still of course, of the global warming impacting on CO2 concentrations, so you are right about that too. As we add more CO2, the world warms up, this makes the Earth's systems less able to take up CO2 just as for the interglacials warmings. That doesn't release more CO2 however because the atmosphere already has more CO2 than it would naturally. In our current situation with an excess of CO2 over what would happen naturally, instead it means that less of the CO2 can be removed from the atmosphere and is the reason why after we add 32 gigatons to the atmosphere each year, 16 tons is removed by natural processes within 50 years, another 8 tons requires an additional 200 years and 4.8 tons is still there thousands of years later, because the warming has made it harder for the Earth to remove it from the atmosphere..

all the people who develop and run the many different climate models think that their models help with predicting future climate? If so, why is it that they think this on your view? This article may perhaps help a bit, on the history of chaos theory and how they moved from early ideas that climate might be completely impossible to predict to the current idea of chaotic fine detail within predictability. Back in the 60s and 70s many scientists did have that view, that the climate could suddenly lurch into an ice age with very little warning and they thought it might happen at any time. When I was in my teens and twenties that was what you read in the science journals about climate change. But then starting with the understanding of how the Milankovitch cycles actually worked, how such tiny effects were able to influence the Earth's climate, they then moved to the current views according to which the climate is far more predictable long term.

Our climate is predictable over such long timescales that they can even calculate when the next ice age will be - with or without CO2 warming - different dates for the two cases - though over shorter timescales you of course still can't predict every heat wave and hurricane. Normally the next ice age would happen 50,000 years into our future following the Milankovitch cycles. But we have already emitted enough CO2 to skip that ice age so the next ice age instead will happen 150,000 years into the future (unless of course we actively remove CO2 from the atmosphere). See Fossil fuel burning 'postponing next ice age' - that page gives a link to the research publishedi n Nature which is available to read in its entirety free online under their sharing initiative. That may seem surprising but the mean half life of CO2 over long timescales is 13,500 years meaning that after 100,000 years there will still be a significant excess of CO2 in our atmosphere, just from the CO2 emissions so far. If we continue business as usual then by the end of the century we could have emitted enough CO2 to end the cycle of ice ages for ever.

This is the most recent article I know, published January of this year [2]. It says that when they selected only the models that most accurately tracked the previous ice ages, using the pre-industrial levels of CO2 of 280 ppm (a bit more than the 240 ppm of the interglacial 800,000 years ago which is the closest model to the present of recent ice ages) they predicted that the next ice ages should be 50,000 and 90,000 years from now (with a possibility of a slowly approaching ice age 20,000 years from now). While with 240 ppm then the next ice age would be as soon as 1500 years into the future. Then, with 500 Gt of emissions, not far off what we have already reached, we may already have enough CO2 in the atmosphere to make a difference to the ice sheets over thousands of years. If it reaches 1000 GT then the chance of an ice age in the next 100,000 years is notably reduced and with 1500 GT of emissions then it is very unlikely that we get an ice age in the next 100,000 years. That's because when you add a new pulse of CO2, it drops off rapidly at first, half gone in 50 years, only a quarter left after 250 years, 15% after thousands of years and long term half life is 35,000 years. You can read it in full under Nature's sharing initiative if you click on the link " published in the journal Nature" in the article in the Guardian here: Fossil fuel burning 'postponing next ice age

CHANGES IN THE ARCTIC

So what's going to happen next to the global temperature with "business as usual"? The most dramatic changes again are in the Arctic: Disappearing Arctic sea ice

And this shows projections for the whole world.

"In the Intergovernmental Panel on Climate Change’s 5th Assessment report, scientists outlined four emission scenarios that serve as ballparks for future climate conditions. In this video, two of these scenarios are compared across the globe. The strictest of the four emission scenarios (left) gives an average global temperature increase of 1°C by 2100; the laxest of the four emission scenarios (right) gives an average global temperature increase of 3.7°C. The scenario on the right shows the strongest simulated change for the Arctic, which is predicted to heat by more than 10 degrees. In both scenarios, temperature increases are more pronounced over land than over the oceans, as water doesn’t heat as quickly and works as a balancing (cooling) agent. "

This is the greening he talks about:

Rise in CO2 has 'greened Planet Earth'

Yes and they also have done several successful long range predictions already. Prediction of present climate in the 1970s and 1980s. Prediction of the exact amount by which the eruption of Mount Pinatubo would cool down the climate and for how long. Prediction of La Ninas six months before the event. And most impressive, detailed predictions of the ice ages based on understanding the Milankovitch cycles - originally seemed far too small an effect to do anything but once they understood the interactions of Antarctic ice, CO2, water vapour etc they could then model them exactly and work out exactly why each glaciation happened. I'll do a list of successful long range predictions for the next article, research into it a bit more.

Longer term using the global warming scenarios, you can predict that if there is more CO2 in the atmosphere, there will be more heat waves, more wild fires, more drought, more heavy rain also, more intense hurricanes and so on.

So - there is no reason why you can't have long term predictions. The question then is - what can you predict and how? So for the how, what they do is they run the model with many different slightly different starting conditions. Lots of runs. Just a bit hotter here, colder there, a bit more wind here, a bit more humidity etc on the day the run starts from. Just vary and tweak those parameters. Then see what happens. Because of the chaos inherent in climate then even a month or two after it starts, the different runs will predict very different outcomes. Maybe one has a severe hurricane in August, another run has it in September, another one doesn't have any severe hurricanes that year. But now you run it for many years, for a decade, say. If all of the models with extra CO2 have more of the severe hurricanes and fewer of the mild ones, and the ones with less CO2 have fewer severe hurricanes and more mild ones, then you have a lot of confidence that adding CO2 to the model leads to more severe hurricanes and fewer mild ones.

They then also will use different grid sizes, in case it is anything to do with the model itself, and different teams will run different models based on working through the problem independently without reference to each other with different assumptions. If those all agree also then they begin to have a lot of confidence. And of course they then validate it using past series to predict the present.

It's the same with all the predictions. This is how it is done.

Until you understand this basic point, you can't hope to understand climate modeling and why and how they do it. And so until you understand this, you don't know enough to criticize them.

Because of chaos theory, it's impossible to predict weather exactly over long periods of time, if you mean the details of every rain shower and thunderstorm, tornado or even hurricane. But over short periods of time it is impossible to predict every cloud that will appear in the sky next day. You can say it will be say 20% sky covered in clouds tomorrow but you can't predict every cumulus cloud even for the next day. We never will be able to do that in every detail. In the same way for longer range forecasts you can say things like "there is an increased probability of wild fires in the US because the weather is hotter on average and there is less rain". But you can't predict every fire.

Yes of course choatic solutions diverge. With the Hyperion example, because its rotation is chaotic, then the solution diverges so quickly that if you plan a flyby a couple of months after your last observation of it close up, you will not have any idea which area of it you will see facing towards you. But you know exactly where it will be in space. Of course the climate modelers know about chaos theory! They model it. Their models of course exhibit chaotic sensitivity on initial conditions. So then they deal with those diverging solutions by doing lots of runs of the model with slightly different start conditions. In that way they can map out the solution space.

Then if they find the same features throughout that solution space - in all of those diverging solutions - then that's what they predict for the future. In the same way that all the solutions for Hyperion have it in the same orbit and same position in its orbit for long periods of time into the future.

So, yes, weather is chaotic, so you can't predict the clouds for tomorrow exactly, you couldn't say "tomorrow you will see a cloud that looks like a swan", say - that is way way beyond what we could do with climate modeling. Nor can you predict hurricanes a year in the future. But that doesn't mean you can't predict anything, just because it is a chaotic system. You can still predict that tomorrow will be cloudy, and you can still predict that hurricanes will on average be more intense over the next decade, if that is what the models show. All that is perfectly consistent with chaos theory.

"Chaotic" in this mathematical sense doesn't mean "totally chaotic so nothing can be predicted". It always means "some things can't be predicted and other things can" so then the job of the scientist is to find out what things you can and what things you can't predict in a chaotic system.

Does that make sense? Please think it over carefully. If you don't understand this point you don't have any chance at all of understanding why people do climate models and why they work.

Notice also - those arguments involving a simple line graph - those are climate models too. They are just very crude ones that treat the whole Earth as if it was a single point so zero dimensional. You get far more accurate results by treating it as a three dimensional complex world with many different interacting systems.

97% CONSENSUS

Debunking 97% Consensus Denial - they may be the origin of the 97% figure and they refer to multiple studies coming up with the same figure and they explain how they asked the scientists with an SAE to confirm their assessment of their abstracts or to change them if they were incorrect.

This is what the authors say:

"Nevertheless, the existence of the expert consensus on human-caused global warming is a reality, as is clear from an examination of the full body of evidence. For example, Naomi Oreskes found no rejections of the consensus in a survey of 928 abstracts performed in 2004. Doran&Zimmerman (2009) found a 97% consensus among scientists actively publishing climate research. Anderegg et al. (2010) reviewed publicly signed declarations supporting or rejecting human-caused global warming, and again found over 97% consensus among climate experts. Cook et al. (2013) found the same 97% result through a survey of over 12,000 climate abstracts from peer-reviewed journals, as well as from over 2,000 scientist author self-ratings, among abstracts and papers taking a position on the causes of global warming.

"In addition to these studies, we have the National Academies of Science from 33 different countries all endorsing the consensus. Dozens of scientific organizations have endorsed the consensus on human-caused global warming. Only one has ever rejected the consensus - the American Association of Petroleum Geologists - and even they shifted to a neutral position when members threatened to not renew their memberships due to its position of climate denial.

"In short, the 97% consensus on human-caused global warming is a robust result, found using several different methods in various studies over the past decade. It really shouldn't be a surprise at this point, and denying it is, well, denial."

Do read what they say and follow through and read their rebuttals of the denials. If you just read what climate skeptics say and don't read the rebuttals you have an unbalanced picture, obviously. As I said in the article also, it may well be that the consensus is much more than 97% because with some of these studies they don't assess articles if they don't say either way what the view of the authors is, which would normally mean they accept the consensus just as for articles on continental drift or the meteorite impact theory of origin of the lunar craters.

the problem is that if we want to stay within 1.5 C then according to the IPCC then the total amount of CO2 we have to emit before 2050 is equal to the amount we would emit in four years of business as usual. It is not certainty but they make it a two thirds probability that we hit 1.5 C by 2100 if we put as much CO2 into the atmosphere by 2050 as we would after business as usual for four more years.

This is the article, that the IPCC worked out various scenarios, published November 2014, and to remain within 1.5 C then we have to produce no more than 6 more years of emissions at the current levels before 2050..It's not an absolute. Depends what we do after that. This is from 2014 and the conclusion is " suggests we have just six years of business-as-usual emissions before the budget giving us a two-thirds chance of staying below 1.5 degrees is exhausted." so that takes us to the end of 2020. So obviously we have to reduce drastically before 2020 to have a chance of keeping the total CO2 emissions up to 2050 the same as 4+ more years of business as usual (now that it is 2016). I.e. not just by 2020 but well before 2020 we have to reduce emissions drastically to have a decent chance of remaining within 1.5 C by 2100. The higher targets of 2 C are more easily achievable. That's why the poor countries are aiming for 0% emissions so quickly. They want to reach 0% emissions between 2030 and 2050. We should all try to follow suit if we want to stay within 1.5 C, to get as close as possible to zero emissions by then, since we have left it so late, and we need to reduce emissions quickly, by substantial amounts already, over the next four years..

That's why it is so urgent, if the climate models are correct. Why we can't just sit around and discuss it but if we are going to act on it, have to act quickly, if we want to remain within 1.5 C

http://www.carbonbrief.org/six-years-worth-of-current...

It is because of projections like that that they feel we have to act quickly. And for many of us things like the Arctic melting, the unusual warm winters here in the UK for several years now, heat waves in the summer retreating glaciers world wide, coral bleaching in Australia - we don't need to wait any more for it to become more obvious that it is going to happen. And the reason for all that is because CO2 is a very long lived gas. Not exponential decay. Of the 32 billion tons we put into the atmosphere this year of CO2, 16 billlion will still be there 50 years from now. But 8 billion will still be there 250 years from now and 5 billion tons will be there pretty much indefinitely, for thousands of years. Indeed because of the very long residence time of some of the CO2, we have already introduced so much CO2 into the atmosphere that the climatologists predicted in a recent paper that this means the next ice age predicted for 50,000 yeras from now will not happen, unless we remove CO2 from the atmosphere we will skip that one and the next one instead will be 150,000 yeas from now.

Of course that's not necessarily a bad thing. But it shows how our actions over just one century have very long term effects. So that also makes it easier to see why we have to act quickly to prevent a 1.5 C rise even though the rise itself wouldn't happen until around 2100.

The half life of CO2 when you take into account the very long residence time of the small part that remains - ove rthose long time periods is 13,500 years. It is due to the way that adding CO2 reduces the ability of the Earth to absorb excess CO2.

That's why every country including the US agreed to the Paris agreement. It looks as if Trump may possibly be rethinking his plans to withdraw from it. If he does withdraw he will be a very isolated figure on the world stage, the only nation not in the agreement.

Well the thing is that now especially now that we can see so many effects of climate change for ourselves - the people saying there is no need to do anything about it seem less and less plausible. She is not a climate change denier, she agrees that it will happen but says that there is more uncertainty than other climatologists think. Yes it does need to be debated. The thing is that in the rest of the world that debate has already happened and we've moved on to action, and there were many actions we could have done, but this was the decision embodied in the Paris agreement. While in the US politicians are still debating it.

This image is being shared on social media as "Politicians Discussing Global Warming." - although that's not the original name of this work, originally called "Follow the Leaders" It was an installation by the sculptor Isaac Cordal in Berlin in 2011. The new title is more powerful though. At some point we have to stop debating and act. The rest of the world have decided they have had enough debate and it is time to act already.

. I think that perhaps once Trump is president, it might be that it becomes more and more difficult to maintain his climate skeptic stance once he finds that he is the only world leader saying this - he would feel like an outsider in a party wherever he goes on the world stage. Also with many cities and states in the US also. Plus as president he will have to give words of reassurance after hurricanes, droughts, wild fires etc, at least I imagine so?? So that may make them more real to him than reading about them and watching them on TV?? Just a thought. Anyway if he doesn't withdraw from the Paris agreement that's a start.

the US has been very strong in clean energy. Still your per capita CO2 emissions are very high. To stop global warming we all have to reduce our CO2 emissions. And the US as one of the top emitters and one of the higher per capita emitters can do a lot by reducing the CO2 emissions. And yes the US is good on clean energy - so that makes it more surprising that you aren't leading in climate change measures. Well of course you have been leading. It's Kerry who was mainly instrumental in getting the Paris Agreement together. But with Trump it seems the US is going to renounce this lead and go into a back seat and indeed get out of the car and leave everyone else to drive on without them.Kerry on the climate change agreement http://www.bbc.co.uk/news/science-environment-38005101

Sadly though it seems that the Trump administration may dismantle much of the NASA research into climate change, changing the US from a leader into the field into a side act.
https://mic.com/articles/160299/donald-trump-plans-to-cut-nasa-s-politic...

CLIMATE SKEPTICS ARGUMENTS

noaa Please read this to understand why they make those adjustments. No climate conspiracy: NOAA temperature adjustments bring data closer to pristine

VOLCANIC ERUPTION MYTH

The first is the myth that a volcano emits as much CO2 as all humans produce in a single year. In actuality humans emit about 32 billion tons of CO2 a year. All the volcanoes put together emit on average 200 million tons each year - estimate in this Scientific American article. So a normal single volcanic eruption is nowhere near the amount of CO2 that humans produce in a year, even all the volcanic eruptions for an entire year produce only 0.625% of the total CO2 added by humans for that year. The volcanic eruption contributions of course are in steady state part of the slow carbon cycle which helps to keep the Earth warm enough for life (otherwise it would be -25 C) and also fertile (plants would have no CO2 to fix if the volcanoes stopped returning CO2 to the atmosphere).

On volcanoes, then they are of course random and unpredictable. But they have produced just the right amount of CO2, millennia after millennia for millions of years to keep Earth at around its present temperature. They don't suddenly all stop for a long time and plunge us into a snowball Earth, nor do they produce huge amounts of CO2 in a big belch. It's because the CO2 is the result of subduction of limestone and chalk and so on as continents move beneath other ones and that is an on going continual process.

As for a supervolcano - then it actually reduces the amount of CO2 in the atmosphere. That paradoxical result is because the dust from the volcano, carried around the world, cools the world down. A cooler world can absorb more CO2 so for a few years the Earth is absorbing more CO2 than usual and so the concentration in the atmosphere decreases rather than increases. . In one study, the researchers

"...simulated Pinatubo-like eruptions (Volcanic Explosivity Index VEI 6) every 5 years in one model run and Tambora like eruptions (VEI 7) every 25 years in another model run for the period 2000 to 2100 where CO2 emissions followthe IPCC-A2 scenario. They demonstrate that the smaller but more frequent eruptions have a larger impact on the carboncycle than the less frequent larger eruptions. At the end of the integration in 2100, the atmospheric pCO2 is 46 ppmv lower than in a control experiment without volcanic forcing."

METHANE EMISSIONS

I'll also cover the Clathrate Gun hypothesis which is often used by climate skeptics in the other direction to say that the climate is going to get unbearably hot no matter what we do so there is no point in doing anything. But the timescale for methane release from the sea floor in a rapidly warming climate is a thousand years, not decades, which rather puts that to rest. As for methane from permafrost - then as it warms up, then the methanogens that love to much on methane flourish and eat the methane so very little of it will actually reach the atmosphere.

Also of course methane has a very short half life in the atmosphere, about a decade, some say a bit more, some a bit less, depending how they calculate it.

BOGUS "SCIENTIFIC CONSENSUS" CLAIMS

Just a warning to watch out for them. This is an example from Forbes magazine from 2013. Claims that two thirds of engineers and geoscientists are climate deniers. But when you read the small print it turns out to be a survey of 1077 self selected subscribers to the PEG Magazine, published by the e "Association of Professional Engineers, Geologists, and Geophysicists of Alberta". One of the big industries in Alberta is the petrochemical industry and the report itself contains a long discussion of awareness of climate change amongst petrochemical engineers.

So actually it was saying that in this small sample from Alberta, two thirds of geochemists and engineers, many of whom who work in the petrochemical industry are climate deniers. That a third accept global warming is the more impressive figure there I think. And that is three years ago. Now many in the petrochemical industry are gearing up for transition to clean energy and finding new ways ahead for their companies, so the figures have probably changed since then.

CONFLICT WITH COAL, OIL AND GAS

Actually CO2 reduction doesn't have to mean direct conflict with coal, oil and gas. Probably the best way to deal with the situation involves transition to a mix that includes a fair bit of renewable energy. But actually, we can continue using fossil fuels as well so long as we use carbon capture and storage technology.

You can burn coal while hugely reducing emissions with carbon capture and storage. It has a lot of potential to help reduce CO2 during transition to renewables, when we are still burning coal, oil and gas.

Our government in the UK short sightedly (in my view) pulled the rug from under a $1 billion CCS competition just days before the Paris agreement

What's more, later on if we use the same technology with biomass, then it is one of the few carbon dioxide negative power generation technologies. Actually takes CO2 out of the atmosphere to grow plants, and then burn them and capture the CO2 emissions from burning them.

What is Carbon Capture and Storage

This is about the Western Antarctic ice shelf. The IPCC has already said that the sea levels will rise by three feet by 2100 if we don't use stronger methods to limit greenhouse gases. But what happens after that in the 22nd century?

In these new studies, a team of glaciologists using satellite and air measurements say that the ice in Western Antarctica has already started a process that is probably impossible to stop. With ice penetrating satellite radar mapping of the terrain beneath the ice (using the EU Sentinel 1 satellites), they say that here are no mountains or hills significant enough to slow the collapse. The fastest melting glacier, Smith glacier, is losing 70 meters thickness of ice a year. It's grounding line - the point at which it starts to float on the sea - is retreating two kilometers a year and has been doing that since 2011, is continuing unabated.

There are six glaciers that will collapse, enough to raise the sea level another four feet. But these may collapse other glacier leading to a rise of sea levels triple that. A separate team studying just one of the glaciers, Thwaite glacier, came to the same conclusion that collapse is inevitable. That is, will happen anyway, based on the CO2 emissions so far.

If so then this would cause a 10 foot rise in sea level. This would cause issues for coastal cities like New York and low lying countries like the Netherlands and Bangladesh which is the area in the world likely to be most affected by sea level rise since much of the country is not far above sea level.

They spotted a new rift which may lead for a large ice sheet to break off again, like the giant 225 square mile "iceberg" of 2015.

Rift in Pine Island Glacier ice shelf, Credit NASA/Nathan Kurtz.

They say there that many think it is inevitable that the Western Antarctic ice sheet will disappear. The main Antarctic ice sheet is still growing. Antarctica has been growing steadily all through the glacial and interglacial periods which is why we have these long ice cores to look at the temperature changes in Antarctica in exquisite detail. If we continue "business as usual" it will stop growing and eventually thousands of years into the future, melt completely. But that is not a risk at present. The risk is just from the Western ice sheet. But because the ice rests on land, rather than on the sea, adding this to the oceans will raise the height of the oceans, and they estimate by 10 feet so about 3 meters.

So what effect will it have if they are right? Not end of civilization. But some major issues. Florida is amongst the most affected since the underlying geology is porous limestone. This means it will be impossible to build conventional flood barriers as the sea will just percolate through the rock beneath them. So it seems inevitable that Florida will be flooded if sea levels rise. Only mitigation possible as a way ahead.

Here it is as it is now.

And after a 3 meter rise

It is already getting affected more than usual by hurricanes because of the one foot the sea has risen by so far. See Goodbye Miami for an article in Rolling Stone magazine about these issues.

In the US, then New York is also impacted and other coastal areas. You can have a look with this global sea level rise map, though be aware - that it just maps a new sea level against the topography of google maps. Places that are inland below sea level of course will not experience sea level rises - you need to trace out to see if there is a connection with the sea. But in the case of Florida, since the underlying rock is porous, then the map above probably pretty much shows what the effect would be no matter what flood defenses are used.

World wide then naturally the Netherlands are amongst the most affected. As a rich country they would be able to increase the height of their flood defenses but it would be expensive. Other coastal areas in France, Belgium, Denmark and the UK (e.g. the Norfolk broads) would also be affected.

Here is the area as it is now:

And after a 3 meter rise. Remember only places that are shown as blue and connected to the sea would actually be flooded. Unless the geology is porous, inland areas below sea level would not be affected, and the Netherlands particularly can be expected to build better flood defenses, though a 3 meter increase in height of them would be an expensive undertaking.

Then finally, this shows the effect for Bangladesh of a 3 meter rise. This is likely to be an overwhelming humanitarian issue for a poor country with a huge population and they would surely need external help to deal with the issues.

Bangladesh as it is now

And after a 3 meter sea level rise such as might happen by 2100 if the western Antarctic ice sheet melts

So, no, a ten foot rise in sea level will not mean the end of civilization at all, but major problems for several particularly vulnerable spots world wide.

They just say land temperatures there, it's not about ocean temperatures. And it is just fluctuations in the temperature of the surface of the ocean. El Nino brings warmer water back to the surface, doesn't warm up the Earth as a whole. You get a better picture from the global heat reserves in the land + ocean.

See Global cooling - Is global warming still happening?

The problem is not so much with the temperature of Earth over the next few years but what will happen by 2050 and then by 2100.