Air pollution was pretty bad in the United States from the 1950s through the late 1980s, and we know that https://en.wikipedia.org/wiki/Global_dimming has kept temperatures down a bit.
So basically from the early 1970s through the 1990s in the United States, the amount of 'global dimming' type air pollution decreased tremendously, because of a number of factors, including all sorts of clean air mandates and regulation.
I don't know how sound the theory is, but i remember listening to the climate podcast Radio Ecoshock where a scientist talked about aerosols in general and that there could be a negative feedback loop that would disappear if we stop polluting heavily, ironically then raising temperatures significantly - this could mean that an extremely agressive decrease towards 2030 could increase temperature weirdly enough.
The effect was seen over the US after 9/11 as far as i remember.
> The effect was seen over the US after 9/11 as far as i remember.
There was a study that claimed that but there's a second study that claims general cloud cover was a confounding factor not controlled for that explains the actual result:
Even with a full set of airplanes in the air contrails are a tiny minority of the cloud cover. Seems pretty hard for them to have any visible impact. Aerosols in general are another matter all together. Those are all over the atmosphere and it wouldn't be surprising if they had a large effect as they accumulate.
Related to cloud cover: one of the biggest uncertainties in climate models is how to cater for the albedo effect also known as cloud feedback. It goes something like this:
1) The most potent greenhouse gas is water vapor (not CO2)
2) Increased temperatures result in more water vapor (simple evaporation)
3) Water vapor forms clouds
4) Clouds are white and reflect incoming solar radiation
A lot of models have to make big assumptions about the strength of this negative feedback loop, and the assumptions have a big impact on the output of the models.
Conversely, further increased evaporation > even more water vapor > increased precipitation > fewer clouds.
Also, clouds reflect sunlight > less warming, but clouds reflect IR > less cooling.
It’s a really complex interaction that doesn’t have a clear global maximum as each region has a slightly different impact and they all interact. Atmosphere dust is one of the biggest drivers here and dust alone involves a lot of non linear feedback loops.
"further increased evaporation > even more water vapor > increased precipitation > fewer clouds." No. You can't get fewer clouds from more water vapor, that's not how equilibria work.
Clouds are dependent upon the particulate matter in the air, which is reduced by precipitation. You still get cloud formation without that, but the air ends up very supersaturated.
My best guess is some natural cycle tied to the https://en.wikipedia.org/wiki/Atlantic_multidecadal_oscillat... is involved. You'll note that the 1960s correspond to the point where the Atlantic was cooling most rapidly, which would be suggestive that the air above the Atlantic was relatively cold at that time.
That said I'm not a climatologist, nor do I play on on TV.
It's worth pointing out that there likely is no such thing as the AMO. The scientist who coined the term in some pioneering work in 2000 has, over the past few years, been extremely outspoken that it was probably an errant interpretation of the data available at the time (e.g. https://michaelmann.net/content/rise-and-fall-atlantic-multi...) and even just a year ago published an article in Nature which more or less thoroughly eliminates the possibility of such a mode of multi-decadal variability actually existing (https://www.nature.com/articles/s41467-019-13823-w).
Whether or not the variation is a predictable cycle notwithstanding, there is a substantial natural variation, and the 1960s coincided with a drop in Atlantic sea temperatures of the right order of magnitude. Which suggests that natural variation could have caused the 1960s to be unusually cool relative to overall warming trends.
Interesting to note that in the 1970's the talk of another ice age was really more of a conversation about a rapidly changing planet. The real anxiety was over pollution and its effect on the climate - not that the planet was cooling naturally or going through a cooling phase:
As they review the bizarre and unpredictable weather pattern of the past several years, a growing number of scientists are beginning to suspect that many seemingly contradictory meteorological fluctuations are actually part of a global climatic upheaval.
They thought it was due to pollution (man made) even back then:
Man, too, may be somewhat responsible for the cooling trend. The University of Wisconsin's Reid A. Bryson and other climatologists suggest that dust and other particles released into the atmosphere as a result of farming and fuel burning may be blocking more and more sunlight from reaching and heating the surface of the earth.
Interesting to note they speculate the trend might be temporary - which makes sense because as we've reduced pollution by leaps and bounds since then, you've started to see more of a warming trend:
Climatic Balance. Some scientists like Donald Oilman, chief of the National Weather Service's long-range-prediction group, think that the cooling trend may be only temporary. But all agree that vastly more information is needed about the major influences on the earth's climate.
This article was never an indictment on the scientists believing we were heading for another ice age. Quite the opposite. They started seeing rapid changes in the planet's climate and wanted to try and explain what the consequences were. They even say they need more data and more research to figure out exactly what's going on.
Note that the set of thermometers and temperatures that make it into the GHCN comprise repeated convenience samples. The techniques used to smear individual observations over vast swathes of territory tend to leave much to be desired.
Here's an animation[1] I made way back when I thought maybe thinking scientifically about how to evaluate the existing climate record for the recent past was acceptable.
I have not touched the GHCN recently nor have I updated the analyses, but it did not take much time to produce the videos many years ago. Anyone curious should be able to repeat the same thing of plotting coordinates of temperature stations with fresh data with no trouble.
So, for the original question, the answer is we do not know. However, data are subject to selection bias and that affects what kind of inferences you can make.
But do compare the distribution of observation sources in the 60s to periods before and after.
>The techniques used to smear individual observations over vast swathes of territory tend to leave much to be desired.
I am by no means an expert, but after farting around with climate data, even today I'm curious how they smear data across comparatively smaller swathes of territory.
I've noticed, in my area (south west), that many of the stations are located in seemingly unique microclimates--like the tops of mountains, on the shores of rivers, etc.
It roadblocked my layman research because distance to a given station was not very indicative of how that area's climate is.
In Step 5: 8000 subboxes are combined into 80 boxes, and ocean data is
combined with land data; boxes are combined into latitudinal zones
(including hemispheric and global zones); annual and seasonal anomalies
are computed from monthly anomalies.
> In 2007, David Jones' team, Climate Code Foundation, reprogrammed the whole procedure in python,
as a product of the Clear Climate Code project and the Climate Code Foundation,
http://clearclimatecode.org/
That's all the source control you'll ever get. Don't bother clicking on that link.
I wonder how much warming and change in precipitation is related to greening of the country from additional CO2 in the atmosphere and resulting changes in the albedo.
Is 0.16 F increase in average temperature per decade a legitimate cause for concern? From my (unlearned) perspective, It seems barely statistically significant.
Yes, because we're talking about a time frame of 200 years. That amounts to 3.2F, and yeah, that's a lot.
It's enough to make a lot of marginally-livable places completely unlivable. It's enough to throw off growing seasons, costing a significant amount of agriculture. It will require more water, already a tough squeeze in a lot of places. It will melt a lot of ice, causing disruptions to sea level that will make expensive storms more frequent. It will throw off rain patterns, meaning that some places will become arable that weren't, but forcing some existing farms to stop. That will be hugely disruptive and expensive.
The real problem is that it likely won't just be 3.2F. That's actually pretty much what the IPCC's target is: they want it to be just 3.2F (2C). But even that much involves drastically cutting CO2 output: CO2 is very stable and will last for many centuries. We will continue to get warmer for decades even if we stopped burning fossil fuels utterly, right this instant.
Instead, the world as a whole has continued to increase its CO2. (The US has actually gone back down to 1990 levels, largely because of a shift from coal to natural gas, but partly due to renewables.) That means we're on a track for more like 6F to 8F by 2100, and that is pretty clearly a lot.
It's misleading, since it sounds like everywhere is just going to be a tiny bit warmer, basically unnoticeable. However, it's really about the massive amount of energy required to warm the atmosphere up by that much, and how that energy isn't equally distributed.
One analogy I use is: imagine sitting in a pool with friends, and one person starts pushing one of those floating lounge chairs up and down. They might only increase the average height of the water by a millimeter or so, but the whole pool starts getting more wavy and chaotic. That's how it is with our atmosphere - more extreme events (both "hot" and "cold") occur and in general more chaotic weather as that energy is spread in waves and troughs throughout the world.
Great way I had it explained to me was that if we imagine weather follows some kind of normal distribution called climate, then a shift of of the distribution makes those extreme weather events far more likely much quicker than it affects the mean weather.
However much temperature is increasing, I find the frequency of extreme weather events far more problematic. How many 100/500/1000 year storms has the United States dealt with over the past few years?
Even more to the point, is it possible that the rate at which these extreme weather events is increasing is also increasing?
Without getting into the weeds about particular numbers, it's known that given a global temperature increase, the poles experience the majority of that change. One reason the United States has experienced so much extreme weather of late is because the northern hemisphere polar jet stream has gotten a lot more erratic, in large part because of a declining temperature delta between the north pole and its adjacent temperature bands.
Effect size isn’t what statistical significance refers to. It’s only in reference to whether or not the measured effect appears to differ from zero. The effect itself may be minuscule to the point where it’s practically inconsequential in real life.
Right, the comment surely should have said "substantively significant" and not statistically significant, since statistical significance is a product of not just effect size but also sample size and estimator efficiency (and, indeed, chosen alpha -- anything can be significant if you make the alpha large enough)
Thanks for clearing things up. How would we know if measured climate change is statistically significant? It seems difficult to measure these sorts of cyclical long-term events.
There are a couple of ways to interpret your question, one is purely statistical “is the climate/temperature actually changing?” And the other is “does it matter?”. I’m not really equipped to answer the second one properly.
Re the first question, statistical hypothesis testing requires a null state of the world that you want to test, in this case “is the global temperature changing”. What that requires, more rigorously, is some sort of assumed probability distribution that generates the data we observe. We would probably assume a normal distribution/bell curve with mean equal to the long term average temperature during the last few thousand years, or whatever period is relevant (again, im not a climate scientist), and variance deduced the same way.
To test that null hypothesis: “the temperature over the last decade isn’t meaningfully different from the long term average”, you then take your recent data, say the last and ask “what is the probability we’d see this data, assuming the state of the world we assume given our null hypothesis”. If the observed data is sufficiently unlikely to be generated by that null distribution, you say there is a statistically significant difference.
Now, That’s not the only way to analyze the data (putting it into two buckets and asking if they differ). But I don’t want to get too into the weeds without better understanding what your wondering about/where the disconnect is.
Edit: looking back on this answer, there are some glaring issues that need to be addressed, because they’re both bad analysis and you hinted at it in your question.
Obviously the issue here is that cyclical data clumps together so to speak. Sot he temporal binning i outlined here is a big issue, because you are going to get data with lower variance that you would expect.
This could conceivably be adjusted for by raising your significance threshold, though off the top of my head I’m not 100% sure how I would approach it personally.
Thanks for the reply. It seems like a pretty complicated subject, and I'm sure I'd need to learn more of climate science if I want to feel good about interpreting data like this.
Part of the issue for me is that I’m attempting to wrangle statistical significance into the discussion, when I haven’t properly stepped back and frames the issue properly. Like you I need to know more about the science as well in order to give a helpful answer.
Suffice to say that statistical significance is only relevant when there is a specific statistical parameter that you want to make a determination about. And in this case I’m not entirely sure what that would be.
Looking specifically at the value you referenced, I think a more useful tool you could look at would be trend decomposition.
Your comment may sound unrelated - but it's exactly at the heart of the problem.
Both climate change and viral growth increase exponentially. Not only do we release more CO2 into the atmosphere every year, but also the growth rate accelerates.
No, the effect of a linear increase of CO2 on temperature is logarithmic. Though the speculation is the there will be catastrophic exponential feedback mechanisms.
I guess CO2 is increasing exponentially, so maybe CO2-dependent temperature change is a somewhat linear with respect to time.
From what I understand it is feedback loops that cause concern for exponential climate change. For example, snow reflects a large amount of the sun’s energy back into space. As you lose the snow cover you absorb more heat causing more snow to melt, leading to snow melting faster, etc. etc.
Another example is in the Arctic tundra. The top layer of the ground was always frozen, i.e. a permafrost. However now that layer is melting and it releases a ton of methane. Methane causes the temperatures to rise, leading to the permafrost melting faster and more methane being released.
> Why does climate change increase exponentially? For reasons other than our collective output increasing exponentially?
I'm a decades long 'climate change alarmist'.
We really don't know if the temperature is really increasing exponentially, considering the proper definition of exponential: "a value increases in proportion to its current value". Also consider that "exponential growth" doesn't necessarily mean "extremely fast"...at least in reasonable time horizons.
I and many others think that the amount of energy in our planet's atmosphere is increasing, and we also believe the rate of increase is increasing. This might be technically 'exponential', but it very well might not. It could 'just' be quadratic...which, given the 'right' constants, could be even worse.
> For reasons other than our collective output increasing exponentially?
In short, all chaotic but meta-stable systems, such as our planet's climate, have built in many 'tipping points', where a small change in input can have a very large change in behaviour. Under ideal circumstances, such tipping points are quite hard to identify with any precision, and virtually impossible with the most complex system in the world: earth's climate.
Even shorter: we don't know for certain, but there's a large and ever growing pile of data that shows here and now impacts of a warming climate. Per my other comment, for example: the rate at which 100, 500 and 1000 year extreme weather events are occurring.
In summary: us not knowing with certainty does NOT mean we should not make it a top priority to start reversing the things we do know are causing damage: enormous releases of CO2 and CH4.
For example, warming the climate tends to melt ice. Ice reflects sunlight back to space better than liquid water or dirt. So the reduced ice cover makes more sunlight get absorbed by the Earth. Absorbed sunlight becomes heat, which raises the temperature. The increased temperature melts more ice.
Technically all of these loops can only result in a sigmoid curve, not an exponential. The temperature won't become arbitrarily large. Trivially it can't get any hotter than the input (the sun). But in the ranges we care about (temperatures suitable for human habitation without massive adaptation and migration needed) it's effectively exponential.
Yes, that's what I mean when I say it's effectively exponential in the ranges we care about. It's effectively indistinguishable from exponential growth (locally).
This is an excellent question. It is a very popular idea that there is some "tipping point", plus positive feedback loops, that are definitely going to make this problem worse, someday. That they are popularly imagined by non-scientists, and that the effects are always promised for the future rather than identified in the past, are cause for skepticism.
Judging by the fact that for the billion years after life has dominated, Earth climate has always stayed decidedly non-Martian and non-Venusian, the evidence is strong that negative feedback loops are more powerful than the positive ones that people may publicize. I would like to see more substantiation behind such assertions.
It's not the average temperature change that is significant, its the effect of the temperature that is significant. For example, 33F vs 32F is very significant... If there are more days above freezing, which leads to extra ice melt, it doesn't matter if the average is up only a tiny bit.
I think some kind of error bars would have been very enlightening, yes.
I find it quite hard to evaluate what is being shown, when the total change in temperature over a century is smaller than the temperature difference between different parts of the room I'm currently sitting in.
Yes. The effects of increasing average temperature are highly localized. It's not that at every place in the world, every day is 0.16 F hotter, but that in some places, some days are much, much hotter.
CO2 also has an impact on human cognitive function, and acidifies the ocean.
Air pollution - usually linked to CO2 emissions - has a whole slew of other problems.
We absolutely have the technology to get rid of coal and oil for transport and electricity production - we need to start banning the production of new ICE cars and new coal power plants now, and tariffing any nations who don't follow.
Since precipitation that falls as snow is released much later in the spring but precipitation that falls as rain causes additional melting and immediate runoff exactly where the freezing level is can have a dramatic effect on glacial melt And how dry forests get in the summer, with the ensuing forest fires.
The change is accelerating. Projections that held true in the 80s and 90s are being revised. We’re talking about 6-12 degrees Celsius warmer on Average globally in 100 years. https://en.m.wikipedia.org/wiki/Climate_change#/media/File%3... That means some days/months will be significantly warmer and our crops, food supply, water supplies and other industries won’t have adapted quickly enough to survive.
The worst case scenarios are the breakdown of civilisation and human extinction.
It is enough to change weather patterns, raise sea levels significantly, its a pace of temperature change that causes extinction events, when it has happened this fast in the past.
See this comment a lot, and it's wrong on so many levels.
1. No, this is not a "cyclical perturbation", unless your cycles are long enough to include mega impact events like the ones that killed the dinosaurs[1]. The last time methane levels were this high, it was at least 800,000 years ago. The last time CO2 levels were this high was 3 million years ago. Guess what. The Earth was 2 to 3 degrees Celsius hotter back then than now. That means that parts of the Earth that are currently inhabited by hundreds of millions to billions of people will become uninhabitable.
2. Humans have never existed under the climactic regime that we are unleashing. Humans evolved over the last 400,000 years through a period of glaciation in a little cuccoon in Africa.
3. Civilization has never existed under any climactic regime other than the warm inter-glacial period from about 12,000 years ago when agriculture was first developed. Before that, it was a looong ice age that almost killed humans off completely. Before us, Earth was cyclical between glacial and interglacial periods about every 100,000 years. That cycle is now pretty much obliterated by what we did in the past 150 years.
4. Changes in CO2 concentration change have generally been very slow over Earth's history, on the order of thousands or tens of thousands of years. Such an abrupt CO2 concentration change is generally not seen short of massive volcanic activity or a planet-killer asteroid like the one that killed the dinosaurs and the majority of animal species 65 million years ago. Abrupt climate change means mass extinction for life on Earth. Buckle up, lots of shit is going to die and whole regions will turn to desert. Earth probably will not support our food production systems. Lots of us are going to die, too.
5. There has never been anything remotely like a CO2 "cycle" in Earth's climate history. There is no mechanism nor natural variation that causes CO2 to just go up and down in cycles that are predictable. If you want to educate yourself with some Science, here's a reconstruction of Co2 for the past 550 million years: http://earthguide.ucsd.edu/virtualmuseum/climatechange2/07_1...
If you can spot the cycle, I'm sure they'll award you an honorary doctorate in planetary Science, and you'll be hella famous.
[1] It's a stretch to call meteor impacts cyclical; they follow a power law distribution and are random enough that they hit about as regularly as rolling the dice. When the universe hits a hard 9, we take a beating.
> The last time methane levels were this high, it was at least 800,000 years ago
The climate went through an ice age for most of the past 1 million years. As you point out in #3, an ice age glacial is infinitely worse for life on earth than mild warming. The climate prior to the Quarternary period was better for life in almost all respects. The ice age was brutal, caused mass extinctions, and nearly decimated humans as well. It also froze up so much water that virtually all of Canada and most of Russia was buried under a mile thick ice sheet. So of course, sea level was 400 meters lower than present. Much of that melting occurred with the onset of the Holocene, which has been the prevailing climate regime for only the past 12k years or so. At some point during the ice age, CO2 levels also dropped dangerously low, to 180ppm, a near record low for life's entire existence on this planet. Below 150ppm, plants die entirely. Their productivity and growth is hampered by both freezing temps and low CO2. It was so low plants could not grow at all above certain elevations that are covered in alpines today.
CO2 is not dangerous for life on earth. It is a fundamental nutrient for plant life. The optimal periods for life on this planet have all been marked by an abundance of CO2. The only mild risk from mild warming is rising sea levels. That has been documented to be holding steady at 1 foot of sea level rise per century for the past few hundred years. Much smaller than the meltwater pulse at the Younger Dryas 12k years ago. Humans have continually delt with sea level rise since we invented agriculture. It's not the doomsday event it is painted to be and the trend has not accelerated in some sort of feedback loop as was claimed 30 years ago. The trends are linear, not exponential.
I wish I could say with certainty that we have disrupted the 100K year glacial cycle. That would be incredibly good news.
Unfortunately, volcanoes have a way of ruining the party. There is absolutely no evidence that a volcanic cycle cannot kick in at any moment and plunge earth in a continual regime of thick dust and particulates that increase condensation, reduce solar irradiance, and cause temps to drop by 5-10 degrees globally for the next 100K years.
The rate at which you post wildly inaccurate facts far exceeds my ability and patience for rebutting them. But this utter gem stands out above the rest:
> The only mild risk from mild warming is rising sea levels. That has been documented to be holding steady at 1 foot of sea level rise per century for the past few hundred years. Much smaller than the meltwater pulse at the Younger Dryas 12k years ago. Humans have continually delt[sic] with sea level rise since we invented agriculture.
That is just completely wrong. Global sea levels have been documented and reconstructed and cross-checked better than radio-carbon dating. There are still extant coastal cities that have existed for more than a thousand years, and they are all in agreement that this claim of 1 foot per century just cannot possibly true. I literally have no idea where you got this, I've never seen anyone even suggest this number. Even documented sea level rise since 1900 has been only 5 to 8 inches. Did sea level rise drastically slow down the last century? It's funny, that number, because since the 1990s the incredibly accurate modern gauges of sea level have noticed that the rate of rise has actually tripled to 3 millimeters a year and is still accelerating! That's funny, that's roughly a foot a century. So you just happen to suggest that the exact, current rate of sea level rise is the average over the "past couple centuries". Uh, no. If we were to project that 1000 years into the past, sea level would have been 10 ft lower! Corals and sediments don't lie. Sea level has been stable for 2000 years, just the perfect amount of time for modern civilization to up and build megacities on coasts the world over. Now, project that 3 millimeters per year into the future. In 1000 years, 10 feet! Goodbye all of Florida, Manhattan, and pretty much every major coastal city. But it won't be 1000 years. Things are picking up. We'll get multiple feet of sea level rise this century.
So one thing to keep in mind, showing 20 year averages every 10 years and comparing them to century long averages for the same time period is a really good idea, it shows that there is a trend and it can show shorter temporal anomalies.
One thing I found very interesting is the information on precipitation. I've heard it said "a warmer world is a wetter world" and that would appear to be supported by this data.
We don't have a century of homogenous temperature data. We didn't have weather satellites until the late 70s, early 80s. And temps were only recorded in some cities and some towns, with uneven coverage, with inconsistent global quality practices. Sea temps were also inconsistently recorded.
So... we have a half century of homogenous high quality temperature then. Does the good data we do have not support the same conclusion? Seems like it does to me.
If you want numbers that sound large, use a scale that's numerically larger for the same dimension.
The interval shown goes from -1 degree F to +1 degree F.
1.8 degrees F = 1 degree C. In other words, the same scale in C would go from -0.56 C to +0.56 C. In other words, the entire range is about 1 C.
The other issue is that the color representation of the maps grotesquely exaggerate the magnitude. The visual impact from deep blue to dark brown is far greater than the 1 C range it represents. Such color range would be more accurately used to represent the delta in temperature between, say, deep winter at the cold end and a very hot summer at the hot end. I other words, it visually represents a difference that could easily be 50 C, rather than 1 C.
An honest scale would represent the range of temperatures (minima and maxima) across the US from winter to summer in a scale from, say, -30 F to +120 F and present maps colored using this scale. The problem with doing this is that this +/- 1 F delta across 120 years would be just-about imperceptible.
In addition to this, the image has no information whatsoever on what the color coding truly represents. One legend reads "Difference from average" while another says "30 year normal compared to 1901-2000".
If you look at the website, they have graphs that compare the "global temperature anomaly compared to 1901-2000". In other words, they average some temperatures (undisclosed) over that period of time, this becomes "zero" and everything is compared to this artificial zero. The temperature over this period averages out to some number and everything before a certain year looks colder than this average while everything past a certain point looks hotter.
Huh? What am I actually looking at?
Where is the map/data with absolute measurements, not this weird averaging over a period that stops twenty years ago. Why does the reference average stop in 2000, twenty one years ago and does not include the last twenty years?
Do these measurements correspond to exactly the same date and location each year? Average of what? When? How? What was the tolerance, error, reliability of the measurements? They are comparing measurements made 120 years ago on a scale with a full range of 1 C.
Let's put these numbers into context:
Hot day in Las Vegas: 48 degrees C.
Let's say we have a measurement error of +/-1%. What would that range be? Well, that's + or - 0.48 degrees C from true temperature.
And this graph is showing a range of +/- 0.56 degrees C over a period of 120 years? An measurement error of 1% across such a vast number of measurement sites, with different instruments, people and systems recording them isn't only plausible, it is very likely. In fact the error could be significantly greater than that.
Read a thermometer, not a digital one, a mercury thermometer of the kind they used to use for much of the last century. Can you read that thermometer to an accuracy/reliability of 1%? Likely not. The thing would have to be massively large for you to be able to visually interpolate that scale to such accuracy. Where are you reading it? Top of the meniscus? Bottom? Middle? How did the guy in New York read it? The one in Dallas? How was it read in years 1, 5, 10, 20, 30, etc.? Same instrument for over a century? Likely not. How were they calibrated, if at all? Who calibrated them? How often? To what standard? Etc.
I don't have a problem for making a case for environmental awareness and more responsible behavior. However, proponents need to understand that extreme honesty and accuracy in every statement being made is of paramount importance. Graphically distorted and numerically exaggerated graphs such as this one only serve to feed deniers and those who aim to increase their numbers. In other words, the reaction could be "I got you! You are lying to me!". Not a very useful outcome.
Either we are doing science or marketing. I propose only one of those will result in changing hearts and minds. Don't lie. Don't manipulate. Be accurate, precise, accountable, transparent and backup your data.
> And this graph is showing a range of +/- 0.56 degrees C over a period of 120 years? An measurement error of 1% across such a vast number of measurement sites, with different instruments, people and systems recording them isn't only plausible, it is very likely. In fact the error could be significantly greater than that.
I think we should be careful to judge metrics from other fields based on our own technical experience. There is no basis for a non-climate expert to assume whether +/- 0.56 deg C over 120 years is a big deal or not, or within margins of error or not. To think that climate scientists as a whole don't know about instrument calibration, measurement-biases, or statistical noise filtering is silly. This is like thinking scientists measure sea-level rise by putting a measuring stick in the ocean.
This has nothing to do with being a climate scientist. In the real world, being able to compare field measurements across an entire country going back over a century to a small fraction of a degree C is almost impossible.
Why are you so eager to accept the data without questioning its validity and demanding proof?
Note that I am NOT denying climate change at all. Deniers are fools. It’s very real. Zealots also fall into the fallacy of believing everything they are told, because “it’s science”. Well, science requires proof and verification. If someone is going to claim to be able to make these kinds of comparisons, they better be ready to support their claim with evidence. Believing anything anyone says isn’t science, it’s a cult.
Show me where, in that entire website, they talk about and provide data on instruments and methods and you might have a point. If you can’t find this information, why do you believe their claims? Based on what? Because they are scientists? Since when is that sufficien proof?
> 1.8 degrees F = 1 degree C. In other words, the same scale in C would go from -0.56 C to +0.56 C. In other words, the entire range is about 1 C.
F is the standard measurement that most people in the U.S. understand intuitively.
> An honest scale would represent the range of temperatures (minima and maxima) across the US from winter to summer in a scale from, say, -30 F to +120 F and present maps colored using this scale. The problem with doing this is that this +/- 1 F delta across 120 years would be just-about imperceptible.
Why would you want to do that with this data set? The point of the data is to show deltas in temperature over a long period of time. It's not a weather map to help people figure out whether they should wear a t-shirt today or not.
> Where is the map/data with absolute measurements, not this weird averaging over a period that stops twenty years ago. Why does the reference average stop in 2000, twenty one years ago and does not include the last twenty years?
Again, if you want to show deltas, you have to chose some fixed point and normalize against it. This is common in all sorts of data visualizations, e.g. charts of inflation that choose some arbitrary point in the past as '100'.
> And this graph is showing a range of +/- 0.56 degrees C over a period of 120 years? An measurement error of 1% across such a vast number of measurement sites, with different instruments, people and systems recording them isn't only plausible, it is very likely. In fact the error could be significantly greater than that.
With a large # of measurements over a large period of time, you would expect cancelling errors, not a secular trend in one direction over time.
It also doesn't require fractional precision in individual measurements to create fractional differences in averages. e.g. if I have one set of people that are 50, 51, 52, and 53 inches tall, and another who are 51, 51, 52, and 53 inches tall, I can say that the second set is on average 0.25 inches taller than the first.
> Do these measurements correspond to exactly the same date and location each year? Average of what? When? How? What was the tolerance, error, reliability of the measurements? They are comparing measurements made 120 years ago on a scale with a full range of 1 C.
You're basically asking for a full explication of decades of work to build a global temperature record data set. It's not feasible to recapitulate that in every discussion of global temperature changes. To get an idea of how the temperature record was constructed, you could start with Paul Edwards, A Vast Machine. (edit: here's a fairly succinct summary of how the raw data is adjusted - https://www.carbonbrief.org/explainer-how-data-adjustments-a...)
> if I have one set of people that are 50, 51, 52, and 53 inches tall, and another who are 51, 51, 52, and 53 inches tall, I can say that the second set is on average 0.25 inches taller than the first.
Thanks for making my case with a very simple example.
You can say that. Of course. However, the issue here is about whether or not what you are saying has any real meaning.
I'll use the second data set as an example of how what this paper is reporting might raise questions.
51, 51, 52, 53
minimum: 51
maximum: 53
median: 51.5
mean: 51.75
partial mean: 51 (only considering the first two samples)
standard deviation: 0.96
Let's look at how we might be able to report the differences between these four measurements.
First, the fact that the measurements were reported as "51", "52" and "53" implies the experimental data could not be recorded to an accuracy of more than one inch. If the measurements were more accurate we would have numbers like "51.1". Whatever the case may be, a measurement like "51" does not imply an infinite number of significant digits. This is important.
Here the difference between "accuracy" and "precision", which are different things, might be worth reviewing:
Accuracy: How far a value is from the actual value
Precision: How close measurements are to each other
You can have incredibly good precision and still be way off. To use a war analogy, you intended to land ten missiles within a 1 km radius of city A. Your missiles landed within a 1 km radius, however, they landed in city B. This was very precise, yet highly inaccurate. The other end of the scale would be that all missiles land in city A, yet they land within 10 km of each other, accurate, not precise.
And so, when discussing experimental data and using it to generate conclusion the nature, scale and kind of errors in the data must be accounted for. Without this information you might think your mission is successful when you actually bombed the wrong city. In other words, unless you have the full picture the conclusions are invalid, or, at the very least could be questioned and deemed to be worthy of skepticism.
Back to your data (which has no error information at all).
How can we report the differences?
Difference from minimum: 0, 0, 1, 2
Difference from maximum: -2, -2, -1, 0
Difference from median: -0.5, -0.5, 0.5, 1.5
Difference from partial mean: 0, 0, 1, 2
Difference from standard deviation: -0.78, -0.78, 0.26, 1.3
So, if I look at these results and I intend to show as dramatic a change as possible, I restrict my mean calculation to the first two samples and produce a graph showing 0, 0, 1, 2. In other words, the increase in the samples outside the partial average looks more dramatic.
This map (and possibly the entire website) uses this weird partial mean that LEAVES OUT the last twenty years and then compares the entire period to that partial mean. Huh?
To make things worse, we don't even know what kind of mean they used. For example, it could have been a weighted average calculation and, yes, this could be designed to add even more dramatism to the charts.
Looking at the standard deviation is likely to be a far better indicator of where each sample lies. Again, without error data some or all of this is meaningless. Why?
Let's say the measurement error is +/- 1 inch. Why? Hair, being in a hurry, using a tape measure and no reliable way to actually measure to the top of the skull, etc.
I don't feel like doing a bunch of statistical analysis this morning. I'll just say that, once you include such an error in the calculations you can end-up with results that anyone could consider to be "normal" and certainly within one standard deviation from the mean. In other words, no news at all.
I mentioned "significant digits" above and said this is important. This is an often misunderstood reality of calculations, particularly when it comes to real-world experimental data.
Say I machine a thousand pieces of metal to a diameter of 1.000 inches. It's a pin that is supposed to go into a hole with a diameter of 1.001 inches. After manufacturing half of them come out to 1.005 and the other half 0.995 inches. The average is 1.000 in. What does this mean? NOTHING. Why? Because fully half of the pins I machined will not go into the hole because they are too big! That's 50% waste due to the error. The average is meaningless.
If I report this as a difference with respect to the mean of a partial population the numbers that will come out of that will be utterly meaningless. Useless. The correct specification is likely to be a measure acceptable standard deviations of the pins from the standard deviation of the population of achievable hole diameters.
Back to significant digits.
You can't fabricate data where it doesn't exist.
If you have three numbers, [10.5, 9.3, 17.3] you can calculate the average to be 12.3666... However, if these numbers came from experimental data, you can't claim you have a valid result to six or more significant digits. You just can't. In this case your best hope for a meaningful average is 12. Not 12.3, 12. Two significant digits. This is limited by the value with the lowest number of significant digits.
I won't go into the details on significant digits. Here's a short document from Yale on the subject and how it applies to calculation:
And, BTW, this takes on a much more important meaning when you are dealing with laboratory or experimental data. Dry theoretical calculations must follow these rules. Once the values being manipulated become connected to the real world this become far more important and significant, because findings must have real meaning. Without that you might assume 100% of the pins you machined are good, when, in reality, you have to throw away (or fix) 500 of them.
Here's a good set of videos on the subject from Khan Academy:
From the video: "to ensure"..."that the result isn't more precise than the things you actually measured".
If you are not familiar with this kind of analysis, please watch the videos. This speaks directly to the issues in the kind of reporting we see on this subject all the time. We can't fabricate data and not expect to be challenged.
As I said in a different comment. This is NOT to say I am one of these delusional anti-climate change folks. The effect is very real. What I am saying here is that if we are going to gain the hearts and minds of those who might not understand or want to accept what's going on we have to make sure we don't leave easy holes in the claims we are making. The "and then a miracle occurs" kind of math and claims you see all the time in climate change reporting provides very easy attack vectors for anyone wishing to discredit the entire discipline.
In my day to day work, if you showed me a noisy graph of a periodic signal measured over some 1/10th to 1/10000th of a typical cycle period, I would tell you that you don't have nearly enough data to establish a trend. Especially when the underlying phenomena are chaotic, complex, impossible to model accurately, and incompletely understood.
I also believe that the orthodoxy is overly negative. Who's to say that earth won't be better off with the newly arable northern latitudes? The OP seems to indicate a trend toward increasing rainfall as well - and plants thrive in increased CO2 environments.
Then there is the supposed looming socioeconomic catastrophy as sea levels rise and displace coastal inhabitants. That also sounds like positive economic churn to me, especially if you go by the metric of GDP. This will be a gradual migration over 50-100 years and potentially a boon for various economies - jobs for infrastructure and construction.
I'm just not sold on the idea that the world is ending (literally or figuratively) if we don't act.
You seem to be imagining that the sea level will smoothly rise mm by mm over a century and that people will gradually and peacefully abandon the shoreline and move inland.
It seems to me that a more likely outcome is that people will build defenses to protect existing cities (which are after all massive investments) and and that those defenses will work well right up until they catastrophically fail in a major storm (e.g. see Katrina).
How is it that you can predict the defenses are likely to fail catastrophically but the people building them won't? Is your concern really that other people don't know as much as you about how to build sea walls?
I didn't actually say it was likely, just more likely than a gradual upslope migration in advance of any danger, which seems contrary to all human social behavior that I've ever witnessed. When major storms come through and wreck places that are bound to be wrecked every time the water rises, the political drumbeat is to rebuild. To give up on those places would be to show weakness (See "Jersey Strong / Stronger than the Storm" slogan after Hurricane Sandy).
If we could somehow build strong enough defenses to protect all the major coastal population centers from rising waters indefinitely, that would be great, though obviously very expensive. I'm doubtful though. Lots of money and effort was poured into protecting New Orleans, and, in the long run, it didn't work.
So it really is that you believe people will make the wrong decisions in adapting to the change while you personally know what's right? It makes sense that the general population isn't as smart as you personally because any common ideas they hold are limited to those ideas that are acceptable to a wide range of people. But we do use experts to help make these kinds of decisions for us. It's not all direct democracy with people voting on individual engineering calculations instead of voting to hire an engineer.
It my country, we had an earthquake a few years ago which slightly increased the flood risk of a suburb. Instead of building a wall, they forced everyone out and flattened the whole place. I think we're capable of writing off areas of land when they become too risky or expensive to keep using. There are actually some small towns which are expected to be killed off when the only road access to them is inevitably blocked by a major earthquake one-day because it may not be worth rebuilding the road.
Oh yes, I heard about that. There will surely be mistakes but I don't see how it will be a general problem people will keep shooting themselves in the foot with an never learning from. How many times can you rebuild in the same spot before you realize you're wasting money?
Surely when you say "This will be a gradual migration over 50-100 years" you are imagining a process by which this will happen.
You didn't ask me any questions, you asked nerdponx, and I chose to join the conversation. What I posted above was my answer to "Is it wishful thinking though? How exactly?"
But given the way are responding, I'm not sure you're interested in having an honest conversation.
> It seems to me that a more likely outcome is that people will build defenses to protect existing cities
Honestly, this part right here is what makes me very skeptical about the magnitude of the problem as it’s communicated to the proles. We’ve heard countless times how Manhattan will essentially be underwater in XX years, a foregone conclusion, yet no one has made a serious proposal to build sea walls around it or other valuable coastal areas.
That and the fact that the US is obsessed with importing countless numbers of foreign citizens who doubtless would produce far lower carbon footprints in their home countries than they would in the US. I mean it would only make sense to limit third world immigration to first world countries as long as possible if we’re concerned about carbon footprint, wouldn’t it? The population decrease in first world countries would be an added bonus as far as climate change goes. Yet we do the complete opposite. Curious!
Throws stuff out of equilibrium that our fragile human world depends on. Could stack with unleashing the trapped methane gases in Siberia, to form a positive feedback loop.
Your response is so sophomoric, but not funny. You clearly don’t understand the tight relations of human systems to the environmental ecosystems.
Have you read a single page of any IPCC report or any major climate science publication?
The rhetoric you're spewing is uninformed and dangerous. Even if you are exactly right (and well-founded scientific arguments would suggest you are definitely not), you are speaking completely out of turn without any evidentiary basis.
Like, "plants like CO2 so more CO2 might be good" is naive, elementary school reasoning. You might as well say "cardio exercise is good for the heart, so maybe medically inducing tachycardia is good for me". It's a complete non-sequitur with no empirical grounding.
Like, "plants like CO2 so more CO2 might be good" is naive, elementary school reasoning. You might as well say "cardio exercise is good for the heart, so maybe medically inducing tachycardia is good for me". It's a complete non-sequitur with no empirical grounding.
Somewhat off-topic, but this is exactly the kind of reasoning I see promulgated by a lot of conspiracy theorists, and more broadly by people who refuse to let go of outdated policies and ideas about the world. There is a certain arrogance inherent in the belief that you can understand anything just by thinking about it.
And climate denialists, who generally fall back on their weak intuitions. It'd be overgenerous to compare them to Newtonian physicists failing to grasp Relativity--the former was at least a respectable, experimentally verifiable theory that one could get to with basic calculus and the straight lines our minds draw for us, and not an inaccessible mathematical theory requiring graduate study to grasp that warps your mind just like it warps space time.
A lot of people are like Flat Earthers who won't look through a pair of binoculars to watch a ship disappear over the horizon. It's both right in front of their face and yet beyond their reach; it needs just a smidgen of education, and they'd see it for themselves.
It is this way with climate change denialists. But things are looking up! Monarchs are basically extinct now, the Arctic will be ice-free in summer sooner than expected, and floods in Florida are picking up! Some people are gonna get an education soon!
>Have you read a single page of any IPCC report or any major climate science publication?
As a matter of fact I have, though not in a few years. The papers and the commentary are both far less certain in their assertions than you would ever guess judging by the common fervor against "denialists".
>Like, "plants like CO2 so more CO2 might be good" is naive, elementary school reasoning. You might as well say "cardio exercise is good for the heart, so maybe medically inducing tachycardia is good for me". It's a complete non-sequitur with no empirical grounding.
No, it's based on studies of modern [0] and prehistoric plants. The Carboniferous period for example had CO2 levels comparable to today's and an explosion of coal forming plant growth.
>The rhetoric you're spewing is uninformed and dangerous. Even if you are exactly right (and well-founded scientific arguments would suggest you are definitely not), you are speaking completely out of turn without any evidentiary basis.
The "rhetoric" I'm spewing is scientifically rational skepticism. The current science as practiced is one sided, primarily because of reactions like yours, which equate differing opinions with heresy. Climate science is far from settled and so called denialism is nowhere near the same level as flat earth/chemtrails/antivax skepticism to which it is dismissively compared.
Also you haven't made a single argument against my point that we are looking at a tiny fraction of a periodic signal and presuming a trend - and that's because there fundamentally is no argument against it.
> The Carboniferous period for example had CO2 levels comparable to today's and an explosion of coal forming plant growth.
This is not what I read here[1].
> Atmospheric carbon dioxide levels fell during the Carboniferous Period from roughly 8 times the current level in the beginning, to a level similar to today's at the end.[14]
> No, it's based on studies of modern [0] and prehistoric plants. The Carboniferous period for example had CO2 levels comparable to today's and an explosion of coal forming plant growth.
So back then they were "rapidly" sequestering it out of the atmosphere rather than doing their darndest to put it back in the air?
Do we also need to resurrect all those prehistoric plants to help deal with it instead of burning and clearing the tropical rainforests we currently have?
If the Carboniferous is how nature deals with too much CO2 over millions of years, we certainly aren't letting it take the same approach this time.
The IPCC doesn't produce scientific research as you are implying. The purpose of the IPCC is to justify the basis of "global climate action" so their opinion has no relevancy when it comes to debating the scientific upsides and downsides of rising CO2.
What few people seem to realize is that during the last ice age CO2 levels @180ppm was so low life on earth for mankind nearly collapsed - and we know CO2 levels on earth have reached 8000ppm in the past - so there is significant reason to be at ease with a rising CO2 considering how close to extinction level CO2 was a short while ago.
> and we know CO2 levels on earth have reached 8000ppm in the past
The last time that CO2 levels were around 8000ppm was also the time when the solar flux was several percent lower. So this argument doesn't really work the way you thought it worked.
This is a bit misleading as the data is US only but the climate and weather doesn’t care about our made up borders. So to get a real picture of climate change you need to look at global weather patterns and temperature changes.
https://www.climate.gov/sites/default/files/30yrNormal_Temp_...