CO2 absorbs infrared as logarithm of its concentration. So, graphs showing actual CO2 concentrations may be a bit alarmist. These concentrations are known quite well since 1958, and earlier, from (for example) Law Dome ice-core-entrapped bubbles. Time since 1850 is of interest, as before then thermometers were not well deployed to estimate global air T. As global compilations of surface thermometry don’t distinctively differ since 1850, I used HADCRUT4. Look first at log10 CO2, the black line. It shows slow increase until about 1960 and faster afterwards. This confirms what we generally know about the pace of fossil-C burning and industrial development. Red dots are HADCRUT4 annual T anomalies and blue dots are the UAH TLT annual (derived) T anomalies. One reason those don’t match is because of different basis periods, but these are all only reported values. The part of this graph that most weakens a ‘strong CO2 control’ argument, in my opinion, is from 1911 to 1945. T was then on a steep upward slope but CO2 was not. During that time solar (sunspot) cycles increased. ENSO cycle spent most years on the positive side, but was not really strong. No basis to suggest heat release from oceans, because data are not available. In short, this is a difficult period to learn from. Two periods of small air-T declines (1880 to 1890 and 1900 to 1910) had large volcanic eruptions to which we may appeal. But a third one (1945 to 1964) did not, and dirty coal burning (sulfate aerosols) is the typical go-to explanation there. In all of these, it may be that the ocean ‘sunk’ heat, but I see no way to get that information now. It would be a large step, and potentially unrealistic, to decide that only post-1960 T & CO2 data should be used to anticipate our future. You have been warned. Also bear in mind that COP21 meeting concluded that CO2 ought not to increase a lot more. But, y’know, maybe it will, because MONEY. Increasing CO2 to 800 ppm (the number I believe we won’t exceed) on this slope would increase air T to 3.5 oC above the baseline (HADCRUT4). Based on UAH TLT full record, by a smaller 2.4 oC. You should realize that models summarized by IPCC would give higher +T, and the reason is supposed to be amplification by IR absorption by atmospheric water vapor. I am not ready to buy that, and we could discuss it later. For now, let us go back to the figure and look at red dots (HADCRUT4) and blue dots (UAH TLT). Blue are much more variable through time, and nobody has told us why that is. I suppose that orbiting radiometers may ‘see’ more warmer, lower air some years and more colder, upper air in other years. Why more temporally variable satellite records are ‘the cure’ for surface-T measurements possibly sited in enlarging urban heat islands, has never been explained. I believe we are stuck with continuing both data sets. Especially surface stations, because they are in the world where we live. It should go without saying (I hope) that in these recent decades, T is anti correlated with solar cycles. Already examined elsewhere. We may be now in a situation where CO2 is firmly in the driver’s seat, with solar and ocean cycles reduced to secondary roles. It was not so when CO2 amounts and increases were less. Considering only 2 periods, before and after 1960, average T increase per log10 CO2 may have been larger in the first. Not significantly so (lots of ‘noise’ in T). This is hard to see from just looking at this graph, but if anybody wants underlying data, just say so. I only argue that +log10CO2 is persistently similarly positively correlated with +T throughout this time.
Hi Doug, I am uncomfortable discussing the UAH data because I read a summary of the code quality. I am currently dealing with some OBD data and it is so easy to have non-linear errors get injected by the code. One reason I like Berkeley is they started from the source data and paid close attention to the quality of the data. In effect they independently re-based the analysis paying close attention to input error detection and resolution. I am OK with ice core analysis. Unlike other paleo-records, there is overlap with the modern era and there are no systematic errors known, yet. One thing I didn't get was use of LOG10. So many physical phenomena are natural log based. Granted one can add a coefficient to scale between the two but that begs another, hard question about the constant. Our shared ancestor, Adam, in hindsight didn't do a great job of teaching his offspring how to understand the natural world ... any better than you and I have been able to bring enlightenment to problem children. Our feet are made of clay. Still you and I and a sliver of humanity are proof that sometimes we can stumble on elightenment. Bob Wilson
Not sure if this counts as a teachable moment or not. To change vertical axis to natural logs, multiply each number by 2.303* All logarithmic bases are interconvertible by fixed multipliers. The constant does not represent a hard question. Yes, it would have been better to use LN; Log(e). But only slightly, and only because Beer's law is a negative exponential. UAH (now) = RSS TLT, pretty much, and except for the latter's slower growth in the last 3 (ish) years. Unexplained to my knowledge. Satellite records need to continue. Their higher variability among years may teach something that climate models would benefit from. Besides, shutting them down would elicit howls of protest. BEST and all the other surface-T compilations are very similar over this interval. I am confident that the same general patterns would appear with any. Variable correlations with CO2 before 1960 and a hard to ignore pattern after. Perhaps lost in the long text is that overall T/logCO2 ratios are indistinguishable in these two periods (before and after 1960). T just 'wandered' more in the first, among years. I am not aware that has been pointed out previously. *It's not 2.303 exactly it is an irrational number close to that
It was unfair of me to claim one set of observations over another. I'm OK with troposphere temperature changes that don't follow surface temperature records because ultimately it is question of water, the liquid and solid types. For that we have sea level and ice inventories. Not to put too fine a point on it, UAH/RSS pretty well have no direct contact except with the water and vapor fraction in clouds. Jason-3 is going through: Once they reach space, satellites are put through a “commissioning phase” that usually lasts a few months. During this phase, the teams check each of the satellite’s systems to make sure they are working properly. They will also evaluate the accuracy of the data coming from the satellite and make sure that the instruments it carries are properly calibrated. Once all systems are checked out and deemed operational, the satellite will be moved into its final destination orbit. Three days after Jason-3 launched, our partners at CNES began to acquire and process real-time data from the satellite. NOAA andCNES will continue to calibrate and validate the instruments and data while EUMETSAT conducts processing trials of the data received at the Usingen ground station. Once this six month phase is complete, Jason-3 will officially begin operations in its planned orbit. These highly detailed measurements of sea surface height, a measure used to study sea level rise, are a critical factor in understanding Earth’s dynamic climate. Sea surface height data are also used to study hurricane intensity, tsunami dynamics, El Niño Southern Oscillation, eddy dynamics, ocean boundary currents, coastal and shallow water tides, as well as weather and climate forecasting. Source: Jason-3 Satellite There is about an 800 times difference between the density of air and water which sternly biases my interest. Surface dwellers, very few of us live in the troposphere. So I don't look for a troposphere temperature reports to determine if I put on a sweater, jacket, or short sleeves. I don't even use troposphere temperature to predict global ice inventory This is not the first time folks have inflated one aspect over others. It won't be the last either. For example, one of my co-workers was advocating a course of action because "<xxx> product support contract is so expensive." Then I pointed out we work on a 'cost-plus' contract and if he kills <xxx> product, the program manager won't get to buy his next luxury car. Things have to be kept in perspective. As for criticism of UAH/RSS, I tend to follow the 'buzz' in skeptical science. It was unfair for me to repeat someone else's claim about poor software quality. Just some of the descriptions are consistent with some ugly software I've had to deal with in the past. So I'm sympathetic to such complaints even though I've not had to deal with UAH software directly. You are right, my persnickety comment about LOG10 was not called for. It sounds like an interesting, back-of-the-envelope analysis. Comparing LOG( CO{2} ) to UAH/RSS does raise an interesting question about how it relates to other common metrics: Berkeley (and other surface) temperature sea level ice inventory Bob Wilson
When the forcing function is talked about as a sensitivity of doubleing green house gas, then we have a story problem that says log base 2 not a natural log. If you use IPCC's guestimate of around 3 for the sensitivity, then, well multimplying log10 by 10 to 2 significant figures gets you there. Easy math from your graph. Note you can subtract the 0 line (you pick the year). You can just relabel your numbers going from -1 to 1 by 0.5 on the left axis. Pretty easy and readable for 3xlog2(ppm ghg) - 25.2. Sorry for teaching something different, but story problem rephrasing makes it logical to do S x log base 2. If we use alley's analogy of ghg being a thermostat, I would add that its in a drafty building with an open door where the air conditioning and heating can't keep up with the leaks (changes in the solar cycle, ocean oscilations, volcanos, etc). We still don't know the sensitivity, so we know we are turning it up, but not by how much.
On the graph as shown CO2 doubling adds 0.301 to the left axis. Every +0.301 is a doubling. There has not bee a doubling in the time graphed here. To extend the analogy @6, the heater got more BTUs after 1960 and now appears to 'beat the drafts'.
I didn't say there was, I multiplied by 3 the sensitivity that IPCC says they think is most likely. You appear to be going from 263-417 ppm on your co2 graph, and plotting log10[co2 concentration in ppm]. I simply was suggesting doing 3 x log2[co2 concentration in ppm] - 25.1. Hey you can center anywhere I picked 25.1 to center on your graph. You then have 3xlog2 going from -1 to 1 on your graph. Neeto. It just happened to work out that your numbers corresponded to nice round numbers when the IPCC's sensitivity was applied. Whether the heater (sun) with poor insulation (ghg) is able to keep up is a matter of scienific inquiry. Using the IPCCs most likely number going from pre-industrial 280 ppm to current 401 ppm should have heated us 1.55 degrees C if you follow my math. Global temperatures have fallen significantly short of this. This leaves 3 possibilities. 1) turning up the isulation takes time, this is leaky stuff that can't keep up 2) sensitivity is actually lower. 3) Gavin Schmidt's new wrinkle is to argue that there is an air conditioner that we don't see (it would be global cooling if not for the ghg) and the numbers add up if you add that factor, but we don't really know what the cooling factor is yet.
Your guys science education, is like pouring motor oil into a fine swiss watch.There has never been any correlation between CO2 and temps except for in your narrow cherry picked time period. Same with Arctic sea ice .The AGW cheerleaders choose the highest year of sea ice, 1979 to begin their comparisons.
Mojo @9 makes reference to the geocraft website, perhaps we did before. Anyway, if you go there you will learn much about fossils and coal etc. and I hope you'll look around there. There are always issues with paleo T and paleo CO2 proxies, and the further one goes back, those increase. For me, the most interesting times are 1) last 55 million years with strong T/CO2 relationships 2) last 5 million years when primate evolution gave rise to several human-like species 3) last 10,000 years when humans developed agriculture and high-density populations 4) last 200 years when much more detailed records became available. Geocraft website, for all its merits, won't help much with those. Many other sources have already been presented at PC for your consideration. But just for fun, consider the notion that Earth T could be know well 100s of millions of years ago with proxies (not updated for 20 ish years). Geocraft idea. Combine that with the notion that recent Earth T cannot be know well from 1000s of surface T measurements. Mojo idea. At the risk of personal head explosion, put those two notions together. If you survive, you well and truly a Mojo with all rights and privileges pertaining thereto.
