The recent surge in temperature anomalies due to the El Nino event in late 2015 seems to be over. Values have fallen towards ‘normal’ in June. One recent claim is that the record anomaly in May exceeding 1 deg.C is in line with the mean projections of CMIP5 models (Gavin Schmidt). However if we compare the H4 data directly to an ensemble of CMIP5 models, then we see that the overall data still lay well below the expected trend.
Comparison of Hadcrut4.4 monthly data with an ensemble of 40 CMIP5 models
It also looks like the El Nino peak is now decaying back towards an underlying trend value of ~0.5C, as happened in 1998 and 1974. July’s data will be interesting to see if this trend is continuing.
Finally shown below is an animation of the spatial dependency for 2015 and 2016 so far. This shows the rise and decay of the El Nino hot spot west of Chile.
Animation of monthly anomaly values showing the growth and decay of El Nino.
I am using a linear colour scale with temperature. Some of the more scary versions of this plot (GISS) use a logarithmic colour scale with a low ‘hot’ threshold for enhanced visual effect. I think my scale is better because it suppresses small scale noise of ± 1C in a single cell.
Clive,
To be honest if I were a modeler I would find the agreement you report between actual data (Hadcrut4) and model projections (CMIP5) encouraging. Also, my visual impression of your animated map is far more of the world is red than blue most of the time. I must say that it seems we spend a bit too much time arguing about how fast the ship is sinking rather than trying to deal with the pretty clear evidence that it is sinking (pardon the analogy).
You’re right that there is a level of agreement between the data and the models. You would expect them to agree before ~1990 since they are tuned to fit the past. However future warming is the crucial unknown factor. In particular the climate sensitivity for a doubling of CO2 levels, or when CO2 levels reach 560 ppm. I would argue the data supports a lower value of TCR within the AR5 limits or ~ 1.4C, eventually reaching ~ 2.3C after a few hundred years (ECS).
If warming is limited to 2.5C then the ship won’t sink. Many places in temperate regions might even be better off. It is very unlikely that we can stop CO2 levels doubling but we might be able to limit emissions so it doesn’t increase beyond that.
Most of the map is reddish because the temperature anomalies are all calculated relative to 1961-1990 period. So any place that is warmer than ~1975 will be reddish.
Clive,
My bet is, if globally there were a full court press to prevent CO2 levels (more correctly collective GHG levels) from doubling the end result would be we might prevent levels from quadrupling. In reality, though, there is no serious effort at all. Efforts to date have been political window dressing.
I am not sure why it matters that the reference year for animated map is around 1975. Pick another reference time frame and the trend is still the trend, though it may not become apparent in the same year.
Bryce Payne:
The trend is very sensitive to where you start the baseline. For example, in the 1970s there were a group of climatologists worried about the trend from 1940 to 1972 and thought that our interglacial was ending, which would have been a true global catastrophe like depicted in the movie “The Day After Tomorrow.” BTW, it is still a worry that warming could trigger the AMOC (Gulf Stream) to slow and result like the movie. There is a real concern for this as our ice cores can see previous interglacials could end in the current orbital conditions. It would take one supervolcano, asteroid strike, nuclear exchange or combination to cause real trouble. We may have been teetering on that edge 400 years ago in the Little Ice Age due to a likely combination of low solar activity, high volcanic and perhaps deep sea overturning. CO2 from fossil fuel may have arrived in the nick of time. We don’t have enough accessible fossil fuel to quadruple CO2.
Electric cars for mass market are just a year or two away. Solar panels are going up now in houses in suburbs across the country. The bright spot is that we will not have severe warming but just enough that everyone can claim victory.
Hope springs eternal……maybe.
Humans, like other animals, seem to do what is necessary, but not a lot more.
Trying to imagine now the new tools we are going to have for alternative energy and mitigation of global temperature in 60-100 years should be like trying to imagine how one would travel to the Moon in the age before flight. Yet, we already have labs working toward direct solutions. As Nobel laureate Freeman Dysan said recently, “Are you trying to tell me we won’t be able to find a way to make it snow a little more in the Antarctic?”
I’m thinking that as the cloud patterns move poleward that problem just might solve itself. But that is admittedly very optimistic. Plus half the problem is sea level rise caused by thermal expansion of the sea. My solution for that would be a surfactant producing algae that would promote sea spoon (foam) that can build into island sized blankets naturally now. Another solution that would cool would be to introduce a non-polluting aerosol to the stratosphere, like the contrail conspiracy theory but for real.
Clive, do you know if Hadley still abstains from doing an urban adjustment? The last documentation I have on the subject is 2013.
I am pretty sure there is no adjustment made for UHI by Hadley. Urban heating is a very subtle effect due to the IPCC adoption of temperature anomalies rather than measuring absolute temperatures. Each station is normalised to its mean value in say 1975. For fast growing cities like Beijing this results in PAST ‘temperatures’ appearing colder than they really were. The use of anomalies suppresses UHI making pre-industrial temperatures appear colder !
As result Global warming relies more on the 70% contribution of SST to global temperatures. Yet before satellite data became available we relied on bucket measurements. How reliable can those really be and how can we audit the various adjustments that were made? The answer is we can’t. It is an act of faith.
Probably the satellite era is the only unambiguous measure of AGW.
Clive, I would agree with you that if satellites and balloon measurements, (the former going back to 1979 and the later to the 1950s,) are accurate then we are seeing a 30-40% bias in the land/SST record. In the land record the non-climate effects of UHI, rural land use, land change LULC and micro-site issues near station temperature sensors are the prime suspects for bias. In the SST I would say it’s faulty adjustments for bucket vs intake vs buoys, along with infilling for the southern hemisphere to make up for the lack of historic data. If there is bias in only one, the land or SST, then it would have to be too be an implausible amount. Also, each is verified to the other according to the modeled ratio of land:SST warming. So if only one is biased then the models are way off in that in addition to temp trend. If both surface records are correct then the balloon and satellites are wrong.
Do you know anything about the
THE INTERNATIONAL TEMPERATURE DATA REVIEW PROJECT investigation into the temp adjustments? I just stumbled upon their press release from a year ago, here.
Also, have you been following the UAH/RSS dual over which satellite is the faulty one and which is good? Both outfits use the same NASA MSU and AMSU (newer) orbiting sensors. The UAH says the newer model is calibrated correctly but RSS is sticking to the older MSU that is warmer but closer to retirement. A good amount of reputation, (as well as GMST trend,) hangs in the confirmation of the existing AMSU by the next launch, I suppose.
Yes. I wrote a paper for the ‘Temperature Data Review Project’ where I compared uncorrected GHCN with corrected GHCN among other things. I was contacted later to ask to be a co-author of a joint paper on this with others. However, as far as I am aware nothing ever came of it. I think I will put my paper up on the blog soon.
I was unaware of any feud between UAH and RSS. I had always assumed that they essentially agreed with each other, although I can imagine RSS have been under pressure to come better in agreement with the instrument data (GISS) !
Using 89 sinusoids I have fit the latest data point. I use a Marguardt scheme in fitting the data that minimizes the sum of the squares error. I have been doing this for some time now and I even include a contribution from CO2 based upon ECS or TCR. The value of ECS I determined from getting this fit was 0.263. This is obviously much lower than current estimates.
Here are the figures that go with the analysis.
https://1drv.ms/i/s!AkPliAI0REKh_T3hNB24tGmnYgDI
https://1drv.ms/i/s!AkPliAI0REKh_T7Am0L78U9DAkNV
The correlation values for each of these figures are shown. I don’t think anyone would choose to argue with me that they are very good fits.
I do find a modest contribution from CO2. Does this qualify me as a member in the 97% consensus?
Below are figures for my estimate of CO2 and the Anomaly value associated with those levels of CO2.
https://1drv.ms/i/s!AkPliAI0REKh_UBRLVgEbvHxqpha
https://1drv.ms/i/s!AkPliAI0REKh_T97J4qthMuitkB9
When I was younger I used to like Peggy Lee. I am reminded of one of her songs, “Is that all there is?”
I follow the efforts of Dr. Evans. He has proposed a model change that would lower ECS values, I believe, to no more than 0.5. I would say my analysis of the data support that.
Wow 89 sinusoids! I only managed to fit 3 – https://clivebest.com/blog/?p=2353
If natural oscillations in temperature such as PDO, AMO, El Nino … all bias the temperature data on different time scales then all bets are off.
My gut feeling is that there is a clear 60y oscillation and an El Nino effect clearly visible. If you take these away then the underlying logarithmic warming term is around 1C. I doubt whether the data are accurate enough to see other natural terms. However, if you are right then we should see temperatures fall in the future to catch up with the underlying CO2 term.
Is this the Dr Evans you refer to ?
http://business.financialpost.com/fp-comment/climate-models-go-cold
His argument is that H2O feedbacks are negative rather than positive. IPCC ‘science’ predicts water vapour boosts warming by a factor 3. If it reduced warming by a factor 3 then you get your estimate of ECS.
Clive
The other day I posted a comment on WUWT that essentially furnished the H4 analysis that I presented here.
https://wattsupwiththat.com/2016/08/03/ipcc-has-at-least-doubled-true-climate-sensitivity-a-demonstration/#comment-2271011
I am not sure anyone bothered to read it. I did see a comment later on about curve fitting but I don’t know whether that applied to me or not.
That did not happen here and I appreciate the time people took to look at what I furnished.
Your paragraph above gave me pause.
“His argument is that H2O feedbacks are negative rather than positive. IPCC ‘science’ predicts water vapour boosts warming by a factor 3. If it reduced warming by a factor 3 then you get your estimate of ECS.”
Add to that, I did see links to your efforts in one of the comments.
https://clivebest.com/blog/?p=3659
https://clivebest.com/blog/?p=3258
New RSS data became available just the other day. Back in April was the last time I tried to introduce an ECS value to the tropics data. Many times the result would come up with a negative ECS value. I quit trying and went with only the cyclic fit.
I tried again and it came out negative. I might just be making a bad guess. Based upon your comment above and the links I mentioned is it even possible that ECS could be negative in the tropics region?
The problem with WUWT and Judith Curry’s blob Climate etc. is that you can easily get lost with the flood of comments, and trying to work out what is a reply to a previous comment.
The tropics do show far less warming – perhaps even zero warming. Temperatures are mostly limited to 30C by evaporation from the dominant coverage of oceans. This is due to the Claudius Clapeyron equation and results in the ICZ is dominated by clouds and thunderstorms . The troposphere is also much higher in the tropics due to the lower moist adiabatic lapse rate.
Don’t forget also that El Nino also plays out in the tropics.
Yes, I am pretty sure that is the same Dr. Evans. His website is here. I do subscribe to his ND theory.
http://sciencespeak.com/
I started looking at cycles a few years ago. I spent 35 tears working on rotating equipment so I was very familiar with the FFT.
If you go to Dr. Evans website, particularly on his ND, you will find a link to a spreadsheet. Dr. Evans created the Optimal Fourier Transform (OFT). I use it to come up with my initial guesses. If I choose to fit the data with 90 cycles that gives me 270 initial guesses to put through my Marquardt procedure. Generally, things get even better after that. The OFT is under the transform tab in his spreadsheet, I think.
The 60 year cycle got me started and initially I used only about 6 cycles. I am way beyond that now.
BTW, I have finished looking at the H4 data for the Southern Hemisphere and the tropics.
SH
https://1drv.ms/i/s!AkPliAI0REKh_UHgmEhRqmAIBvhk
https://1drv.ms/i/s!AkPliAI0REKh_ULmGQNWpHj62KWu
Tropics
In the past I have not been able to come up with an ECS value for the tropics. It even came out negative sometimes and I decide to omit ECs from the tropics.
https://1drv.ms/i/s!AkPliAI0REKh_UMwBSwmiRWPB1vb
https://1drv.ms/i/s!AkPliAI0REKh_UQFI8HP96BhiAxj
Presently, I am looking at the NH. The data has more noise and does come out as well as the others.
Besides evaluating Hadcrut data I also analyze RSS, UAH and the data from all four NINO regions. BTW, the RSS and UAH data are the best and I suspect it is because they haven’t been tampered with. The cyclic nature of the NINO regions is very apparent.
I agree with Dr. Curry. Natural variability is not properly accounted for. From my evaluations it looks like natural variability is the biggest part.
If you like I will add the NH evaluation when it is done.
BTW, I went back to look at your link.
“Wow 89 sinusoids! I only managed to fit 3 – https://clivebest.com/blog/?p=2353”
I posted comments on that one. It is good to be back in contact with you. It has been a while.
I believe the 60Y cycle is really 65.
Yes I also think it is 65y !
Clive
Rather than furnish the analysis of the Northern Hemisphere, I thought this might be a better time to explain myself and what I am up to in spending so much time on analysis of the data from various sources.
I mentioned previously that I spend 35 years on rotating equipment. To put it simply most of what I learned came from instrumented test units. Over my career I implemented many design improvements and they all originated from ideas I got from reviewing test data and making sense of it. In doing this I sort of resembled Colombo. I would take facts from various sources and try to make sense of it.
From that experience, I believe the answers to many questions can be found in the measurements we already have. We need a Colombo to spend time looking at the data and trying to make sense of it all.
I once got into a big fight with a PHD from one of the labs who was an expert in fluid dynamics. I questioned his analysis and theory with my understanding of the test data. His answer was that since the data did not match his peer reviewed paper that I had bad data.
I won that argument and my proposed design improvements were implemented.
I am going to change focus here to Nino region 3.4. To me this is far more interesting at the moment.
https://1drv.ms/i/s!AkPliAI0REKh_UX9nKgWp169HA8a
It is noticeable in the next figure but I switched to the daily measurements in 2014.
https://1drv.ms/i/s!AkPliAI0REKh_UZhMH6BDz2kOEDg
https://1drv.ms/i/s!AkPliAI0REKh_Ue4lDXCbzAWMr7v
Since I have data going back to 1854 I am willing to project forward in time for a short period. In the figure you will see projections for dates. Those dates are the date of the raw data measurements. You can see that the projections have changed over time but you will notice that all projections match the measured data.
The projections may not be precise quantitatively but are more qualitatively correct. It does look like we may be headed to another El Nino.
https://1drv.ms/i/s!AkPliAI0REKh_Ug09NfsEi73NOzb
You will not see the blue line on this chart since it is buried under the dated latest projection. The latest daily update goes back to July 5, 2016. I don’t know when their next update will be but they don’t seem to be on a regular schedule.
I do differ with NOAA on this. I watch Joe Bastardi’s weekly updates. Back in April I think he said that we would have a strong La Nina according to Scripps with a value of around -2.3 C. That is not going to happen. If anything NOAA is moving in my direction now.
This is going to be fun. Who gets it right the GCMs or some redneck with a cyclic analysis. Don’t bet against me (yet). We are near the bottom.
To emphasize again, I believe the answers we seek can be found in the data.
I completely agree that measurement data must prevail over theory no matter how certain scientists believe they are right. That has been the only way that progress has been made in science. You only need to look at the history as to the origins of Quantum Mechanics and Relativity.
You seem to have a sophisticated tool to resolve underlying Fourier components in signals. Your prediction post El Nino looks spot on because warm surface water must now disperse as wind driven water flows westward towards a new La Nina. So you are right that by analysing regular cycles you can do better than medium range climate models, because they do not understand El Nino, PDA/AMO etc. I am interested in the underlying causes of El Nino. I suspect that the latitude spread of new moons in the 18.6y precession cycle may play a role.
Sophisticated signal processing will always win when there are underlying regular physical processes. I would like to understand what those physical processes are. CO2 forcing is just one or them.
Here I will try to furnish some additional details on what I have been doing and I will give you the underlying cycles for Nino region 3.4.
I did not invent this analysis method out of thin air. I am using TKSolver which already had a nonlinear curve fitting model that I have adapted to my purposes. I used this method to come up with fits for the BH properties of some of the motor steels I was using. I started using the DOS version of this program back in 1984. I used it to come up with a 3 phase induction motor design program. The built in iterative solver is something to die for. I would put guesses in for the air gap voltage and other parameters and design an electric motor.
Here is the function sheet where I would place the equations that I wanted to fit to the data. BTW, b0 are the initial guesses and b are the final answers that you will see in a another figure. The stuff at the bottom is where I added in the DC offset.
https://1drv.ms/i/s!AkPliAI0REKh_UuTGXk4gNEnA9kA
Nino 3.4 is fit with 90 sinusoids meaning there are 270 guesses plus an additional DC offset.
https://1drv.ms/i/s!AkPliAI0REKh_UoQOTU0LAgjcjF7
Here is the full table of all the sinusoids. Ignore the values for the OFT that value likely originates when I was using only monthly values for the data. Look it over maybe you will find some periods that make sense to you. Anytime the light bulb is lit there should be applause. Phase is radians.
https://1drv.ms/i/s!AkPliAI0REKh_Um7vuXMVN3spCln
I don’t know whether I will get to it today but I will do the same for the H4 analysis. It is going to take some time to show you how it developed. Like you I started with a very few number of cycle.
charplum, Clive, we have been discussing the land records and adjustments at Lucia’s Blackboard for the last couple of weeks. Steve Mosher started the post and vanished after an error was pointed out. We’ve continued without him and would welcome some new perspectives. http://rankexploits.com/musings/2016/population-ii/
Clive
Like you this all started with a few number of cycles. The talk I was seeing on the 60-year cycle prompted me to investigate. Eyeball examination of the data sure indicated it.
While focusing on the 60-year cycle I became aware of this.
https://1drv.ms/i/s!AkPliAI0REKh3xI4ygzxOZ7QmkjN
I used only 6 cycles in this analysis and two of them were picked from the above figure. One cycle was 350 years and another at around 974 years. Her is the result.
https://1drv.ms/i/s!AkPliAI0REKh_UzKL-mP6v7WgbjU
Even with only six cycles the correlation coefficient was 0.867. Since I only had over 150-years’ worth of data the only place I could look was back in time to see if I could get some kind of confirmation that this was on track.
The next figure told me I was on the right track in that the LIA and the MWP were about where they were supposed to be.
https://1drv.ms/i/s!AkPliAI0REKh_U1kUs1sxNVFuX6l
The resulting waves are in this table.
https://1drv.ms/i/s!AkPliAI0REKh_U4v8yLb-ahXy1-3
Much has happened since then.
I mentioned before that I rely on the OFT to identify the many frequencies I now use but not all of them come from the OFT. Out of only 150 years plus of data the longest period cycle identified was 273 years. I input myself the two long cycles identified above at around 1000 years and 350 years.
Here is the table of values that goes with the earlier H4 analysis. BTW, I was wrong about the 89 cycles. It is 107. I failed to change the name of the file.
I am not sure I need that many waves but you do see the results.
https://1drv.ms/i/s!AkPliAI0REKh_U_jxujLb7YXd2fV
Here is what the near future may hold.
https://1drv.ms/i/s!AkPliAI0REKh_VCUZABOIIzgFfOD
I do communicate with Dr. Norman Page who also subscribes to a delay like Dr. Evans. I am sure he likes the dip in 2022 which he predicts.
There is a conflict between the above chart and what is shown for Nino region 3.4 which has us heading into another El Nino.
This is where the Colombo role is needed. There was a prominent El Nino in 1982 yet it really fails to show in either the H4 data or the RSS and UAH data with any significance. Look at the North Pole RSS data and you will clearly see a prominent contribution from the just passed El Nino yet the one in 97-98 is not clearly apparent.
I can’t explain that. For now, I will keep doing what I am doing and hope the light bulb comes on.
This week I expect to have new RSS and UAH data and, perhaps, a daily update to the Nino region data. The last one was July 5.
What is interesting is that you have added long term climate terms as well. I love this graph as if true would explain the MWP and LIA, although quite what the physical process behind such a strong climate cycle is unclear to me.
You seem to have also added short oscillation terms for El Nino as well. If you’re fits are right then we will surely know by 2018, as you predict a massive temperature drop of nearly 0.7C !
That certainly would cause a few embarrassed faces !
This simply needs further clarification and I do have questions myself.
I have always been inclined to believe that the Sun drives the bus. The cycles that I added are what is in this table.
https://1drv.ms/i/s!AkPliAI0REKh_VWG7ibfxtvJp1hh
Dr. Evans had a figure in his spreadsheet that hit me like a ton of bricks.
He devised his own temperature reconstruction going back to the time of Christ and combined with the knowledge we have of CO2 from Law Dome ice cores.
I could spend time to relocate it but I analyzed it here and came up with my own.
https://1drv.ms/i/s!AkPliAI0REKh_VdDq3FtUJlmFwrj
Observe the green line. That is the temperature anomaly from CO2 and the reason it is so flat is that the Law Dome CO2 record is flat until around 1800.
It begs and answer to the question that if CO2 is now driving the climate then what drove it in those earlier years?
I think this amplifies why some have had to eliminate the MWP and LIA from the record.
You will notice that the record is longer than it should be. That is because at that time I spliced in my projections for Hadcrut 4 beginning in 1850. It does appear we will be getting colder.
The cyclic analysis is a good fit considering the noise in the data.
Clive, you noticed the 0.7 C drop. That is predicted by the ND of Dr. Evans.
Here is the conflict.
The Nino region forecasts show us heading for another El Nino and the H4 projections show us heading in decline. Which is it or could it be both?
As I noted in a previous comment Nino has not always driven the global temperature. Take a good look at 1982. I am also aware that it takes a longer period of time to change the ocean temperature. Might that explain the disconnect?
Clive
Every Monday NOAA publishes a new weekly value for each of the Nino regions. I prefer the daily data but I have analyzed the weekly data too. It gives me a heads up where the daily is going to be headed. The weekly data are now through July 27.
Adjustments in the cyclic analysis are still being made but I think we are done with big alterations in the cyclic fit. We should be close to the bottom.
This is for region 3.4. I am currently conduction the analysis on the other two regions I look at on a weekly basis, those being 1.2 and 3.0. I also look at 4.0 when I have new daily data. I do not do the weekly on that one.
Region 3.4
https://1drv.ms/i/s!AkPliAI0REKh_VLGE5CU0P9sStyF
I am furnishing a wider view this time of the projections.
https://1drv.ms/i/s!AkPliAI0REKh_VHsXm0ujCUOroMM
Since the analysis just completed I will furnish a bonus of region 1.2. These data are noisier and do not furnish the same correlation coefficients but as you will see they are still good fits.
I do not know the reason for the greater noise in the data but I think it might have something to do with region 1.2 being located against the coast of South America.
Region 1.2
https://1drv.ms/i/s!AkPliAI0REKh_VS3fEY5YBk5wJr8
The above figure is interesting between 2016 and 2016.5. Observe the up spike followed by a big down spike. What is that? A slug of warm water entering followed by a big slug of cold water with a bounce back. It almost looks like a damped response to an impulse. With each succeeding response lower in magnitude.
https://1drv.ms/i/s!AkPliAI0REKh_VOeiogvHVDiAq2v
Let me know what you think. I am willing to listen and learn.
If I have anything to add to your previous comment I will get back to you. I thought getting this new analysis out took priority.
You are doing an amazing job with signal processing. I personally wouldn’t push it too far for El Nino because we don’t have a physical understanding of what is happening. The same is true for longer term oscillations like PDO/AMO and others? I don’t claim to have one either but some thoughts are:
The earth spins every 24 hours but has a small Chandler wobble. Heat in the tropics moves air (and water) towards polar regions. The Coriolis force drives westerly trade winds in the North and easterly in the south. The wind piles up water towards Indonesia where it warms before eventually sloshing back towards Chile. This releases heat to colder regions rising global temperatures.
What role does the Chandler wobble play ? What role does the moon play as the position of tidal bulge change with seasons and the 18.6y precession cycle. Does the gravitational force of Jupiter have any effect? Does the solar cycle play a role?
So I think physics has to guide this. I would fix frequencies at known wavelengths and see which amplitudes survive.
This is just my 2 cents worth.
I agree with you about the danger of going to far with it. As you can see by my fuller presentation of the projections the forthcoming El Nino, if true, will be a monster. I am reluctant to believe that.
Let’s say qualitatively what I am predicting is true. I believe as things turn upward the magnitude of the prediction will decline as more data are included. You can see how the depth of the current trough changed with added data.. I suspect something similar will happen as we head up or even if we head up.
I pointed to that sudden drop in temperature in Nino region 1.2. The drop happened in one week. When I saw that I went too far and thought this proves the La Nina will happen. Next week the temperature rebounded.
I jumped when I should not have. Live and learn.
In any case, I do appreciate your thoughtfulness on this.
Pingback: Hinkley, renewables and the social cost of carbon | Matt Ridley
Clive
I have another added bonus. The first of the month is always a busy time for me. New RSS and UAH data become available. I am analyzing the new UAH global, NH, SH and tropics data. One will do for now the global data have been analyzed.
https://1drv.ms/i/s!AkPliAI0REKh_Vi5xxj-DJ7HpFZz
https://1drv.ms/i/s!AkPliAI0REKh_VnNJVsQAULxNUqu
In the second chart it is more apparent but the cyclic fit ran past the new data point. It was essentially ignored. You will see another spur on the other side of the El Nino. This point may be just like that. It may require a few more months of data before the cyclic fit is forced to respond.
The correlation coefficient is 0.93 and the fit clearly passes the eyeball test. The ECS value is much lower which is typical of what I find for both satellite records. The value is 0.155.
With only 35+ years of data I do not project forward. I have done that but the results are erratic.
In both of the satellite records the El Nino event in 97-98 and 2015-2016 are plain to see. Going back to one of my earlier comments what happened to the one in 1982? Maybe you can make a remark that there is a hint of it but it is not like the two mentioned.
I am not sure which way things are headed.
A fit applying 89 sinusoidal components for 35 years of data considering that random non-sinusoidal volcanic events occur? cool.
In this posting I mentioned that I was looking at the other UAH regions separately. That is now done.
From what I can tell I can just see more uncertainty.
Here is Nino region 3.4 on the same time scale as the UAH data.
https://1drv.ms/i/s!AkPliAI0REKh_V0Fm7xHmCBm3sHr
Note the El Nino peak in 1982. It almost reaches the same magnitude as the one in 97-98. We already know that that the latter one clearly impacted global temperature.
https://1drv.ms/i/s!AkPliAI0REKh_Vi5xxj-DJ7HpFZz
On second look you might suggest that the 1982 event may have shown itself in global temperatures but it is not clear like the latter in 97-98.
Here is the NH UAH.
https://1drv.ms/i/s!AkPliAI0REKh_VukYSlFhXK362MI
You might be able to make another similar argument that it is there but not readily as detectable and certainly not prominent.
That picture changes with the SH.
https://1drv.ms/i/s!AkPliAI0REKh_VxSeoVqL5H2P_Tr
It is much easier to see here, I think.
I think it goes back to what Clive has said that the El Nino events are not well understood. Hence, predictions are likely not reliable. I am going to compound that further by suggesting that we don’t really know when they will play a prominent role in the measurement of global temperature.
That peak in 1982, although prominent, played a much smaller role in determining global temperatures.
The more I investigate on my own the more uncertainty I see.
charplum,
Fascinating. Could you make lines on graphs bolder (thicker). Most of them render on my screen as just a thin black line, even if you have more color codes in your plot legend.
Thanks.
Bryce
I can certainly try.
The program I am using, TKSolver, has its own graphics. The practice that I have been using in storing my charts is to copy the chart into paint. I then crop it and store it as a jpg file.
I have another option which is to use the snipping tool and use a screen capture and save that.
Here is a repeat of one of the charts presented. This is as it was when posted.
https://1drv.ms/i/s!AkPliAI0REKh_T3hNB24tGmnYgDI
This is the same chart made with the snipping tool.
https://1drv.ms/i/s!AkPliAI0REKh_VqwZ9T6FPXp5b80
If one is preferred let me know. Otherwise, I will work more with the line widths and colors to make things easier.
I am open to suggestions if there is a better way.
charplum,
Snipping tool is a bit better, but that is not a good example graph in that all color coding is fairly obvious in either version. Have a look at https://1drv.ms/i/s!AkPliAI0REKh_VOeiogvHVDiAq2v. In that graph I can see only black lines with a color code legend. Can get basic point, but would be at least interesting to be sure instead of guessing about what you are seeing.
With the program I am using I have a choice of 9 colors and no choice on line style. I changed the thickness of the lines and it does seem better.
https://1drv.ms/i/s!AkPliAI0REKh_V4mw22yub9bYwPL
The end of the month and the beginning of the month can be busy times.
I had earlier mentioned that the last update of the daily Nino region data was July 5. Wouldn’t you know it, last night new daily data became available. It is dated August 1.
Over night I analyzed all four regions. One of the benefits you will see is that this information I put in a spreadsheet which also allows me to put the monthly data on the same chart. Nothing really has changed on the correlations and I do think they all pass the eyeball test.
Region 1.2
https://1drv.ms/i/s!AkPliAI0REKh_WIsLFdYZ7QoeFTH
Region 3.0
https://1drv.ms/i/s!AkPliAI0REKh_WMZ-KZwHiTVW5Ya
Region 3.4
https://1drv.ms/i/s!AkPliAI0REKh_WE3YTCsuXisetR8
Region 4.0
https://1drv.ms/i/s!AkPliAI0REKh_WDHKqY8isu_MMST
I put more effort into the daily data analysis. I used to only do the monthly data but when you look at the charts I think they tell us more with the daily data.
The only thing left are the RSS data analysis but since you have seen the UAH you have seen about the same.
I can only hope that someone got something useful from what I have furnished. I have been doing this for a while and it just seems to me that cycles explains a lot of what we see in climate measurements.
Just trying to help.
charplum,
Increased line thickness definitely helps. Much better.
Is there anyone who would actually contend that cycles would not explain a lot of what we see in climate measurements.? The concern, and the signal we should be concerned about, is precisely when the measurements indicate sustained divergence from the cyclical patterns.
Some of your plots suggest, along with some of your comments, that we may be on the beginning edge of such divergence. But, data is limited, and the adaptation challenge is that the climate system is incomprehensibly massive with a similarly massive inertia. When the divergence is sufficiently strong to provide a high level of confidence that the divergence is real, the climate will be too far down that divergent path to alter anything on functionally relevant time and mass scales. Stated differently, we, or I should say, I would expect early divergences to be small and erratic with respect to the cycle based projections. We will be inclined to not perceive them. That is, cycle based projections will be heavily weighted to look like the past when what we are concerned about is the likelihood of a future in which past cycles are no longer reliable predictors. There seems to me no reason to expect the climate system will suddenly jump from one general condition to another, though I also cannot rule such a shift out. The only way for cyclical analysis to discern a divergence is to have enough post divergence data, which is precisely the same predicament.
What you have done is fascinating and impressive and speaks in a number of ways to the issue, but at this point does not seem to provide the “soothsaying with certainty” that is necessary to resolve current contentiousness. Please do not stop doing what you are doing, and availing us of it. It is, in my humble opinion, a substantial contribution.
I think the best way for me to answer that is to give you an example of what I went through.
About two years ago I became familiar with Dr. Evans Notch Delay (ND) theory and I down loaded his spreadsheet. Here is a screen capture of what is in that spreadsheet.
https://1drv.ms/i/s!AkPliAI0REKh_WSjQSjaRE6yn7cM
With Dr. Evans spreadsheet you could choose how much of a solar contribution you wanted and how much CO2. In this instance I believe I chose 20% CO2 and 80% solar. The red line shows the precipitous drop in temperature from the ND.
The next thing that hit me is that at that time with the cyclic analysis I was showing a similar abrupt decline. I brought that to Dr. Evans attention.
I went around and around on this in my head. How could the cyclic analysis be showing the same thing?
I finally argued with myself enough that it could predict this drop. How I got there is simple in the end. Solar cycles heat the Earth with various frequencies and phases.and what this represented was a unique combination of these frequencies and phases to produce this.
Note that my current analysis of the H4 data shows this decline.
This countered by the Nino region predictions of heading back to another El Nino.
I did show in an earlier comment that the 82 El Nino failed to noticeably impact global temperatures. It has happened.
It is perhaps too much to hope for but this should present a real dilemma and one that may end this once and for all.
How do you explain declining surface temperatures while proceeding into an El Nino? I think thermal inertia may account for this discrepancy but I am not the physics expert here. I did read a Russian paper fa few years ago suggesting a time constant of 14 years. That would be about right.
It may be, as you suggest, a whole new paradigm or it may be as simple as we just need better explanations. If this happens it will be devastating to the GCMs. The climate change scare will pass on by just like the ice age scare of the 70s.
They won’t give up there easily. There is a lot of well-connected money and interests involved.
I am not planning to fly to Vegas and bet the farm on the dilemma occurring.
On this subject look to Dr. Evans and Dr. Norman Page. They both suggest a dely.
This next year could be very interesting.
My compliments to Clive and his site. There have been many good discussions.
Clive, reading your blog post from Aug 2011 on clouds and humidity, I was surprised to learn that the upper troposphere is trending lower in humidity over the last few decades as well as clouds in the lower troposphere. Both these seem counter-intuitive to Clausius Clapeyron. Could part of the reason for less clouds be quicker water cycle? Could pollutants be seeding clouds? On separate question, do you have an opinion on the quality and value of reanalysis as outlined in the new Cederlof paper ? It shows an upper troposphere hotspot. Also any thoughts on both GIS and HAD having over double the land surface warming relative to SST for the last 37 years?
Ron,
It depends on what is cause and what is effect. If AGW is the driving force then increasing surface temperatures should lead to more evaporation and more clouds. However, the ISCCP data show the opposite and consequently have been dismissed by the IPCC as being erroneous. However if they are correct and something external to the system (cosmic rays?, aerosols?) is reducing global cloud cover then part of the raise in surface temperatures is an effect and not a cause.
The NVAP data also show no real change in relative humidity https://clivebest.com/blog/?p=4517 See also Water Vapor Decline Cools the Earth: NASA Satellite Data
I have only just looked now at the Cederlof paper. Reanalysis data when tied to data assimilation of real measurements is about the only way to get a global 3-d view. The land (H4) and SST (HSST) trends are certainly different. What the paper seems to show is an increase in lapse rate over hot arid areas and that it is here the upper troposphere hot spot appears. I am not sure off the cuff how to interpret this!
Clive or someone,
I intended a while back to take a different approach to analyzing actual, historical temperature data. I am inclined to initiate that effort now, but do not recall the most convenient source of actual historical temperature data. I recall there was a source, supplemented by more than one publication as I recall, where one could obtain both the temperature data as well as site history (location, local land uses/changes, proximity to urban areas, temperature measurement methods, etc.). Can anyone make a or some suggestions on how to access such collection(s) of temperature data?
GHCN uncorrected and corrected data can be found on their FTP site see https://clivebest.com/blog/?p=6534. I think Berkeley Earth claim to provide all raw station data but I never succeeded in finding it. You also need Matlab to run any of their software which costs > $1000 for a license. Individual station data can be downloaded for Hadcrut4 from the Met Office, but these are already corrected.
Monthly Tav values are calculated from the Tmax-Tmin values each day. I seem to remember that it is the Tmin values that are increasing rather than Tmax. This means global warming is happening at night rather than the day. This could well be affected by pollution and UHI.
The short answer is I have never found the raw measurement data for weather stations. Only the results after correction.
Clive
I thought this thread was over and then I watched Dr. Salby’s video on Judith Curry’s website. I remember him from a previous talk and when he was left high and dry in Germany. It is good to hear from him again.
After watching it a few things hit home with me.
For the most part of his presentation he used the Mauna Loa CO2 measurements. I have analyzed them and have a very precise fit to the data. I used this equation for modeling the data.
y[i]=b[1]*sin(2*pi()*b[2]*x[i]+b[3])+b[4]*x[i]^2+b[5]*x[i]+b[6]
It is nothing more than a sine wave on top of a quadratic. For those that might make use of this fit here are the constants that make it work.
B1=2.714413
B2=1.000478
B3=-.0603071
B4=.0120771
B5=-46.48858
B8=45038.15
https://1drv.ms/i/s!AkPliAI0REKh_Wb3HLJk968IDYrN
The correlation coefficient for the above is .9994.
Throughout his presentation Dr. Salby mentioned that CO2 grew at the rate of 1.6 PPMV/yr. The quadratic gives a variable value.
https://1drv.ms/i/s!AkPliAI0REKh_WXg7tdWWdNjAJ1R
I hope someone can make use of this.
However, it was first this chart from his presentation that grabbed my attention.
https://1drv.ms/i/s!AkPliAI0REKh_Wgby-pF8HUPauPK
It is the change in CO2 from 280 to 400 PPM that got my attention. That roughly covers the H4 data. In an earlier comment I identified that the ECS to fit that data was 0.259. Over the time period from 1850 to 2013 the rise in temperature from CO2 alone was 0.145. Not bad and it is roughly comparable to the one identified in Dr. Salby’s figure.
Later he modified this figure to this.
https://1drv.ms/i/s!AkPliAI0REKh_WcQovQYavOSZr2U
The value is much lower 0.01. However, if instead of using the RSS ECS value of 0.088 even that lower number from Dr. Salby is within reason or should we say striking distance.
I guess that one thing can be said is that we are both suggesting that CO2 is not that influential in the climate.
Clive, I did learn one more thing from this and comes from one of your earlier comments about sites like this. I was hoping to learn from the comments whether Dr. Salby would be savaged or accepted. I did not learn that. The long thread went off on a tangent and I stopped reading.
That is a very nice fit to the CO2 data! The quadratic term reflects that there is a slight increase in underlying CO2 increase rate. This is slightly at odds with what Salby claimed (fixed rate of change).
You should try posting your result as a comment on to new post by Guido van der Werf on Judith Curry’s blog https://judithcurry.com/2016/08/12/broad-consistency-between-patterns-of-fossil-fuel-emissions-and-atmospheric-co2/
There is of course some truth in what Salby says in his presentation, but it is not the whole story:
If somehow you could tag each CO2 molecule emitted by humans then on average their lifetime would be around 8 years exactly as Salby calculated. So apparently he is right and anthropogenic emissions do indeed decay rapidly. The problem is however that these extra ‘anthropogenic’ CO2 molecules displace other ‘natural’ CO2 molecules which would have otherwise been removed from the atmosphere.
Some sinks are sensitive to CO2 partial pressure and react quickly to any excess CO2 (e.g. dissolving in ocean). This is why about half the emissions quickly disappear each year. Other natural sinks react only very slowly. The accumulation of net CO2 in the atmosphere is because these other sinks have not had time to absorb the other half, before yet more are added. Eventually though they will respond. In fact you can show that even if human emissions continued indefinitely at a constant rate, then eventually CO2 levels in the atmosphere would equilibrate.
The level would just be very high !
Murry Salbly’s estimate of the radiative forcing of CO2 based on ‘opaqueness’ is too low when compared to line by line calculations. If everything else remained the same then ECS works out at ~ 1C However the biggest unknown is how water vapour reacts and this can change this figure significantly.
Clive
Before I got your reply I was going to make this a comment on the very article you mentioned. I was ready to post the comment and then I hesitated. Normally, when I post comments I check the follow box and I was hesitant to do that. This article had a flood of comments.
Not too long after that I decided to make the comment but not check the box. I was simply going to go back from time-to-time to see if there were any replies.
However, when I went back to post the comment I was unable to post it. The only way at that time to post a comment was through facebook or if you had a wordpress account. So it never got posted.
Just to add a little more, I saw this picture on another site and when I examined it I could see where Joe Bastardi came up with his identification of a strong La Nina. The Scripps prediction is the lowest. Things have changed a lot from just a short time ago.
https://1drv.ms/i/s!AkPliAI0REKh_W7uC5MMEg9QuvjB
Every Monday morning NOAA publishes new weekly measurements for the Nino regions. I have already analyzed three of the four regions. The next image is from region 1.2 and does look like it may have reached bottom and may now be headed up. It will take more measurements to clarify. Things seem to happen first in this region.
https://1drv.ms/i/s!AkPliAI0REKh_W8_R1TbPYRiVA5B
The others look like they are very close to the bottom.
Clive, from past comments there seems to have been some interest on this. I do think that in the next few months we will get clarification on where this is all headed. My question is should I continue to post comments/updates on this particular article or would you prefer something else. I did save a link to this article so I can do that easily. Let me know.
Clive
I know we both agree on the approximate 65-year cycle. Today I thought of a better way to reveal what I am suggesting with my work. I used this in an email to my contacts.
I am sure some of you are aware that I have fit the Hadcrut4 measurements with a number of sinusoids. To be precise I have used 107 sinusoids. The cyclic analysis fits the measurements with a resulting correlation of 0.95. That is close to 1.0 folks which is perfection.
https://1drv.ms/i/s!AkPliAI0REKh_XZcIHhQFxNq-zsG
Note the 65-year cycle sine wave that I included. Notice how it approximated the pattern in the raw data. Where the data reaches a peak the sine wave does too. The converse is also true. It is not exact but remember that this single wave is being combined with 106 more to produce the cyclical fit.
What you are also seeing here is what got me involved in trying to fit the data with sine waves. You can almost eyeball the data and see an approximate 60 year cycle.
There are two added pieces of information available here. I put a line representing the linear trend in the data on the curve and I also included my contribution from Co2 that you have seen before.
From the start back in 1850 the linear trend for the raw data change is 0.92C. For the change in the temperature from CO2 alone the change in magnitude is only 0.16C. It might then be said that CO2’s contribution to the total temperature change of 0.92C is less than 20% of the total.
There is no reason for a coal miner here in Pennsylvania to be put out of work other than the competition he faces from shale gas. Our EPA should not be a threat to him.
Clime, on another front I now have Nino region data available through August 17. I have analyzed all the regions but to just give you a glimpse of what I am seeing I furnish this.
https://1drv.ms/i/s!AkPliAI0REKh_XOFvr_SzGN1wxPt
Unless the direction changes they have bottomed out. The La Nina threshold of -0.8 will not be reached unless it is only momentary.
The next few months should be interesting.
I prefer the true daily data When that becomes available I will furnish a more thorough update.
This sums up the basic problem of climate ‘science’. Good scientists analyse data and if they find a long term problem then they strive to find a workable scientific solution. Instead climate scientists just strive to prove we are heading for disaster, but then wash their hands of any responsibility. They have been so successful that their careers and funding have prospered, while simultaneously over a trillion dollars have been wasted on the pipe dream of renewable energy. Green lobby groups and NGOs have prospered while a huge renewable industry has built profiting from the hype and siphoning off yet more tax and bill payers money.
Renewable energy (wind + solar) simply cannot sustain modern society. Any electrical engineer can tell you that. Unfortunately energy policy has been hijacked by arts and political science graduates all with an agenda. When put on the spot about the random unreliability of wind or solar they claim this will all be solved with cheap batteries and the ‘smart’ grid, as if they even understand the concept. Unfortunately Kryptonite batteries don’t exist . everything else violates basic thermodynamics.
So the choice for the future is simple. It is either a slow transition to nuclear power (eventually fusion) or a massive drop in population before a return to the 18th century living.