Guest Post by Willis Eschenbach
The Berkeley Earth Surface Temperature (BEST) team is making a new global climate temperature record. Hopefully this will give us a better handle on what’s going on with the temperature.
BEST has put out a list of the four goals for their mathematical methods (algorithms). I like three of those goals a lot. One I’m not so fond of. Here are their goals:
1) Make it possible to exploit relatively short (e.g. a few years) or discontinuous station records. Rather than simply excluding all short records, we prefer to design a system that allow short records to be used with a low – but non‐zero – weighting whenever it is practical to do so.
2) Avoid gridding. All three major research groups currently rely on spatial gridding in their averaging algorithms. As a result, the effective averages may dependant on the choice of grid pattern and may be sensitive to effects such as the change in grid cell area with latitude. Our algorithms seek to eliminate explicit gridding entirely.
3) Place empirical homogenization on an equal footing with other averaging. We distinguish empirical homogenization from evidence‐based homogenization. Evidence‐based adjustments to records occur when secondary data and/or metadata is used to identify problems with a record and propose adjustments. By contrast, empirical homogenization is the process of comparing a record to its neighbors to detect undocumented discontinuities and other changes. This empirical process performs a kind of averaging as local outliers are replaced with the basic behavior of the local group. Rather than regarding empirical homogenization as a separate preprocessing step, we plan to incorporate empirical homogenization as a process that occurs simultaneously with the other averaging steps.
4) Provide uncertainty estimates for the full time series through all steps in the process.
Using short series, avoiding gridding, and uncertainty estimates are all great goals. But the whole question of “empirical homogenization” is fraught with hidden problems and traps for the unwary.
The first of these is that nature is essentially not homogeneous. It is pied and dappled, patched and plotted. It generally doesn’t move smoothly from one state to another, it moves abruptly. It tends to favor Zipf distributions, which are about as non-normal (i.e. non-Gaussian) as a distribution can get.
So I object to the way that the problem is conceptualized. The problem is not that the data requires “homogenization”, that’s a procedure for milk. The problem is that there are undocumented discontinuities or incorrect data entries. But homogenizing the data is not the answer to that.
This is particularly true since (if I understand what they’re saying) they have already told us how they plan to deal with discontinuities. The plan, which I’ve been pushing for some time now, is to simply break the series apart at the discontinuities and treat it at two separate series. And that’s a good plan. They say:
Data split: Each unique record was broken up into fragments having no gaps longer than 1 year. Each fragment was then treated as a separate record for filtering and merging. Note however that the number of stations is based on the number of unique locations, and not the number of record fragments.
So why would they deal with “empirical discontinuities” by adjusting them, and deal with other discontinuities in a totally different manner?
Next, I object to the plan that they will “incorporate empirical homogenization as a process that occurs simultaneously with the other averaging steps.” This will make it very difficult to back it out of the calculations to see what effect it has had. It will also hugely complicate the question of the estimation of error. For any step-wise process, it is crucial to separate the steps so the effect of each single step can be understood and evaluated.
Finally, let’s consider the nature of the “homogenization” process they propose. They describe it as a process whereby:
… local outliers are replaced with the basic behavior of the local group
There’s a number of problems with that.
First, temperatures generally follow a Zipf distribution (a distribution with a large excess of extreme values). As a result, what would definitely be “extreme outliers” in a Gaussian distribution are just another day in the life in a Zipf distribution. A very unusual and uncommon temperature in a Gaussian distribution may be a fairly common and mundane temperature in a Zipf distribution. If you pull those so-called outliers out of the dataset, or replace them with a local average, and you no longer have temperature data – you have Gaussian data. So you have to be real, real careful before you declare an outlier. I would certainly look at the distributions before and after “homogenization”, to see if the Zipf nature of the distribution has disappeared … and if so, I’d reconsider my algorithm.
Second, while there is a generally high correlation between temperature datasets out to 1200 km or so, that’s all that it is. A correlation. It is not a law. For any given station, there will often be nearby datasets that have very little correlation. In addition, for each of the highly correlated pairs, there will be a number of individual years where the variation in the two datasets is quite large. So despite high correlation, we cannot just assume that any record that disagrees with the “local group” is incorrect, as the BEST folks seem to be proposing.
Third, since nature itself is almost “anti-homogeneous”, full of abrupt changes and frequent odd occurrences and outliers, why would we want to “homogenize” a dataset at all? If we find data we know to be bad, throw it out. Don’t just replace it with some imaginary number that you think is somehow more homogeneous.
Fourth, although the temperature data is highly correlated out for a long distance, the same is not true of the trend. See my post on Alaskan trends regarding this question. Since the trends are not correlated, adjustment based on neighbors may well introduce a spurious trend. If the “basic behavior of the local group” is trending upwards, and the data being homogenized is trending horizontally, both may indeed be correct, and homogenization will destroy that …
Those are some of the problems with “homogenization” that I see. I’d start by naming it something else. It does not describe what we wish to do to the data. Nature is not homogenous, and neither should our dataset be homogeneous.
Then I’d use the local group, solely to locate unusual “outliers” or shifts in variance or average temperature.
But there’s no way I’d replace the putative “outliers” or shifts with the behavior of the “local group”. Why should I? If all you are doing is bringing the data in line with the average of the local group, why not just throw it out entirely and use the local average? What’s the advantage?
Instead, if I found such an actual anomaly or incorrect data point, I’d just throw out the bad data point, and break the original temperature record in two at that point, and consider it as two different records. Why average it with anything at all? That’s introducing extraneous information into a pristine dataset, what’s the point of that?
Lastly, a couple of issues with their quality control procedures. They say:
Local outlier filter: We tested for and flagged values that exceeded a locally determined empirical 99.9% threshold for normal climate variation in each record.
and
Regional filter: For each record, the 21 nearest neighbors having at least 5 years of record were located. These were used to estimate a normal pattern of seasonal climate variation. After adjusting for changes in latitude and altitude, each record was compared to its local normal pattern and 99.9% outliers were flagged.
Again, I’d be real, real cautious about these procedures. Since the value in both cases is “locally determined”, there will certainly not be a whole lot of data for analysis. Determination of the 99.9% exceedance level, based solely on a small dataset of Zipf-distributed data, will have huge error margins. Overall, what they propose seems like a procedure guaranteed to convert a Zipf dataset into a Gaussian dataset, and at that point all bets are off …
In addition, once the “normal pattern of seasonal climate variation” is established, how is one to determine what is a 99.9% outlier? The exact details of how this is done make a big difference. I’m not sure I see a clear and clean way to do it, particularly when the seasonal data has been “adjusted for changes in latitude and altitude”. That implies that they are not using anomalies but absolute values, and that always makes things stickier. But they don’t say how they plan to do it …
In closing, I bring all of this up, not to oppose the BEST crew or make them wrong or pick on errors, but to assist them in making their work bulletproof. I am overjoyed that they are doing what they are doing. I bring this up to make their product better by crowd-sourcing ideas and objections to how they plan to analyze the data.
Accordingly, I will ask the assistance of the moderators in politely removing any posts talking about whether BEST will or won’t come up with anything good, or of their motives, or whether the eventual product will be useful, or the preliminary results, or anything extraneous. Just paste in “Snipped – OT” to mark them, if you’d be so kind.
This thread is about how to do the temperature analysis properly, not whether to do it, or the doer’s motives, or whether it is worth doing. Those are all good questions, but not for this thread. Please take all of that to a general thread regarding BEST. This thread is about the mathematical analysis and transformation of the data, and nothing else.
w.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I have a question about annual global averages and the number of sites used for the calculations …
If year X has 1,000 data points and year Y has 800 data points can’t they generate an average for year X using the 1,000 data points and for year Y using the 800 data points or do some or all of the 200 “extra” data points in year X get dropped because they don’t have the same datasets ?
Since they are calculating an annual global average shouldn’t they only care about contiguous data within a single year ?
[“Snipped – OT” – see article body]
I too do not understand why the first objective is not to identify any trends in the many climate regions, by region, including micro-climates.
The concept of an overall global trend is kind of meaningless since it will wash out at lot of information.
I think that creating a database by Station (with gaps) that included a quality indicator with a algoithm reference to how the quality indicator was determined should be the first step. This would include hourly temp, humidity, wind speed and direction where available, where only min/max were recorded then only two hourly rows would exists the remaining 22 would exist but contain a NULL value (NULL does not mean zero is means No Data), the data should include a relative altitude to mean sea level.
Then add a fewof columns for each remote sensing device that has information for the station location, again with gaps and quality indicators and references, hourly.
Then the same again for the oceans, grid the oceans in 100km squares and use any data from any device with a set of columns per-device, for mobile devices use the data from when the mobile device was in or sensing the grid square, for bouys include the relative altitude from mean sea level.
It would be no less of an undertaking then the human genome project, but once completed, we would have a maintainable public database that could be used for many purposes.
It should become the one and only input source for any and all computer models, a common public frame of reference containg the totality of historical information.
It will be somewhere in the range of 2,680,927,920,000 rows.
Large but manageable, on par with the scale of a historical transaction database for a very large stock exchage.
Raw unadjusted data, all of it, in one place, all public, warts and all.
crosspatch says:
March 23, 2011 at 12:17 am
Yes, anyone that lives next to a body of water can attest. Also regional weather anomolies at the same site might be difficult to handle, such as a Chinook wind in southern Alberta where you could have -20C to + 20C in a matter of hours… it might seem impossible, but those who have experienced it will assure you it’s real.
Will,
they are doomed because they are starting with bad raw data … its that simple … you can’t fix that with math …
I think there’s a much deeper problem: BEST is doing the best they can, and we ought to be thanking them because it’s a job someone should have done long ago, but no one should expect to get clean data out of this.
On the contrary, by the time they get done their conclusion will, I predict, be that we don’t have valid data. Consider, for example, their decision to remove gridding because it adds imaginary data – that’s certainly true and therefore a good decision, right? Right, but absent gridding, there’s no large area applicability, so every researcher who wants to draw a conclusion applicable beyond the limits of each station is going to be adding some kind of gridding stand-in- and so making up data to cover what he doesn’t have.
Bottom line: if BEST demonstrates – by meeting about 3.5 of their four goals – that the best data is pretty useless for large area policy and planning purposes they’ll have done something very valuable, but it will be the opposite of what most of the people commenting here seem to expect – taking away a basis for conclusions instead of providing one.
Re: UHI: John Tofflemire says: March 23, 2011 at 10:10 am
With respect, I did not ask about the assertion. I asked about “requirements / specifications”. That would include the derivation for each location.
Willis,
I have to ask about the use of the Zipf Distribution.
A quick check of Google shows that you are about the only person who uses this distribution with respect to temperature (yet you seem very confident that it is THE distribution for temperature data). Typically this distribution is used to look at things like how often different words come up in language, how often cities of various sizes occur, how much traffic the 7th most popular web page gets compared to the 8th most popular.
This distribution is based on rankings of items. From Wikipedia on “Zipf’s Law”:
A similar statement for temperature would be something like:
Of course, the specific percentages would be different, but I don’t really see how a ranking of temperature frequencies will be an effective way to analyze the data. For one thing, it eliminates the sign of the deviation from the most common data (the mode).
“It is also deplorable that the BEST team seek to hide the adjustments they make, rather that letting the raw data stand”
BEST might take a page from the lessons learned in IT. The best practice in Data Warehousing is to never overwrite (update) or delete data. The only operation allowed in the warehouse is an insert operation.
Think of the old manual accounting ledger. You never cross out or erase any entry. Even when the entries are in error. You remove errors by inserting a reversal. You correct errors by inserting an adjustment. Each entry then becomes part of a history, leading from the current state of the data back to the original state of the data, providing a full audit of all changes.
As soon as you start deleting or updating entries in place, usually under the excuse of “saving space”, or “more efficient”, you have no way to audit the changes and you cannot rely on the results. You can never certify that your data is correct.
The mistake that CRO made, as revealed by ClimateGate was to adjust the data in place, and thus overwrite the original data. This effectively destroyed the audit of changes, so that no one can determine what was actually done. As a result, the CRU cannot certify their results as accurate.
Financial transactions have been done this way for years, allowing banks to certify that their financial data is correct. There are now large corporate initiatives underway to apply this same technique to non-financial data, to allow it to be certified as well. The typicall buzz word is Master Data Management. The aim is to make non-financial changes auditable by keeping a full history of all changes.
The beauty of this technique is that by maintaining a full history, you can always correct past errors. If you make an adjustment, and later you find out the method was faulty or the calculation incorrect, you simply insert a reversal today, effective the date of the error and the data is automatically corrected, with a full history still in place. Banks use this method all the time to correct past errors with minimal disruption to existing data.
BEST should apply the same techniques to control of their data. Otherwise the correctness of the data will always be open to dispute, no matter how good their mathematical approach. Data quality requires much more than good analytics. It requires fundamentally good data management techniques, including a full audit of all changes.
This is brilliant, Willis. Be sure you share it with the Berkeley team.
The purpose of the project is to try to determine to what extent the record has been skewed by air conditioners added next to the weather station, increasing use of asphault, and other aspects of the Urban Heat Island Effect.
That is a very difficult challenge indeed, and only a truly rigorous procedure has any hope of getting it right.
If the Berkely tem share ALL the raw data and calculations, we will have something useful. But if they screw up the math, it will get a reputation akin to Mann’s hockey stick.
Actually I think this problem should be approached from the other end. It would be better to have a hundred series that you trust than ten thousand that you don’t. So I recommend to search for data series which are well documented, check changes of measurement techniques over the duration of the record, average out or ignore obvious omissions (months when no data have been recorded) and arrive thereby at the best record for that station. Then you apply (or not) UHI corrections for that station. De Bilt in Holland is an example, where they have been doing this since 150 years or so. You can then also communicate with those stations if necessary, obtain copies of the original records when data are uncertain. If you find say one hundred stations like this with a reasonable spread over the globe you can draw useful conclusions about global temperature trends. It is of course a lot of work, but not more I think than the BEST approach. It would lead to an understanding of what has been happening around the world (before you apply statistical techniques) and could provide a basis to expand the system to recorded humidity levels as suggested in earlier comments.
Evert
Willis I think you’ve misunderstood empirical homogenization. Further, I’m not at all convinced that temperatures at a given location are universally described by a power law distribution, third I do not see the kind of discontinuous behavior that you speak of. 4th, the 1200km correlation distance is an integral part of the error calculation due to spatial sampling. There will be, I suspect, some refinement of that namely to account for the known variations in this figure due to latitude, season, and direction. Temperatures are correlated in space because the atmosphere is a flow, so its more than mere correlation.
All that aside, we know this from tests with synthetic data. Methods like theirs, perform vastly superior to the method you have preferred at times in the past : the first differences method. See jeffId’s destruction of that method back on the airvent
Hear! Hear! Dr. Curry! To be right-minded, researchers should be flattered when their efforts are questioned, critiqued, and improved upon. Important research MUST be questioned, duplicated, shown in other ways, and improved upon.
My thoughts: I believe that the effort to develop temperature data sets is still in its infancy. I am convinced that multiple broad regional sets made up of a running overlapping three month average in much the same way oceanic and atmospheric oscillations are presented would be another viable (and in my mind better) method. The single global monthly average hides way too much important information be it anthropogenic or natural, or both.
First time poster. Not a meteorologist, not a scientist, not a physicist, and my degree in Oceanography, earned at a small trade school on the Severn River in Maryland, is almost 40 years old.
That said, if my admittedly porous memory serves, somewhere I read or heard or was tested on the fact that the oceans are the ‘engines of our weather.’ If that observation is valid, then all this surface temperature analysis is akin to taking your own temperature when you have a cold. That temperature is but a symptom of something else that is forcing the temperature.
All we learn with surface temperature observations is that, yep, something significant is going on. We still aren’t any closer to understanding what is causing the temperature variations. Seems like a lot of wasted effort that might be better focused on understanding ocean currents and ocean temperature variation effects of atmospheric temps.
The point being that two towns (any region probably has their own) might be close by when one looks at a map but live in completely different climate zones. The purpose of explicitly mentioning the names of the towns was so people so inclined could look them up on a map and see how close they are. The difference being that there is a mountain range between San Mateo and Half Moon Bay and so the two are in completely different climate zones even though they are within a short drive of each other
Same with Truckee and Reno. One is an alpine climate that gets several feet of snow in the winter, the other is desert even though they are only 40 miles apart.
Jeff Carlson says”If year X has 1,000 data points and year Y has 800 data points can’t they generate an average for year X using the 1,000 data points and for year Y using the 800 data points
That won’t work. Having the same set of stations from year to year is crucial. The average max temperature across all weather stations in Australia for the first four available full decades are as follows:
1860-69 : 28.53 deg C
1870-79 : 29.07
1880-89 : 29.94
1890-99 : 30.94
Do you think that the average temperature in Australia really rose 2.4 deg C in 30 years? That’s 8 deg C per century (using the IPCC way of thinking).
[Data from http://www.bom.gov.au]
Rather then trying to homogenize the data – how about we start over with a detailed, calibrated surface temperature sensor network? Place sensors to measure what needs to be measured, instead of by other considerations, and calibrate to satellite data.
At some point the finite, definable cost of that exercise would be less then the continuing efforts to integrate bad data. We could then draw a line, reference all of the old data (proxy and sensors) with error as a historical value (and fight over that forever). The data could be evaluated over the coming decades to determine if the “signal” of carbon emissions exists, or solar influences, ocean cycles, and land-use changes are predominant. The density of the network would determine its accuracy and cost. If this is really important, the effort should be politically feasible.
Might be a way out for the politicians who have been hoodwinked into being AGW proponents – get better data and study it some more is a time honored way for politicians to kick the can down the road.
I’d urge an accounting-transaction style approach to the dataset. Explained more fully here. That way, for a given station/day/time, the ‘layers’ of raw data, adjustments etc can be represented as separate transactions, with source and other categorisations – date/time stamps, process and version used, type of adjustment and so on – added to ensure traceability. That way, even if there are kludges and mash-ups, they show up as separate transactions in the station/day/time ‘account’ and can be aggregated across all stations, excluded with a line of SQL magic, or filtered out in a cube approach. Transparency and workability, please!
Surely these weather stations have kept track of other weather data than just temperature? How about BEST include ALL the data and meta data from all the weather station records? What information has been recorded by these stations? Is it even the right data needed to make sense of the climate?
Does it even make sense to “average” temperature data into one number for the whole planet. I’ve never seen any explanation of that that makes sense. Doesn’t it make more sense to monitor each station’s local temperatures and look for trends in that? Some will stay the same, some will have upward trends and some will have downwards trends. Having an average of the “anomaly” data doesn’t even make sense, at least I’ve never seen any good explanations as to why it makes any sense.
In computer science we deal with discrete data sets all the time. Averaging or “homogenizing” data aka “fabricating data” just doesn’t seem wise as that just continues the statistical games that have been played. Somehow a discrete method must be developed that incorporates error ranges and uncertainties intact throughout the computations so that outputs also show the combined error ranges and uncertainties for any computed data. Computed data should be clearly marked as such and should NEVER be presented in a graph in a way to imply that it’s anything but computed data.
Bottom line it won’t be the best BEST if it only uses temperature data alone, it seems that Surface Air Moist Enthalpy should also be computed.
How can climate science advance when it’s stuck with limited methods of measuring Nature (e.g. Average Surface Temperature) when more accurate methods exist (e.g. Surface Air Moist Enthalpy)?
just checked Romm’s (is he really not Roem after all?) site…they are incensed and worked up about something that Mr Watts might have done. If we could harness their negative energy to power our nations, the problem is solved. They seem completely beyond the bounds of reason. i feel I should feel sorry for them and try to lock them up somewhere. they obviously cannot handle real life.
Sorry, that was my first dip into climate progress: it is like dealing with a horde of hysterical lunatics.
Like a few of the other posters have mentioned the question which is trying to be answered is that of energy budget. Temperature, while correlated for sure is a pretty poor proxy for energy, especially without pressure and humidity levels. All of the effort should be going towards ocean heat content and proxies there of (SST). In addition to it being largely free of urban heat and other contaminating and confounding effects it actually is the direct measurement of the question which needs answering. The top few meters or so of ocean contains as much heat energy as the entire atmosphere. In addition the variance in ocean temps is way way lower than atmospheric surface temps so the underlying data is much more stable and fewer measurements will give far far better approximations to reality.
It seems bizarre that we can easily get daily satellite temp data, and yet the OHC data from the argo floats is available only with great difficulty. Just business as usual in the world of climate ‘science’.
Willis, thanks for a thought provoking post, once again.
I’m with the “don’t change the data group “(which seems to be almost everyone). If the data is questionable, drop it from the analysis.
I, like lots of others on this thread, have an anecdote about real weather producing a temperature reading that would be questioned. A few years ago, here in San Antonio, Tx ,we had a 100 degree F high one day in February. Of course it was an all time record high for February. Temperatures were fairly normal on either side of that day. If one was looking at the temperature records, one would probably question the validity of that reading. Yet it was real.
Regarding data distributions. I too have always wondered about the general assumption that the data have a Gaussian distribution. I was not familiar with the Zipf distribution. (Thanks to Tim Folkerts, and someone else whom I could not find again in the thread, for explaining what it is.) My favorite expression for distributions is the Weibull distribution, since the expression for a Weibull distribution includes a shape factor that will cause it’s shape to vary from exponential to normal to log-normal.
And on the subject of distributions, why do we always use the mean of the distribution for the “average?” It would seem to me that the median (50% above, 50% below) would be more descriptive of temperature behavior than the arithmetic mean.
One last comment. I hope that Dr. Curry is right. I hope that BEST is just the being.
Many have suggested that instead of discarding outliers they simply include this in the error bars. I don’t see how we can even calculate error bars given the nature of local temperatures that Willis points out here and the limitations of the data set. Can someone give me a primer in one paragraph or less on how we can even quantify error given the limitations of the data? If it is possible what are the assumptions included in that error calculation?
It seems to me that we needed to know our methodology before we began collecting data. Obviously, that is not the case with the historical dataset.
It’s the trend of change in temperature at each measurement location that matters, not the average temperature over some area, and therefore not the trend of the change in average temperature.
The methodology should be to compute the trend of temperature changes over time at each measurement site, then multiply the number of risers by the average rise and the number of decliners by the average decline. That would provide a far better synopsis of the trend.
Consider, for example, a stock market: Which methodology would be a better measure of what the market is doing as a whole? 1) Computing the mean price of all stocks every N units of time, and then plotting the trend of that average price over time? Or 2) Computing the trend of each stock price individually over time, and multiplying the number of rising stocks multiplied by the average percentage increase, and multiplying the number of declining stocks by the average percentage decrease?
Heat capacity of the immediate environment at each station is a major player. Global average temperature may not be a meaningful concept, as indicated by Essex & al., 2007.
Journal of Non-Equilibrium Thermodynamics. Volume 32, Issue 1, Pages 1–27
ISSN (Print) 0340-0204
DOI: 10.1515/JNETDY.2007.001
February 2007
Does a Global Temperature Exist?
Christopher Essex, Ross McKitrick & Bjarne Andresen
However, heat content of the entire climate system or of its subsystems, being an extensive quantity, is meaningful. Now, average of temperatures over the subset of stations with a high heat capacity environment (basically the ones close to a large water body) is a proxy to heat content, because heat capacity of water is the same everywhere and is much larger than that of any other substance in the climate system.
Therefore it would be an interesting exercise to compare seaside temperature trends to inland trends. As the latter ones lack any obvious physical meaning, if trends of averages differ significantly for the two subsets, that would cast serious doubt on any global trend.
Willis, you are dead right.
The BEST effort is worthy and should be praised and encouraged. However, unless they are magicians they will not be able to overcome the fundamental problem which is that the source data is just too lousy for the job. So I fear the BEST team may have created a rod for their own backs. If their results show less warming than the current official data their methodology will be roundly criticised by the warmists. If their results confirm existing estimates, the skeptics will be all over them with objections.
For me the irony of all of this is that it is irrelevant anyway. What people seem to be blind to is that, even using the official world temperature data from NASA or Hadley, THERE IS NO DANGEROUS GLOBAL WARMING ANYWAY.
Here is a plot of the fficial HadCRUT3 data straight from the University of East Anglia of Climategate fame:
http://www.thetruthaboutclimatechange.org/temps.png
Perhaps somebody could explain to me why a long term trend of 0.41degC per century is an alarming problem? The 30 year period from 1970 to 2000 looks to me likely to be just the upswing of the well documented ~67 year oceanic AMO temperature cycle, which shows up clearly in the 11 year running mean line (red). During the upswing section of any roughly sinusoidal variation, the slope is going to be 4 or 5 times steeper than the long term average – hey, that’s the nature of a sinusoidal variation! So it is surely no coincidence that climate alarmism flourished during that 30 year period when Hansen et. al. believed fervently that they were witnessing the clear and alarming signature of man-made global warming. Now that we are moving over the top of the natural cycle and into the downswing, they are having an increasingly tough time maintaining their position.
Reworking the world’s temperature data is most unlikely to resolve this issue. But the next 10 to 15 years will surely resolve it one way or the other just by observing whether or not the current apparent downturn is maintained. While we wait patiently for the outcome, perhaps we should take on board Ronald Reagan’s famous maxim: “Don’t just do something, stand there”.