From the Cliff Mass Weather Blog
Cliff Mass
Yesterday, the temperature at SeaTac Airport rose to 89F, beating the all-time daily record for the date (86F)!
Today, even if the temperature is the same, it won’t break any records.
How could this be?
It turns out that not all records are equal.
Let me explain.
Below is a plot of observed temperatures (blue bars), average temperature range (brown band), and record highs (red) and lows (blue) for June at Boeing Field in Seattle.
Look carefully at the record highs. A large variation of record highs in June from 81F to 104F! Generally around 90F.
.png){kind=link}
So why the variability?
It turns out that to get the really high temperatures in western Washington, the atmosphere needs to organize itself in a very specific way, generally with a strong upper-level ridge and offshore-directed flow at low levels.
And on some days, by the luck of the draw, the atmosphere gets the right setup for maximal heat. After many years, the needed flow pattern occurs, and the temperature climbs to record levels.
Global warming plays very little role in these records—the key is getting the right atmospheric flow situation.
Would you like me to prove this to you?
Here is a plot of the highest maximum temperature each year between June 7 and June 21 since 1950 at Olympic Airport (a less urbanized location than SeaTac). No apparent upward trend.
.png){kind=link}
Turning back to SeaTac, below is a plot of the highest temperature on June 14th at that location (below).
2025 was the warmest year on record for that date. But note! There is no trend in the record temperatures for that date over the entire period of record.
In fact, the highest temperatures on June 14th are trending down! (brown line).
.png){kind=link}
So the global warming claims for the origin of such records should be taken with a large grain of salt.
“It turns out that not all records are equal.”
Well, yes. I think the enthusiasm for records for the calendar day are pretty much a US thing. And as shown, flaky.
Yes. Daily temperatures in a specific location are an extremely noisy metric.
That’s why climate scientists focus more on long term trends across large geographic areas rather than isolated calendar day records.
They provide a much more reliable estimate of the underlying climate signal.
Deniers, on the other hand, like to cherry pick this noise because the large scale, long term trend is far less favorable to their position.
“19 Mar 2026A California desert community tied the highest March temperature ever recorded in the U.S., amid a record-breaking winter heat wave in the Southwest.’
https://www.cbsnews.com/news/north-shore-california-ties-march-temperature-record-southwest-heat-wave/
You mean those lying, denying bastards? 😉
As I said:
“That’s why climate scientists focus more on long term trends across large geographic areas rather than isolated calendar day records.”
The word “more” does not imply that one metric completely disregards the other.
Rather, it indicates that one metric receives greater emphasis than the other. The second metric is still considered and given attention, as the evidence in your linked source demonstrates.
You were talking about deniers. Deniers using simple things like the earliest temperature record in a month. That means the various new outlets that spruik that rot, are the deniers. 😉
Gottcha. Long term trends are what is to be used except when point statistics are more convenient; like when the long term trend shows no trend as in the above article.
‘ever’ in how long?
I say the Medieval Warm Period when Vikings farmed southwest Greenland was warmer than today and climate was stable.
Oh! you say _March_ and a _single day_. Think….
As the human body does not have one temperature, neither does a large geographical area, particularly land.
Catch up – the Northern Hemisphere does not have the same climate as the Southern Hemisphere.
Carbon dioxide does not control the global climate, many other things are the primary causes of changes such as precession of the equinox, alteration in the elliptical orbit around the sun, distance of earth from sun as altered by the huge planets Jupiter, Neptune and Uranus with moons and periodic changes in the magnetosphere of the sun modulating the cosmic ray nucleation of ice crystals forming clouds. There is even present in stratigraphic studies for the passage of the solar system through one of the arms of the Milky Way.
And, btw, the overall warming has stopped and the cooling period has already commenced.
I accepted it as a fact that climate changes from my youth, eighty years ago and probably before you were born
“neither does a large geographical area, particularly land.”
Large geographic regions have average temperatures that can be measured, calculated, and monitored over time.
“Catch up – the Northern Hemisphere does not have the same climate as the Southern Hemisphere.”
I did not say there was.
“Carbon dioxide does not control the global climate, many other things are the primary causes of changes such as precession of the equinox, alteration in the elliptical orbit around the sun, distance of earth from sun as altered by the huge planets Jupiter, Neptune and Uranus with moons”
Precession mainly changes when seasons occur relative to Earth’s position in its orbit. So its climatic effects are strongest at regional and seasonal scales. There isn’t a large effect on global mean insolation.
It’s also operating on timescales that are MUCH SLOWER than the RAPID WARMING we’re currently observing, just like the other variables you mentioned.
“And, btw, the overall warming has stopped and the cooling period has already commenced.”
No, it hasn’t. The longterm UAH trend is currently about 0.16°C/decade.
Back in 2002, I saved a cooling prediction by John Daly:
http://www.john-daly.com/press/press-01a.htm#iceman
Since then, there have been countless predictions of imminent cooling. Virtually all of them have either failed outright or are performing very poorly against observed temperatures. Given this track record, it’s ironic that climate skeptics often accuse mainstream climate science of making poor predictions.
GHGs are expected to remain a persistent warming influence, and the temperature record continues to reflect that.
You can’t “measure” an “average temperature”. period.
Yes, to paraphrase Feyman –
“if it’s a concocted average, it ain’t science”
In fact, 2002 was cooler than 1998, and temperatures cooled through 2015.
See the UAH chart.
Meanwhile, the Liars at NASA and NOAA were claiming 10 years between 1998 and 2015 were hotter than 1998.
You have a lot of confidence in a cyclical temperature movement stopping and becoming a continuous warming trend.
You should study temperature history. It doesn’t say what you think it says.
No, they don’t. This is a climate science myth. Temperatures across a large geographic region depends heavily on geography and terrain which can vary considerably across a large region. An extreme example is the large area encompassed by the area including Pikes Peak, Colorado Springs, and the High Plains area east of Colorado. A large area encompassing mountains, foothills, and flatland. Widely varying temperatures on a daily, weekly, monthly, and annual base for both the area as a whole as well as in smaller parts of the area.
Even in a smaller region, say NE Kansas, you encounter geography and terrain differences such as river valleys, surrounding river plateaus, prairie, and transition forested areas. Large differences in things like humidity, elevation, evapotranspiration, etc. Meaning a wide variation in temperature over any period of time, both regionally and locally.
The average temperatures in NE Kansas vary widely, depending significantly on weather patterns in the SE US, Texas/Oklahoma, plus Nebraska and the Dakotas. Those weather patterns have their own dependencies. The variance in “average” temperatures can be easily seen by tracking grain harvest/acre figures from year to year.
This climate science myth perfectly illustrates the lack of physical science used by climate science. Averaging intensive properties of different things simply doesn’t work. You cannot build a larger system by combining an intensive property of different things. It’s like trying to find an average temperature of 100 water baths in 100 separate laboratories where each water bath has a different heating method, a different size, a different control system, etc. That “average” can certainly be mathematically calculated. But it has no physical meaning in the real world. You can’t use that average value to predict anything about the phantom larger system. The variance of the average for the phantom larger system can be significant since the variances of all 100 variables adds.
What you call “noise” is not noise. It is the result of trying to define a larger system by combining an intensive property.
Then how can you combine their temperatures into a larger system? Climate science doesn’t even weight the temperatures based on their variances.
Seasonality causes significant stationarity problems in the time series for the global mean insolation and its effects *are* different in each hemisphere because of the overall makeup of ocean/land. The globe is not a homogenous, isothermal sphere.
The “rapid” warming you are seeing is not any different than the rapid warming seen over the millenias. It’s just happening while *you* are alive. It’s a form of the argumentative fallacy known as Argument by Age. I give you the 1920’s/1930’s as a prime example. That rapid warming ended. You’ve provided no physical reason why this rapid warming won’t end as well.
This violates radiative physics. If at time T0 the earth radiates 100 joules and CO2 returns half of that then the earth will have still cooled by 50 joules. That’s not “warming”, it’s “cooling”, just at a slower rate. Radiative physics is a *time* function and must be analyzed over time, not at a single instant. The only way for CO2 to WARM the earth is if it is a source providing heat rather than acting as a reflector. Even then, CO2 would have to send back at least 100 joules in order to just maintain a stagnant temperature. It would have to send back *more* than 100 joules in order to warm the Earth, i.e. CO2 in the atmosphere would have to be warmer than the earth.
Yes, the Pacific North East region of the north American continent, which is Prof Mass’ area of specialized meteorology and climates, in no way lends itself to being “averaged” with equivalently sized areas in that part of the globe, but are East of the Rocky Mountain range.
“Averaging” of probity-poor temperature numbers is abject nonsense.
Only in climate science, can you take the average of bad data in order to create good data.
“can you take the average of bad data in order to create good data.”
The Olympia, WA temperature records that Cliff Mass is using are averaged and aggregated into the final dataset.
As a result, it is unclear why there has been no pushback on the assertion that Cliff Mass is using poor quality data.
(Actually, it becomes clearer when I consider the context of the blog.)
The article is about OBSERVED highest temperatures reached on particular days.
Not derived constructs.
So, comparisons of singly recorded numbers.
“The article is about OBSERVED highest temperatures reached on particular days.”
I know, but MarkW is one of the people who changed that conversation’s context.
I’m just pointing out his implication about Cliff Mass and his data, and how people don’t challenge it.
Re the warming of the 1920’s and 1930’s Hubert Lamb notes in his book ‘Climate , History and the Modern World’
“In Spitsbergen the open season for shipping at the coal port lengthened from three months in the years before 1920 to over seven months of the year by the late 1930s.”
A study of worldwide glacier length fluctuations in the journal Cryosphere 8 p659 – 672 (2014) found that
“despite increased global temperature in the 20th Century this retreat is strongest in the period 1921-1960 rather than the last period 1961 – 2000”
‘A data set of worldwide glacier length fluctuations’ PW Leclerq, J Oerlemons, HJ Basagic, I Busheva, AJ Cook and R Le BRIO
It’s *ALL* cyclical. And you can be fooled if the observation period is not long enough to encompass the cycle with the longest period (i.e. the lowest frequency). In fact, while Nyquist only implies that the sampling frequency be high enough to resolve a single cycle, that’s only for a perfectly repeating cycle. If the period of the cycle is variable you may need to look at more than one cycle to resolve the differences.
Every word in your response to Eldrosion is correct and to the point and, unfortunately, wasted.
“No, they don’t. This is a climate science myth. Temperatures across a large geographic region depends heavily on geography and terrain which can vary considerably across a large region.”
Yes, we’ve already seen from your blah blah blah that you don’t understand how temperature anomalies are intended to be operate.
“What you call “noise” is not noise. It is the result of trying to define a larger system by combining an intensive property.”
No, it’s noise because temperature records from a specific location are affected by numerous confounding variables.
For Washington specifically, the influence of the nearby Pacific Ocean and its various long term oscillatory patterns can introduce substantial variability into its local temperature records, making them a poor proxy for global warming trends.
“Seasonality causes significant stationarity problems in the time series for the global mean insolation and its effects *are* different in each hemisphere because of the overall makeup of ocean/land. The globe is not a homogenous, isothermal sphere.”
Which has nothing to do with the original argument of whether it is more currently more important in contemporary warming than carbon dioxide.
“This violates radiative physics. If at time T0 the earth radiates 100 joules and CO2 returns half of that then the earth will have still cooled by 50 joules. That’s not “warming”, it’s “cooling”, just at a slower rate.”
The key issue is flux equilibrium.
Thank you for proving Phil R’s point.
Incorrect!
Those confounding variables are called influence quantities that affect the measurement uncertainty. They have standard uncertainties determined from either statistical analysis (Type A) or other documentation (Type B). They are added together to obtain a combined uncertainty.
Please learn what an uncertainty budget actually is.
Do I need to give you the GUM definitions?
“Do I need to give you the GUM definitions?”
No, because I don’t dispute anything you’re saying. What I was referring to is completely different from measurement uncertainty.
You just made my point! Ignoring measurement uncertainty is ignoring the most fundamental quality of what you are analyzing. If you don’t acknowledge that it exists and include as a part of any and all analyses, then you are not doing science, you are doing advocacy.
I understand perfectly how they work. Anomalies inherit the variance of the components of used in calculating the anomalies. The variances of the components ADD. Since the variance is a metric for the uncertainty of the value, that means the anomalies are MORE uncertain than the individual components.
Anomalies created at the top of Pikes Peak versus anomalies created for Colorado Springs have different variances. Statistically you can’t “average” those anomalies without weighting them properly.
The anomaly with the smallest variance is more accurate than the anomaly with the largest variance. Why would you average the two by giving the less certain result equal weight with the more certain result?
Averaging anomalies of intensive property values for different things is NO different than averaging the intensive property absolute values. Neither the absolute values or the anomalies can be combined into a physically meaningful larger system. If you can’t combine them into a larger system then the average is physically meaningless. You can calculate it but it is still physically meaningless. Look again at those 100 water baths at 100 different laboratories with 100 different heating elements, 100 different control systems, 100 different measuring devices, etc. Averaging the absolute values given for the intrinsic property known as temperature for those 100 water baths will tell you NOTHING that is physically meaningful. If the average changes how do you identify what caused it to change from trending the statistical descriptor known as the average? Using the anomaly is no different. The anomaly is just a linear transformation of the data using a constant. In other words it is just scaling the data along the x-axis. It doesn’t change anything except the position on the x-axis. You *still* can’t tell from the average what is going on with the data. In other words the average of the anomalies remains physically meaningless.
Noise and variance are two entirely different things. If you have a function f(a,b,c,d,e) and a,b,c,d,and e are confounding variables, i.e. they are factors used in determining the value of “f” then they are *NOT* noise, they are part of the signal. You can’t just throw away the variance in “f” caused by the component values of a, b, c, d, and e.
Even shot noise in the electronics of a modern measuring system is a part of the signal, it is included in the statement of the measurement uncertainty of value: estimated value +/- measurement uncertainty. You don’t know *what* that contribution of the shot noise is, it is random and changes from moment to moment – and it is UNKNOWN. Unknowns get included in the measurement uncertainty budget.
When making multiple measurements of the same thing under repeatable conditions the variance of the data values makes up the measurement uncertainty associated with the best estimate of the measurand. That variance includes any random effects such as reading of an analog scale, e..g parallax, needle width, lighting, etc. YOU CANNOT CANCEL THOSE RANDOM EFFECTS BECAUSE YOU DON’T KNOW WHAT THEY ARE! They are included in the variance of the data.
And your assertion does *NOT* change the fact that you are trying to average intensive property values. A physically impossible thing to do. Just because you can do the calculation for an average doesn’t mean that it is physically meaningful!
So what? Each time you take a measurement of that Pacific Ocean you are measuring a different thing – different salinity, different density, different temperature, different contaminates, etc. Those factors have to be accounted for in the measurement uncertainty part of each measurement – what you call “variability”.
The influence of the Rocky Mountains introduce substantial variability in the temperatures in Colorado Springs. It makes the temperatures on the east side of the mountains substantially different than the temperatures on the west side and the mid-range temperatures on each side don’t tell you much about the climate on each side because of rainfall differences.
It has a LOT to do with it! One more time, variance is a measure of uncertainty. Estimated values with less uncertainty should be given more weight in an average than estimated values with more uncertainty. CLIMATE SCIENCE DOES NOT DO THIS. It can’t be fixed by spatial gridding. It can’t be fixed by anomalies. The variances tell you whether or not you can even identify the small changes introduced by CO2. If the differences you are trying to find are smaller than the measurement uncertainty interval then you don’t really know if you have identified a difference or not. And when you calculate the difference of two values THEIR UNCERTAINTIES ADD! It doesn’t matter if you are adding them or subtracting them, they ADD.
The measurement uncertainty of a baseline average calculated from mid-range daily temperatures is huge because each measurement included adds measurement uncertainty. Just finding the average of two annual values with measurement uncertainties of +/- 1C gives a total uncertainty of +/- 1.4C or 2C depending on how you add them. That means you don’t know if year1 is different than year2 unless the difference is greater than 1.4C! Now do the same thing for a decade! (hint: +/- 3C)
Bullshite! Since incoming flux occurs over an interval of 12 hours and outgoing flux occurs over an interval of 24 hours you better hope the flux values never balance! Earth would become an ice ball of it radiates for 24 hours with the same flux as the sun insolation over 12 hours!
The equilibrium that is important is JOULES-IN and JOULES-OUT! And this has to be calculated over an interval of time long enough to encompass the cyclical thermodynamics of the system – at least centuries if not longer. A decade won’t do it. Three decades won’t do it. Heat entering the ocean may not show up as Joules-out for a LONG TIME. And once you know the total joules-in and joules-out there isn’t any reason to normalize the values to the same time interval. The balance will be shown by the joules-in and joules-out values!
Climate science today is a mish-mash of garbage using assumptions like “all measurement uncertainty is random, Gaussian, and cancels” or “incoming flux and outgoing flux from the system must be in equilibrium” or “averaging the intensive properties of different things is physically meaningful” or “you don’t need to weight temperature data based on the variance” or “anomalies don’t inherit the variance of the parent components”. I could go on and on with these. you would think that the real scientists today involved in climate science would be aghast at the assumptions climate science makes. But that doesn’t seem to happen much. I guess money talks.
The averages also have a large variance that you never quote. That means the “average” temperature can vary significantly over time.
“…Large geographic regions have average temperatures that can be measured, calculated, and monitored over time…”
Learn a little thermodynamics. “Average Temperature” is an arithmetic construct with no physical basis. Completely meaningless even if you could accurately compute it.
Thermodynamics! This team can’t handle Arithmetic.
Truly a persuasive and rigorously substantiated argument, and completely free of rhetoric!
It’s clear that you and your ilk EXPECT GHG’s to drive warming.
Spare us the repetition, please.
eldrosion comment – “It’s also operating on timescales that are MUCH SLOWER than the RAPID WARMING we’re currently observing, just like the other variables you mentioned.”
Eldrosion – you comment is one of the most dishonest talking point in the “climate science” arena. The resolution of the paleo data is way to low to make that determination either positively or negative.
“The resolution of the paleo data is way to low to make that determination either positively or negative.”
Not all paleo data:
https://www.carbonbrief.org/factcheck-what-greenland-ice-cores-say-about-past-and-present-climate-change/
What a load of gibberish. Your carbon briefs are full of BS. !
If it is unreasonable to obtain seasonal temperatures there is no way to estimate monthly or annual temperature.
Your graph is exactly what Micael Mann did. Stick current temperatures to proxy estimations whose time resolution is far lower.
https://www.nature.com/articles/s41586-024-08181-7
It’s true that the temporal resolution of ice core records are lower than that of instrumental temperature records.
But the instrumental record can be temporally smoothed so that each data point represents the same averaging interval as the proxy record (in this case, 20 years). From the CB article:
“The six ice core sites used by the reconstruction are shown in the figure below.
The temperature reconstruction produced using data from all six ice cores is shown by the blue line in the figure below, and spans the period from 9690BC to AD1970. It has a resolution of around 20 years, meaning that each data point represents the average temperature of the surrounding 20 years. So, the end of the record – 1970 – shows the average temperature between 1960 and 1980.”
Why didn’t you show this graph from the paper you gave.
Why do you always deflect?
I am not deflecting. I am showing that this paper you sourced determined that we are not currently near the warmth that occurred in the past. Any conclusion that warming is now unprecedented is false.
That’s not what the CB article concluded. Here’s what it actually said:
“While periods during the early Holocene – 7,000-11,000 years ago – may have been warmer in Greenland than the present day, if the present rate of warming continues, the Earth should pass well beyond any temperatures experienced in Greenland during the Holocene by 2050.”
See that word “if”. That is not a scientific description of what WILL happen. It is a word like might, maybe, perhaps, possible, that is, a declaration that no one knows what will occur. The paper doesn’t even give a probability estimate of the conclusion happening, and neither do you.
Science requires a hypothesis that has a mathematical relationship that can be falsified. Using weasel words in the conclusion is nothing more than the guy standing on the street corner with a sandwich board predicting the earth will go poof tomorrow.
How can you be sure that the mechanical mixing of melted ice/snow, and diffusion of gases through the firn doesn’t vary over the last 12,000 years so that 20-years is too small?
If they do, it’s highly doubtful their influence is significant.
The proxies in the CB article already reconstruct the Holocene Thermal Maximum earlier in the record (physically consistent with higher Arctic insolation anomalies) followed by a cooling trend and then the hockey stick pattern of instrumental warming, which is well constrained.
So it seems like everything is accounted for.
Here is a far more realaistic temperature for Greenland since 1850
How exactly is that “far more realistic”?
Because you say it is?
And the Greenland GISP data showing that the Arctic is still very much on the cold side of the last 10,000 years.
That little uptick at the end, according to Mickey Mann, starts in 1900, ends in 1940.
Data from all around the region shows that the 1940 was as warm or warmer than the first 20 years of this century.
Interesting that the “Years Before Present” label on the x-axis has disappeared from your chart.
Otherwise, readers would be able to see that the record only extends to about 95 years BP (1855).
Was that omission merely a coincidence?
When I spent a month in 1966 in Greenland supervising the last annual closure survey in the ice tunnel at Camp Tuto, I observed that the top of the ‘ice’ at the terminus was slush for at least a meter. There is no way to know if during the last 12,000 years that was typical or not. And you expect me to believe one can resolve temperatures with something like annual averages. We probably can’t even resolve reliable annual averages let alone seasonal variations unless one can be certain that the sample represents ice than never melted during the Holocene Optimum.
I guess you didn’t get the memo about RCP 8.5
HAW HAW HAW HAW HAW HAW HAW,
you liked that dishonest chart because you don’t understand the error it shows they are grafting a higher resolution data set onto lower resolution set and you are using the discredited RCP 8.5 scenario that has been dropped by the IPPC because it is pseudoscience embarrassment.
There is no climate change in Greenland it was polar 12,000 years ago it is still polar today.
mmmm. you’d need enough data across the region (number of readings evenly distributed by area), to get an average.
The Puget Sound area is quite variable because of ocean water and mountains and the Fraser Valley which cold air comes through in winter. (Its mountains include the Olympics which some winds blow around and the Cascades with volcanic peaks like Baker and Rainier.)
He doesn’t know about the convergence zone either.
“Large geographic regions have average temperatures that can be measured, calculated, and monitored over time.” Using how many data points? And the other problem is that then these few data points are “averaged” and then assumed to be the “average” temperature of that large land mass, which as another commenter has pointed out, could contain substantially different geographical features and essentially be non-homogeneous. Take the Palm Springs area as an example. You can start down in town at 100+ degrees, but take the tram up to 8500 feet and the temperature will definitely NOT be 100+ degrees. But you would say that the “average” temperature would be the sum of the two temperature measurements divided by 2?
Especially daily maximums at single locations, which have a very high signal-to-noise ratio.
Anyone seriously looking for a warming signal would use long term, regional monthly averages.
Here’s NOAA’s 1950-present June average temperature record for Washington State with linear trend. The warming is right there.
Why not start in the 1940s 😉
Because the article specifies a 1950 start date for June temps at Olympic Airport.
What a cop-out, and where are Washington temperatures measured. !
Please show us the weather site this data comes from.
You can’t trust NOAA. They have an agenda.
That would be an opinion.
You can’t trust Tom Abbott. He has an agenda.
And you don’t 😉
One of the things I learned in calculus and in engineering is to evaluate a signal in piece parts. Here is a graph that I just evaluated by eye and separated it into three pieces. It suggests that somethings occurred to vary the baseline of temperatures. That is what you should be including in your analysis.
I also learned in business that linear regressions are the bane of trying to analyze underlying conditions. In this case, using your graph, there is no way to analyze whether Tmax is rising, Tmin is rising, or some combination of both. In other words, it useless to draw conclusions from it. For example, a +0.3°F/decade increase tells you nothing about what is actually occurring.
That would already be taken into account by the NOAA algorithms.
Was evaluating charts by eye something you learned in calculus?
“Was evaluating charts by eye something you learned in calculus?”
Exactly what has NOAA defined as the cause of the baseline jump? I can’t find it anywhere.
Evaluating charts by eye is a PRIME example of what engineers do when confronted by a data distribution. E.g. You can calculate an average for a multi-modal distribution without every realizing it is multi-modal. Identifying it as a multi-modal distribution by eye, however, can be done in many cases. The average of a multi-modal distribution tells you little about the physical reality the data represents – so the “eye identification” is a big clue to investigate further.
Identifying the piece-parts of a distribution by eye is exactly the same. And as Jim points out, you need to evaluate WHY the piece-parts happen as well as what happens in each piece part. A part of the data analysis you simply gloss over by using an end-to-end linear regression line. You *miss* one of the biggest needs for further analysis – why the different piece-parts exist. Did a piece fly off of a fly-wheel? Did a drive belt break? Did a volcano happen? Did a stagnant pressure system occur? Did a mud-dauber wasp build a nest in the air intake of the measuring system?
And yes, much of this was learned in developing engineering calculus skills – by learning how to “see” changing slopes in signals. Calculus skills were taught in all 4.5 years of my EE training. ALL 4.5 years. Not just in the first three semesters of basic engineering calculus. I would point out that in the 60’s and 70’s when I learned all this we didn’t have digital data acquisition systems and computerized graphing capabilities. You learned to identify what was going on by watching an oscilloscope screen or a line trace on a strip-chart recorder. Calculus is the study of slopes and learning to identify slope changes *is* part of calculus. There is no reason why this can’t be done by eye as a first approximation.
When I see statements like yours I know immediately I am talking to a statistician or mathematician with little real world experience.
I had an outdoor booth using a canopy this past weekend at a renaissance fair. We had a major wind storm come through just before dark. After the storm I could *LOOK* at the slope of the canopy roof at each corner and immediately identify there was a problem and where it existed. I didn’t need to write an equation for the elevation of each roof section and do a first derivative to figure out the slope of one was different than the other three. I estimated that first derivative BY EYE!
Jeeesh! I am just flabbergasted sometimes by the criticisms leveled by so-called climate scientists.
It shows what you want to see, so you ASSUME it must have been properly created.
Look up “integrating by parts”. The idea is what is important here, not the integration itself.
And yes, when dealing with waveforms, sometimes it is useful to break it up into pieces to analyze what components cause different parts of the waveform.
Your ability to analyze complex problems is lacking.
That’s what computers are for! 🙂 And why younger academics are often wrong. Their crutches are the wrong size.
Evaluating charts is something you learn to do when dealing with REALITY. !!
And I also learned to estimate by orders of magnitude when good empirical data were not available.
Yep, 2016 El Nino step. ! It is in the USCRN data too.
Incidentally, that jump is the ONLY warming in the whole of the USCRN data (we haven’t the info yet to see what final affect the recent El Nino will have)
There appears to be a step function imposed on the June temperatures, resulting in a net upward change. There is no reason to believe that CO2 is causing a step change. In other words, we don’t know why it is changing like it is, but it doesn’t seem to be what the consensus claims.
“The warming is right there.”
What warming are you speaking of? Trending averages only tells you the trend of a statistical descriptor. It doesn’t tell you what is happening with the data itself. Where is the warming happening geographically? Where is the warming happening timewise? Are these really averages of base data or averages of averages that aren’t averages but daily mid-point temperatures?
The +0.3F per decade linear warming since 1950. Says it on the chart.
What utter nonsense. Every global temperature data producer adds trends to average temperatures (more often to anomalies, but the trend would be the same either way).
Washington State. Says it on the chart.
The month of June, between June 1950 and June 2025. Says it on the chart.
That would be a mix, I imagine, with more recent data using automated hourly averages and the older data daily =(Tmax+Tmin)/2, averaged over the month then averaged between all the reporting stations in Washington State.
WHAT IS WARMING? What object on the earth is warming? What part of that object is warming? Where on the temperature profile of the object is the warming occurring?
Total malarky! First off, you aren’t using AVERAGE temperatures, you are using daily mid-range temperatures which cannot, by definition, define climate. Second, it doesn’t matter who or how many people are trending averages, IT IS STILL TRENDING A STATISTICAL DESCRIPTOR. And trending a statistical descriptor tells you nothing about what is actually happening to the data itself!
If you have ten data points where the first five are 10 and the second five are 20 you will get the same value for the statistical descriptor known as the average if the first five become 20 and the second five become 10. Yet the data itself will have changed significantly. Trend the average and it will have a slope of 0 (zero). Trend the data and the slope will change from positive to negative!
Washington state has SIGNIFICANTLY different geography and terrain across the state. That means significantly different temperatures, precipitation, clouds, etc. EXACTLY WHERE in Washington state is the warming occurring? At the coast? In the river valleys? In the mountains ranges (since they have multiple ranges which ones?)?
June is a seasonal transition month from late spring to early summer. WHEN in the month of June is the warming happening? Is the growing season lengthening? Is it getting shorter?
In other words you aren’t even trending the same thing!!!!!
Your answers are a prime indicator of the issue of trending averages.
You can’t tell me a single thing about what the actual data is doing. Which means you have no way to actually evaluate anything concerning a change in climate like the length of the growing season or the amount of precipitation in June over the years. You can’t even locate where geographically any changes might be happening!
It’s garbage in-garbage out from the beginning to the end.
Yep.
My layman example of the folly of averaging critical conditions is the car that is required to have tire pressures of 40 psi, but the vehicle about to be used for a cross-continent trip has 20, 60, 30 and 50 psi respectively in its 4 tires.
Average pressure then in each tire is 40 psi.
Just as the manual says.
So totally good to go for that 60 mph dash across the continent?
That “warming” is not real. Period.
Yes, this isn’t real:
Arctic sea ice since 2005.
And of course current levels are far higher than for most of the last 10,000 years.
So you have no idea how NOAA creates their data or where it is measured.. Just take it as is.
… how not-sciency !
I’m sure that the warming religiously observes the state lines.
For someone who is quick to disparage others (“utter nonsense”) I’m not impressed with your demonstrated thinking.
This is a classic misleading chart because it doesn’t show WHEN it is warming being a 60-year Washingtonian, I have observed that most of the warming is at NIGHT especially in the cold part of the year.
“This is a classic misleading chart because it doesn’t show WHEN it is warming”
If someone wants to separate daytime vs nighttime trends, NOAA maintains diurnal temperature records as well. And they are publicly available.
El Drosian, you are assuming that it is constant over time, which it isn’t. Albeit, generally the warming is greater at night and in the Winter.
http://wattsupwiththat.com/2015/08/11/an-analysis-of-best-data-for-the-question-is-earth-warming-or-cooling/
I have seen and known about it for many years, your reply fell flat on its face.
There is NO climate change happening in my state as the climate zones are still the same now as it was 10,000 years ago after it slid into the interglacial period we are in the waning days of.
Doesn’t indicate where it was measured either.
Jim has shown the El Nino step in 2016..
… not much other warming, but what there is could easily be from urban or airport growth effects.
There is no underlying climate signal.
Indeed. Nothing much happens between 1950 and 2010. Most of the positive gradient is driven by what’s happened from the 2015 spike onwards. Some people are obsessed with straight lines.
Spikes in the data, if not corrected for, continue to appear in the trend line for a long time. It pulls on a linear trend line until future values can overcome its contribution.
Trending a statistical descriptor only tells you what the descriptor is doing. The trend line doesn’t give you a lot of knowledge of what the data distribution itself is doing. If you have 10 data points from time T0 where the first five are very high and the next five are very low, at time T1 they can flip to low/high and the statistical descriptor known as the mean won’t change. You start trending the “mean” and things look stagnant. So what does the trend line actually tell you about reality? And it gets even worse when you start trending averages of averages of averages where the base average isn’t even really an average but a mid-point value.
“Spikes in the data, if not corrected for, continue to appear in the trend line for a long time. “
That is why the alarmista just love El Nino events. 🙂
We could take UAH and “correct” for the spike+ step of the major El Nino events..
But that would totally destroy the “human caused warming” myth.
Both CoPilot and Grok discuss that UAH does not decay the shocks properly. Grok says it is up to the data users to address them. That is exactly what you have done, albeit in a slightly gross manner. A GOOD model could fill in the blanks, Lol.
Bob Tisdale showed that the El Nino events caused a step up in adjacent oceans 6 month after the main event.
You are misusing the term “noisy”. Noise is an extraneous signal that is similar to but separate from desired signal. The noise signal is added/subtracted to the desired signal.
The appropriate description that you should be using is VARIANCE. The signal you are examining has no noise, it has a high variance.
Describing the signal using a noise description is another way of justifying modifying the time series to obtain a desired value. That may not be the case here, but in general climate science does this constantly with “bias” corrections and homogenization.
This is also wrong. Noise is the correct term to use for the unpredictable component of a datum point in time series data, such as an exceptionally warm or cold June in any given year.
Variance is the mathematical spread of the typical deviation from average.
Sorry, what you are describing is statistical noise where a MODEL does not or can not predict a data point. You are trying to defend that your linear regression is the correct model to use to predict measured temperatures. Temperatures that aren’t predicted by the MODEL are therefore “noise” artifacts.
Good luck on convincing folks that you know what you are talking about.
Sorry, but if your model can’t predict real physical measurements, then the problem is in your model.
You really need to avail yourself of Internet sources concerning time series analysis.
Metrological noise is extraneous information superimposed on an expected signal. It is part and parcel of measurement uncertainty.
It is only unpredictable if one does not know what is causing the exceptional temperature(s). Obviously, a well-mixed gas like CO2 would not be expected to produce episodic spikes in temperature.
Noise by any other name is just as malodorous. It is the component of a waveform that is not of interest and interferes with extracting the signal of interest.
Jim, to be fair, if noise is present, and has a large amplitude, it will increase the variance. Typically, noise is random, like radio interference from lightning. However, something like a 60 Hz bleed- over from a power supply can be an annoyance that can make it difficult to extract a weak signal and will be considered as noise by most.
In image processing, it is often seen as a pattern in the image. I’ve found that an efficient way of dealing with it is to do a Principle Components transform, and when dealing with multispectral data, find the component that obviously shows the periodic pattern. Do an FFT of that and suppress the 60 Hz pattern. Then do an inverse PC transform, substituting the filtered Principle Component for the original. The result is a suppression of the periodic ‘noise’ in all the bands, which results in cleaner, sharper images when combined to make natural or false-color composites. Something that I have found interesting is that while the higher component images typically have more ‘noise,’ sometimes there are indistinct patterns that the eye can see that represent geochemical targets for exploration geologists!
That is true if the noise is superimposed onto the expected signal. As you say, lightning has frequency components that become superimposed on an AM signal. That does increase the apparent variance. However, measurements of temperature should include an uncertainty in measurement that covers the normal effects that introduce measurement noise. Things like wind, shade, drift, housing, etc. These are described in an uncertainty budget and allow one to know the possible values due to that noise. It is an admission that one can not extract the noise components and arrive at a true value.
Statistical noise is when the model you have created is unable to predict the actual physical values that have been measured. If you are confident of your measurements, including uncertainty bars, yet your model won’t come close, the fault is the model and not the measured values.
It is why linear regression of a time series of temperature explains nothing. Time is not part of a temperature functional relationship. Therefore, time will never predict accurate values of temperature. Calling that failure a result of noise is an incorrect conclusion.
This isn’t really directed to you, I just wanted other folks to have an explanation that is meaningful.
“Typically, noise is random, like radio interference from lightning. However, something like a 60 Hz bleed- over from a power supply can be an annoyance that can make it difficult to extract a weak signal and will be considered as noise by most.”
You’ve just defined the difference between random uncertainty and systematic uncertainty. Random uncertainty is almost impossible to remove because it *is* random. It’s why you use the standard deviation of the observations as the measurement uncertainty instead of just assuming it all is Gaussian and cancels. Systematic uncertainty is impossible to remove if you don’t know what it is. E.g. if your tape measure is off by an inch how do you remove that if you don’t know it is off by an inch?
The best noise blanker I ever found in a shortwave radio was one designed to remove the entire signal when it’s amplitude or rise-time surpassed a critical point that could be adjusted. It got rid of random noise like lightning by just blocking the amplifier chain (required the use of a delay line on the signal path). You lost both the noise *and* the signal. The ear/brain filter would “fill” in enough to make sense from the remaining signal before and after the blanking period.
There isn’t any reason climate science couldn’t do the same thing with modern temperature data that consists of more than a Tmax and Tmin observation. A perfect example is a measuring station that has been moved to a new microclimate, just delete the old data and start fresh. If that one station is 1 out of 1000, so that you wind up with 999 data points from the past instead of 1000, it isn’t going to make a hill of beans difference in a properly handled statistical analysis.
‘Deniers’? You mean the apocalyptic mentality that keeps bleating despite none of their dire predictions having come true in 3/4 of a century? Zero, nil, nada, NONE. :-o)
Not only a US thing but a U.K. thing too.
Also a UN thing.
Oh, yeah! Nick doesn’t know what he is talking about.
Don’t sell him short. He knows exactly what he is doing when he tries to throw you off the scent trail.
Mr. Spencer: You are on the scent!
Not sure of your point but it’s not “pretty much a US thing”. It’s a climate histerics and climate science (but I repeat myself) thing. It’s the warmers who glom onto and scream apocalypse every time there is a June or July high temp record. A couple years a go I saw an alarmist post about a temperature record of something like 54°C in Turpan, China. And boy, did it generate a lot of histerical climate change scare comments.
If curious, I’ll let you figure out what was wrong with this “record.”
Funny how, whenever a particular, frequently used metric, is shown to be faulty, Nick’s response is always the same: Nobody uses that.
Chosing a particular day at a particular location makes the record meaningless.
There are hundreds of such records ( hottest, coldest, wettest, driest…..) every day around the world.
It is pure sensationalism and statistically meaningless.
Today my brother hadthe least money is his pocket for any June 16th in his entire life.
Does that tell us anything about his financial “climate” in general? Of course not.
So who does it? Data, please!
As always. Warm temperature records = climate change; global warming. Cold record temperatures, well that’s just weather. Never fails. I appreciate the facts and data presented by Cliff.
Looking at one date certainly tells you nothing about the presence, or not, of a warming climate.
Such records throw up all sorts of weirdness. In the Central England Temperature series of nearly 400 years, there was one day in June (I think) that stood alone having never reached 90F until a couple of years ago.
There is still one day in the middle of June that is a clear 2C colder on average than all the rest. This is nothing but pure chance. Clearly if the record is maintained for long enough, it should ‘catch up’.
What is sure is that currently the overall maximum envelope is being nudged up year after year.
One day in June 1975 saw a cricket match in Derbyshire abandoned not because of rain but because of snow.
And if I remember correctly. at that time the grave concern was global cooling and the coming ice age. Ahh, memories…
Yes – and through a slightly different intervening mechanism, THAT was supposedly due to human activities, specifically industrialization and the use of fossil fuels. Sounds familiar.
When the “environmental catastrophe” du jour about faces and yet the supposed “cause” and supposed “solutions” all remain the same, you know you’re being conned.
The solution for every problem always involves de-industrialization in the west and more power for the socialists in government.
I was in Buxton that day and was planning on going to the game, the weather just before had been warm and the snow storm moving through was quite a shock! That was on the 2nd June, by the 6th a heatwave had started leading to a long hot summer which was followed by the hot drought of 1976.
But nudged up in steps. I have not seen a good explanation (or working hypothesis) for why it is in steps when it is being attributed to a well-mixed gas that is obviously varying smoothly in concentration because of biological agents.
From the article: “Global warming plays very little role in these records”
Don’t you mean CO2? That should be what you mean because there are lots of sources for global warming. You should be more specific.
A Request to Dr. Cliff Mass and the climate research staff at the University of Washington:
Climate activists say that American leadership in reaching Net Zero by 2050 is crucial for convincing other nations, especially India and China, to quickly reduce their own carbon emissions.
As these climate activists state the problem, the world as a whole will not reach Net Zero unless American leadership shows the way.
Here in Washington state, we are under a legislated mandate from the Climate Commitment Act (CCA) to achieve the following carbon emission reduction goals:
— By 2020, a reduction to 90,500,000 metric tons (1990 levels)
— By 2030, a reduction to 50,000,000 metric tons (45 percent below 1990 levels)
— By 2040, a reduction to 27,000,000 metric tons (70 percent below 1990 levels)
— By 2050, a reduction to 5,000,000 metric tons (95 percent below 1990 levels)
These targets are unambiguous mandates codified into law. They are not in any way aspirational. Since no progress has been made in achieving the 2030 target of a 45% reduction below 1990 levels, something has to give as 2030 approaches.
The state either abandons the CCA targets formally or informally, or else the state imposes a draconian fossil fuel rationing plan on the state’s citizens and on our economy. An honest discussion concerning energy policy is therefore in order, and that discussion must include debate concerning the validity of today’s mainstream climate science.
Suppose we assume for purposes of discussion that the United States reaches Net Zero carbon emissions by 2050, and that most of the world’s other nations including India and China follow suit and reach Net Zero by 2075.
What is needed for purposes of an honest and informed debate concerning US energy policy is a set of climate model runs which embody mainstream climate science’s current thinking about the impacts of CO2 on the earth’s climate, and which predict:
1 — The concentration of CO2 in the earth’s atmosphere in the years 2075, 2100, 2125, and 2150 if the United States reaches Net Zero by 2050 and the entire world reaches Net Zero by 2075.
2 — The earth’s global mean temperature in the years 2075, 2100, 2125, and 2150 if the United States reaches Net Zero by 2050 and the world reaches Net Zero by 2075.
The University of Washington does extensive climate research using its access to mainstream climate modeling codes.
The UW climate modeling staff could produce the necessary pair of climate model runs for each of the years 2075, 2100, 2025, and 2150.
Furthermore, and just as important, the UW staff could document the assumptions and the uncertainties associated with those particular model runs in ways which would serve to greatly increase the transparency of today’s mainstream climate science thinking.
Dr. Mass, how about it? Are you game for this proposal?
You missed the third option. The state ignores the CCA targets (either formally or informally) because…the state.
I missed nothing. For the state government to ignore the targets is to abandon them, for all practical purposes. Even if that’s being done in direct violation of an unambiguous law.
Sorry, no offense intended. Comment was supposed to be a little tongue-in-cheek.
The state of New York is currently being sued for not meeting similar commitments. I fully expect this to happen to Washington State as well. The whole idea is BS but it happens to be law. We can only hope some future voter initiative over turns it.
Isn’t their recent efforts to drive the productive out of the state part of their Net Zero strategy?
I make the point to the local climate activists that the jet airliners Boeing manufactures in the Puget Sound area, and in South Carolina, are responsible for roughly one percent of the world’s total carbon emissions.
And so I ask them why they aren’t putting pressure on Boeing to produce a hydrogen fueled airliner; and if Boeing chooses not to do that, to lose all the tax incentives and all the other perks they enjoy in Washington state by keeping their manufacturing operations here.
Their response is always a blank deer-in-the-headlights kind of look.
What, you are expecting them to think?
Temperature is a piss poor metric for climate. I’m not sure why it is a main focus of climate science. Precipitation is the main factor in making crop species and variations decisions, not temperature. Temperature only defines the boundaries of what is possible. Precipitation is the main factor determining what survives. Miami and Las Vegas may have the similar high temperatures but vastly different climates because of the difference in precipitation as well the difference in soil types. When is Freeman Dyson’s criticism of the climate models going to be taken seriously by climate science? The climate models need to be VASTLY more holistic to define climate and climate change properly. There is a reason why we aren’t seeing the mass starvation, climate-forced migration, species extinction, etc that the model outputs predict by looking only at temperature.
The original climate research was to understand the climate, both natural and anthropogenic.
That got rudely changed to determine how temperature rises with rising CO2 levels.
The rest of the insanity is history.
Climate model:
CO2 is input.
IR is transfer function.
Temperature is output.
That is not an energy system.
Ah…never for $1,000, Alex.
I guess this would have been better as a “Carnac the Magnificent” joke.
Holds envelope to head: “The answer is Never.”
Opens envelope: When is Freeman Dyson’s criticism of the climate models going to be taken seriously by climate science?
Humor – a difficult concept.
— Lt. Saavik
Maybe because it has proven more difficult for the models to predict precipitation than temperature, albeit the temperatures tend to run warm. As I understand it, the precipitation forecasts for climate are a lot like weather forecasts for precipitation. Therefore, they emphasize their strong points.
So climate science just takes the easy way out and creates climate models that don’t really tell you about climate? That’s pretty much where climate science is today.
Tis a variable region with ocean and Olympic Mountains to the west and varying direction of wind hitting them, Cascade mountains to the east.
I recall 84F a few times in the 90s.
Abbotsford BC might be a rough comparison to check.
(Direction of wind hitting the Olympics creates different conditions north-south in the area, for example Stevens Pass at 3500 ft ASL can be warmer than Snoqualmie Pass at roughly 2000 ft ASL. Makes a difference to XC skiers.
It is pure sensationalism and statistically meaningless.
Today my brother had the least money is his pocket for any June 16th in his entire life.
Does that tell us anything about his financial “climate” in general? Of course not.