Climate Change - The Scientific Debate

[TheExcession]

turbobloke said:

TheExcession said:

How many times a day do weather stations measure temperature?

By that I mean - is a mean temperature for any one day calculated from a min/max reading or is it an average reading of several measurements taken each hour/minute?

When (if ever) did we move from a min-max to an average measured over the entire day/night?

I'd like to put aside any talk and mathematics from here for the moment, purely because I cannot find any reference to a document that says 'If you would like to contribute to the measuring of temperatures this is the equipment you need to purchase, this is how you get it certified and this is how you submit your data'.


What am I missing here?

I am willing to invest in a certified temperature recording data collector that will record and submit certified and calibrated temperature data to a certified data centre....

Where do I sign? What are the terms? What equipment do I need to purchase?

(In the past I worked for a Pharmaceutical company where I was checking balances 'weighing scales' to class E1 and F1, where are the classes for measuring temps?)

Ask Phil Jones, though he might have lost the info or deleted it, possibly.

Going back to that paper over in the Politics thread. You have 5% global coverage in part of your oceans dataset. So you use local means from the 5% to fill the 'gaps'. That'll be the gaps making up 95% of the data. Where there's more data, the standard error of the mean is ~0.5 deg C and often SEM > 0.5 and yet your final error bars are narrower than an amoeba's armpit and your result is in the second decimal away from zero. I've had postgrad work to assess in both science and social science over the years and I've seen some absolute carp in both, but nothing that would come close. I'm not convinced a half-decent 10-year old primary school student would consider it a plan, maybe they got 8-year-olds on work experience doing peer review.

Hey TB, I appreciate that reply to my post, but either I didn't quite ask the right question, or you answered a question that I likely didn't quite ask properly.

I'll try again - with no pressure on you to answer.

SO:

If a person decided that they wanted to monitor local temperatures on a day-by-day basis, with a view to contributing to some globally and rigorously monitored, calibrated and controlled dataset which could be paramount to policy forming decisions regarding climate issues.

How exactly would one proceed?

As I said, I've worked in Pharmaceuticals, so if anyone reading this is taking an ACE Inhibitor, maybe Zestril? - (sorry about the cough) I likely measured the milligram weights that were used in the assays for all the blood and urine samples that were tested.

Now, every single bit of kit we worked with during those trials was CALIBRATED.

I just want a heads up on the temperature kit that is used for measuring global warming.

Does anyone here actually measure global warming? I just want a quiet chat in the corner to understand what kit you are using....


Is it so much to ask?

[TheExcession]

turbobloke said:

Totally understood, your question was clear enough, I can only assume that Phil Jones hasn't got back to you as yet! I was still reeling from the claimed error in that oceans 'result'.

Meanwhile on the less than universal standard front, you may be interested in the couple or three insights available in these articles if you haven't already seen them.

http://wattsupwiththat.com/2010/03/09/accuracy-of-...

http://www.drroyspencer.com/2012/05/our-response-t...

and

turbobloke said:

And, on another aspect of your question, this article:

http://www.appinsys.com/globalwarming/GW_Part2_Glo...

It's interesting reading the links you have posted above. However, the CRITICAL issue I have is that they (the links) all seem to talk about errors and accuracy.

I'm utterly certain that we can indeed measure a temperature at a given time with a certain accuracy.

However I've often said that we can not measure sensibly to a 0.0x degree at a point in time.

Despite working for a major Pharmaceutical company where every thing was measured in milli and micro grams, I've also worked on a farm where fertilizer was measured in kilos per acre. If I was to go take a sample of soil it might show 100% above or 100% below the 'average' for the NPK optimum for that square hectare of soil.

I still await a guide for measuring temperature - I've got a guide for measuring in pharmaceutical terms, I've got a guide for measuring in farming terms.

So, not aimed at you TB => But aim me at the 'Measure Global Temperature Policy' document - if it is so f**king important then we'd all be wearing a measuring device - there'd be millions of smoothed data sets all showing - well I don't know what - because there aren't millions of data sets.

This is the science thread after all and yet there is nothing telling me how the data should be measured - just lots of reports of how the data massaged and twisted.

Polar ice big, polar ice low, winds and storms unprecedented, sea level up and down, flooding on flood plains higher than recorded utter utter GIGO.

And still - no one can tell ME how to measure global warming - Well I have a measure - my measure is that I haven't had the overnight heating on in my apartment this year. ERGO Global Warming surely.

Do you, anyone reading this, understand the problems with collecting this data?

I'm not actually convinced or interested in a GLOBAL temperature - because I'm simply not convinced that there is an agreed and calibrated method of recording local temperature. We can talk about spreading temp records over 5km squares or 10km or 100km or 1000km squares - just tell me how to record the temp over the 1metre square that I can monitor - and tell me what to do if they build a motorway past my house or an airport.... grhhhh

Edited by TheExcession on Wednesday 19th February 23:09

[TheExcession]

Variomatic said:

The requirements for WMO reporting stations, including instrumentation, are here:

https://www.wmo.int/pages/prog/gcos/documents/grua...

part 1 Chapter 1 (p. 15 on of the PDF) and Annex 1B of the part (p.33 on) refer to site and instrument requirements.

For temperature the technical requirements are briefly 0.1k resolution, with max error of o.1k over "normal" ranges, but with a note in the requirements that the achievable accuracy with typical sensors is only 0.2k - so outside the requirements.

There's also a warning in the text that an "inefficient" radiation screen (which includes one that's not correctly maintained) can introduce an error of up to 3 deg C, even if the station is otherwise well sited - which few of them are.

According to the consensus none of that - bad siting, bad design (including power supplies mounted \inside the thermometer radiation shields!), bad sensors, bad maintenance etc - matters because all the errors will average out.

Thanks for that - I'll take a read through it all. Again my point is how often should a measurement be made.

For example when I first made my post earlier on this thread all was quiet, now the wind has really picked up and the temperature has risen as the warmer winds are now blowing in from over the Atlantic.

Interesting that you say 'all errors will average out' - i think my point has been missed once again.

I'm not looking for errors, I'm looking for a methodology that accounts for my anecdotal evidence that temperature in my apartment can swing between shivering under the duvet to kicking it off to cool down, in the space of the hours from 12:00 to 08:00.

Earlier here there was little or no wind, right now it's picking up a bit, my guess, there's a front coming over. Temps have gone up a bit.

My problem is that I've regularly witnessed a frost that started around 3am and was gone by 7am when a front came over and dropped all its rain.

So what was the average temp? It did freeze during the night - but was warm and wet at bed time and warm and wet at wake up time?

Edited by TheExcession on Thursday 20th February 00:15

[TheExcession]

I'm asking these questions to try and provoke some thought about how this data is collected.

Simply put - Either I am (still) really lacking in any understanding of means/medians and temperature records or the current method is so flawed it beggars belief.

Please not earlier I was asking 'How to measure temperature', not 'What do I need to measure temperature'

Finally we might have an answer as to 'how' and possibly why it is flawed

Variomatic said:

The daily average is calculated as a simple mean of daily max and min.

Take a look



Now these numbers are made up - but I've regularly seen a frost at 03:00 or 04:00hrs that wasn't there at 07:00hrs.

My point is that SURELY the median approach to measuring temperature does not accurately reflect the actual average temperature.

Look at day 1 vs day 3 - Min and max the same - medians over the entire set are miles apart from medians over the min/max.

How is the method of min/max valid?

[TheExcession]

Jinx said:

And you don't see a problem HK?

I do - see above.

[TheExcession]

hairykrishna said:

TheExcession said:

My point is that SURELY the median approach to measuring temperature does not accurately reflect the actual average temperature.

Define "actual average temperature". Does the trend in this actual average over time meaningfully differ from the trend in (Tmax+Tmin)/2?

To be honest, I think you are probably correct, however it just doesn't sit right in my head.

The issue I'm trying to explore is that I don't believe (Tmax+Tmin)/2 is valid in terms of of assessing mean temperature.

(Tmax+Tmin)/2 is not an average temperature - it's a median temperature that tells us NOTHING about what the average temperature was.

And as I've shown above - means and medians are very different.

Let's think about another example. Imagine that I record my heart rate - twice a day - morning when I get out of bed and evening when I go to bed.

Both readings are valid, and I can calculate the min max.

Now let's consider I wear a heart rate monitor which measures my lowest heart rate at 60bpm and my highest bpm at 130 for the 20 minutes that I was playing a game of squash.

I only played squash for 20 minutes - my heart rate was high - off the chart but - but realistically would you consider my average heart rate to (Tmax+Tmin)/2?

[TheExcession]

Variomatic said:

Which means that (Tmax-Tmin)/2 is completely useless

This is the issue that I am concerned about, either we are are completely wrong in our understanding of means and medians, I've shown earlier how these numbers do not 'add up'.

Perhaps someone with better mathematical skills can explain why we are wrong and 'climate science' is correct.

Guam

[TheExcession]

Variomatic said:

People with mathematical skills would agree with us because, mathematically, a simple average from Tmax and Tmin can't tell you anything about the bits in the middle

But, again, "Climate mathematics" is a specialist branch that only climatologists are entitled to comment on. Cos they said so.

(Hope you don't mind the Edit.)

Relax.... chill out - let's try and get a few more ideas about what a 'mean' really is.

We can talk about regressions to that mean later.

I'm still concerned that the likes of McIntyre haven't pulled everyone up on this.

Means and Medians? What are we missing?

[TheExcession]

Variomatic said:

I think the problem is that there are just so many holes in the "settled science" that the people actively challenging it in the literature have to stick to one or two of them at a time. Besides which, the meaninglessness of averages (as used) is so straigtforward that it wouldn't warrant publication in a journal!

Well, I find my self completely disagreeing with you on this point!

There are no holes in settled science - honestly this boils down to a couple of really quite simple questions.

So,

At any station recording and contributing to recording temperature and submitting said data to a climate monitoring 'body'

Q1 - Is temperature reported as a mean or a median value?
Q2 - If measured as a median is that valid? - we know this
Q3 - If measured as a mean, how often are measurements made collated/reported?

[TheExcession]

SkepticSteve said:

I do apologise EX, but I have found your last few pages of protestation amusing.

You suddenly “falling-in” on this did put a smile on my face.

I think you will find that Guam, TB and others pointed this out about a 100,000 pages back.

However, I was just as astounded as you are now, back then when I “feel-in” on what/how the CRU team were Strangling this interesting, but limited temperature record.

We can’t blame those early gathers of this data, they had to use what they had.
The STUPIDITY, is ALL in Prof Jones lap! IMHO of course.

:;

I'm not suddenly "falling-in" and you are correct - it has been mentioned many times - it's one of my bugs to bear - right up there with claims of a 0.0x degree rise/fall/decline/incline (delete as appropriate).

Every single time I watch/read/hear a climate claim of unprecedented/record-high/record-low I simply despair at the fact the some people take this as gospel without even the slightest understanding of where the numbers are coming from.

The truth is out there I'm sure - and TB's mantra of "Any causal link to human activity?" is step one on my decision tree when I evaluate any data or report.

However - I'm angry that an entire industry and politic has been set up and is seemingly blooming over numbers that simply don't add up.

To me, it is insanity defined. Right there in front of us and so many simply accept it. O-level mathematics is all you require to see this.

hairykrishna said:

...when looking at long term trends, people have to work with what exists.

I'm 'pointing out' (that's a Tibetan Buddhist term btw) that what exists probably isn't good enough - and perhaps should be ignored.

Anyway - glad I made you smile

[TheExcession]

durbster said:

I read an interesting article about this recently which made the point that if Stephen Hawking had been the one to present the IPCC findings...

Ah, the old Brian Cox - David Bellamy gambit!

[TheExcession]

QuantumTokoloshi said:

I am still trying to get my head around his Fourier work, which is heavy going.

My old man is a Prof in cardiology, he was doing research into arterial wall sheer stress analysis, and whilst he will easily shame me on my math knowledge, he admitted he had to go talk to the other Math Professors in the University when it came to applying Fourier treatments, so I feel your pain.

(Mind you I can wipe the floor with him when it comes to CDMA )

[TheExcession]

Here's something I stumbled upon the other day.

CIA Special Collections Release In Full 2000. USSR: The Impact Of Recent Climate change on Grain Production

So a paper dated 1976 with not a hint of CO2 AGW stating weather changes and grain yields go up and down.

I'll just quote the first line (not for effect) 'During the past 15 years a significant fluctuation, which strongly aided grain production... may mark the end of a favourable climate trend....'

An interesting read!

[TheExcession]

Variomatic said:

(b) the whole "fraud" thing around climate science is frankly absurd.

Absurd because the 'fraud' is plainly there or absurd because there is no fraud?

The 'Climategate' email leak certainly pushed me off the fence on that one. From the top to the bottom it was corrupt. Everyone from high ranking politicians down to the lowest data analysts were implicated as basically not having a fking clue what they were doing, looking at, or recommending.

If you'd said "(b) the whole 'religious' thing around climate science is frankly absurd." I'd be right there with you. But hey, religion/fraud one word is longer but the meanings are very similar.



Just to add;
Great to see people discussing the science again with out the need for three full time mods working 24/7 trying to keep the everyone in order.

[TheExcession]

rovermorris999 said:

So why is temperature flatlining while 'carbon' is increasing?

Because it hasn't been adjusted yet?

[TheExcession]

Jinx said:

Space is cold. Very very cold. The sun sends a lot of energy at us that is absorbed by the oceans and the earth beneath our feet; which evaporates the ocean - providing nice warm H2O - and vibrates the earth's surface which warm the molecules of air around us keeping away the cold of space. The warm air at the surface is less dense than the colder air above so rises to allow colder air access to the warmth of the surface (convection) creating a "heat pump" that regulates the temperature - add more energy and the faster this pump will work (the atmosphere is adiabatic - change in internal energy is a function of work done against gravity).
It has been suggested that by slightly changing the composition of the gases in the atmosphere this will have a catastrophic change to the temperature at the surface - and by comparing the mid point of the tmin (minimum temperature in a 24 hour period) and tmax (maximum temperature in a 24 hour period) from a small number of sensors over time somehow has shown this to be true (when logic would suggest the actual mean temperature would need to be used for any comparisons and that the mean temp is a function of the amount of gaseous movement not energy in).
By using a static energy balance model with huge amounts of assumptions and little understanding of the triple point of H2O a small number of people have taken a very complex reality (and therefore pretty unknowable chaotic process - weather) mixed it with the precautionary principle and determined that capitalistic economic success based on easy to obtain high energy density raw materials is a danger to the entire planet and therefore must be stopped at all costs..... some of us disagree but get called names for it.

That is one of the finer summaries of the topic I've read.