Diurnal radiation, three UK sites compared

Posted: August 30, 2014 by tchannon in Analysis, radiative theory, Surfacestation, weather

This article is cross posted from Tim’s blog as of interest to some Talkshop readers with a few extra sentences likely to raise discussion. 

August 2014 there was a meteorological gift of both exceptional conditions and good data. What can be learnt?

Three Met Office sites showed a signature of exponential cooling. This requires clear sky and a calm. Given somewhat limited parameter hourly data the following shows the commonality. The computed terminal conditions are shown later in this article.

Image

Benson and Santon Downham data has been normalised to Katesbridge[3], which has the least noisy data or the three.

Achieving a close overlay requires taking earth rotation into account, dusk and dawn move relatively both by geographic location and the peculiar movement throughout the year as night length changes, these do not move together [1]. Fractional delay (less that the sample period) was used to equalize diurnal time. (see the two blog articles here)

Dusk appears to be the important factor, a surprising finding, I assume cooling is time from dusk, dawn terminates cooling.

General information, under essentially calm conditions wind drops for a period during the night then reappears just after dawn. (not shown here)

Temperature normalisation defined is for the cold period, not as accurate for Benson where the better site exposure (more open) led to more wind at times.

Lapse rate reduces Katesbridge by about 0.5C relative to Santon Downham which is close to sea level. Normalisation will include this in compensation. (91m vs. 6m)

I expected dawn to be the arbiter of time shift but I found a larger shift. Initially my reaction was site solar exposure at low angles: Katesbridge is surround by low hills; Santon Downham is effectively a large clearing in a forest. I know the Benson site and that is not far off the middle of an RAF base, very open even though urban housing is encroaching. [typo fixed, said Newton Downham, thanks Stephen Richards –tim]

3658 Benson England
51.620266 -1.098695 56 Bing Google
99014 Katesbridge Northern Ireland
54.296694 -6.111149 88 Bing Google
99118 Santon Downham England
52.459201 0.671145 7 Bing Google

Benson tipped my thoughts. The time seems to follow time from sunset. More data is needed before this is a clear matter, a large work, would need automating.

Wind

Benson is a fairly complete Met Office site, the other two are minor.

Image

Here we can see there is a certain amount of wind and how it affects the temperature. Visibility reduces as it gets cold, lot of humidity limiting cooling. Humidity sensors are notoriously very poor, one of the most difficult to measure parameters, Benson one is fairly good but will be slow responding. (dew point is shown)

This site includes “weather” information. Benson is a manned RAF, Police rotary wing base (possibly air ambulance too), so the weather data might be human logged.

Completeness

Image

Image

Above from all station PDF  created from Met Office Datapoint service under OGL [4]

Conclusion one

Given no wind there is a common cooling profile between met stations. This has been shown from real data, not theory.

Terminal state and radiative figures

Image

Katesbridge data is fairly clean so a good exponential match was possible, therefore a fair guess at the terminal temperature if dawn was delayed. There is a “flat” on the data at 0.0C. This together with another outlier were removed for law match purposes.

For over a year I have been collecting high resolution data from the Chilbolton Observatory, Hampshire [2]. This includes pyranometer (short wave including visible) and pyrgeometer (longer wave) instruments. More recently full duplex instruments data became available, currently used to sanity check the earlier, which it does.

By visual inspection of the hundreds of day plots the radiation values for the kind of climate we have and under high humidity clear sky conditions at the temperatures involved here allowed me to place a guess on the outgoing radiative loss.

Image

An example day which contains a variety of factors. Was a frost, probably melts leading to the rain gauge tip (one at 12 hours is normal). Essentially calm during the night, dawn flat calm, odd time of day. Pressure was falling, cloud and water was moving in through the day. Plot labelled IR shows the two figures, with the difference the heat lost by radiation, see net_IR_out, a computed figure.

The same exercise was done for Santon Downham and Benson (on raw data, not normalised)

Image

The data is noisier, outliers were removed.

Image

Benson data is somewhat noisy but wind data is available. I suspect the terminal should be a little lower.

Conclusion two

Given suitable meteorological conditions estimates of final cooling and outgoing long wave is feasible. Whether this is useful is moot.

At terminal conditions there is steady state and ought to mean the Stevenson screen is at equilibrium.

The dominant heat loss factor apart from ground temperature is water vapour and droplets in the atmosphere. Probably primarily reflection. Colder clouds still manage to kick back essentially 100%  of outgoing although it is a fog at IR and localisation of sources is not done, probably impossible. CO2 is far below anything detectable here.

More useful is showing the commonality between significantly different sites.

Santon Downham

Sounds like Sandtown might be an origin

The soil is light sand, and in 1688 in consequence of the sand being blown by continuous high winds from the hills of Lakenheath (distant 5 miles south-west), several houses were buried and destroyed. The river was choked up. One farmer of the period; named Wright, had all the approaches to his residence so blocked up that in order to reach it he had to pass over a wall eight feet high. The sand at one time filled his yard and reached to the edges of his roof.

http://www.santondownham.org/history1900.html

Sandy soils seem to lead to a higher diurnal temperature range, Santon Downton appearing often in UK maximum and minimum figures, probably more often regionally. (I don’t collect them)

Norwich Airport is another with a light soil and that too is warm.

Quoted has a clue for this station, the former RAF / USAF Lakenheath is close by.


1. The Solex ephemeris program was used here to find the time of dusk and dawn for a particular geographic location.

2. Articles using Chilbolton data can be found on Tallbloke’s Talkshop

3. Belfast Telegraph brief article http://www.belfasttelegraph.co.uk/news/local-national/northern-ireland/katesbridge-the-legend-of-northern-irelands-frosty-hollows-29014781.html

4. PDF of all stations, file here (PDF 1.4MB)
Numeric data imported to portable XLS, file here (XLS 1MB)

5. Original article here (is almost the same)

Comments
  1. tchannon says:

    Oh dear, I think I’ve used a wrong word in the title. Should have been nocturnal radiation. Changing it after publish tends to break links. Life I suppose.

  2. Richard111 says:

    A personal observation.
    “”I expected dawn to be the arbiter of time shift but I found a larger shift. Initially my reaction was site solar exposure at low angles:””
    Last year I placed a 3mm thick steel plate, half square metre in area, on a larger 4 inch thick polystyrene sheet. This was raised about a foot above ground level on trestles. A sensor on the end of a six foot lead from a digital thermometer was bolted to the centre of the steel plate. this was arranged on the west side of my house which NEVER sees dawn sunlight. Over several clear calm dawns the thermometer recorded up to 2C temperature rise BEFORE air temperature. The temperature increase appeared to start as soon as the sky overhead started to turn blue.
    This is an easily repeated experiment.

  3. Tim,
    How do these folk distinguish between radiative and convective heat transfer?

  4. ren says:

    Escape the heat also depends on the temperature at an altitude of about 3500 m. This is associated with the reach of the jet stream. This in turn depends on the circulation in the lower stratosphere.
    http://earth.nullschool.net/#current/wind/isobaric/700hPa/overlay=temp/orthographic=27.38,92.91,365
    http://earth.nullschool.net/#current/wind/isobaric/250hPa/orthographic=27.38,92.91,365
    http://earth.nullschool.net/#current/wind/isobaric/70hPa/orthographic=27.38,92.91,365

  5. Stephen Richards says:

    Newton Downham? Santo Downham is pure sand with some sparse heathland plants. It is a “frost hollow”.

  6. tchannon says:

    Richard111, yes dawn seems to have some surprising properties. Dusk might too.

    Maybe there is enough data now to look and see. I have access to data at an exact location with a time resolution of better than 30 seconds.

    That leaves trying to find suitable weather conditions or less ideally by normality given the extreme volume of data (gets awkward with samples every few seconds over months or years)

    As I imagine it the high atmosphere will intercept solar so something radiates. My first guess is ozone, which has access right through the atmosphere.

    If the A/D is dithered, will be by accident, some extreme detail can be extracted. (how come I was looking at uC and nC some time ago, was not set up for kelvin, and discovered the international standards are broken, no-one sane uses thermocouples like that, I iz crazy).

    I expect someone will be along to remind us about why the wind does that.
    Also the nocturnal jet problem.
    And the matter of temperature inversions.

    I’ve been looking for obvious evidence of positive heat IR flow from atmosphere to ground, nothing so far but the instruments are awful. Looking at dawn might allow coherent detection.

  7. tchannon says:

    Will,
    “How do these folk distinguish between radiative and convective heat transfer?”

    Which folks? The article is my own work, independent from anyone else and hence mistakes, finger trouble.

    If you look at the Chilbolton plot, after dawn it goes into convective water vapour regime, all goes vague and very noisy. There are more instruments than those I show, not decoded, its a lot of work. There is nothing stopping someone else also using the source.

    Most of this is wheel reinvention yet for me that is how to get a true understanding and open up the possibility of a different take or making a new path.

  8. tchannon says:

    Stephen Richards, thanks, I missed that typo. Goodness knows why I originally wrote Newton in several instances.

  9. Richard111 says:

    Tim, that experiment I conducted as described above was to confirm, in my mind, that the blue sky at dawn is scattered light, not re-radiated energy from the atmosphere. The metal plate has a very dark grey surface which I assume might approximate a ‘black body’. It certainly warmed up, but not by the same amount each time. The slightest sign of haze in the air and the plate wouldn’t warm up.
    I very ‘iffy’ experiment, but fun. We are due to have some fine weather next week courtesy of ex-hurricane Cristobal according to the BBC so I will repeat the experiment and see what happens. The forecast is high pressure over the UK so hopefully no wind.

  10. tchannon says:

    No dawn glitch I’ve seen in the IR. This suggests a close check on short IR/visible against exact sun position might be worthwhile.

  11. tchannon says:

    Richard it looks like a synchronous experiment is possible if you are located close enough. I think we had better retire to email to discuss that. (publishing where people live is unwise)

    A snag is that I can’t produce data until the following month: after the Chilbolton data is made available. There are a host of instruments there with full resolution data, typically sampled every 10 seconds.
    eg.
    :title = “Chilbolton Pyrheliometer” ;
    :history = “Recorded using Microlink 3000 series DAQ\n”,

    http://www.microlink.co.uk/3000.html

    Checking their timing is accurate might be fun. If you record accurate timing then allowing for geographic location can be made.

    All weather permitting. If you can get the effect then it sounds a fun escapade.

    One point, a possible boring explanation. Look very closely at the plot I supplied

    At dawn there is a period of *total* calm, the dawn effect. This will affect temperature.

    I can decode the mountain of NetCDF to dumped text. Parsing the text is incomplete. A lot of work involved.

    Even if this is merely repeating well known work I think it is valuable since it is independent.

  12. tchannon says: August 30, 2014 at 4:35 pm

    Will, (“How do these folk distinguish between radiative and convective heat transfer?”)

    “Which folks? The article is my own work, independent from anyone else and hence mistakes, finger trouble. If you look at the Chilbolton plot, after dawn it goes into convective water vapour regime, all goes vague and very noisy. There are more instruments than those I show, not decoded, its a lot of work. There is nothing stopping someone else also using the source.
    Most of this is wheel reinvention yet for me that is how to get a true understanding and open up the possibility of a different take or making a new path.”

    OK How do you distinguish between radiative and convective heat transfer?
    I agree that in still air convection will be lowest and the exponential time constant longest.
    At air velocity of even 0.3 m/s convective heat transfer would greatly outweigh any surface radiative effect. You seem to be playing into the “claim” that radiative effect from the surface dominates. You could use the claim of upwelling to estimate the temperature difference of the surface and and recording thermometer to get a guess of surface convection. The thin green line in your “radiative” charts has no meaning, nor will any difference in the so claimed up down “radiance”.

  13. tchannon says:

    There is a lot of background I guess you haven’t seen.

    Radiation does not dominate but does do things, my view on that has appeared, I’m scathing. Clue: why did I recently mention a hard vacuum in relation to lunar dust as critical to heat flow? (a paragraph I added)

    I am accessing the only instruments which seem to be available, making the most of that.

    A while ago I showed conclusively the Kipp & Zonen instruments are used in my view irresponsibly as recommended by the manufacturers. The static radiation field is added on and published as factual.

    The work with the green line involved a great deal of work creating the software. The collects published plots daily, parses the plots, includes dedicated OCR to read scales on autoscaled plots, produces the data and replots thumbnails you see. Originally the underlying data was not accessible, it is now but appears a month or so later. I haven’t written the software to handle this yet. It is over course more accurate.

    These radiation instruments are simple as in 19th century passive technology. The only modern thing are the filtering covers. The instruments turn heat flow into or out of the body via a surface into electricity. The body temperature is also sensed and externally in a computer the static thermal field is computed which is then added to the heat flow data, giving the published result.

    The thumbnail with the green line contains two traces, actual radiative flow is the difference between the two and shown in the bottom right plot. To do this I have reverse computed from the published data using the air temperature as a measure of the instrument body temperature.
    Subsequent access to the actual data confirms the results. I’ll stop there.

    In this instance we are dealing with nocturnal radiation (see the Angstrom articles) for which no filtering cover is needed.

    Under no wind night conditions and with humidity high or rising there is no conditions for convection. These are also conditions where an atmospheric temperature inversion will be forming.

    Of interest here too are the articles mentioning nocturnal jets, winds above ground but close to the surface.

    Clues are also present in night radiosonde data which often shows a close surface inversion although unfortunately the instruments barely have time to record data. The high resolution narrow beam radar instruments show this better. (don’t know if Chilbolton have anything useful, are various instruments I haven’t considered)

    Nighttime inversions are one of the “climatic” studies elephants in the room. Day and night are very different in many ways but the mainstream community ignores this, tries to average and call that reality. Even weather GCM have enormous trouble, papers out there.

    So I can’t *know* convention is absent but the conditions are apt.

  14. tchannon says: August 31, 2014 at 5:59 pm

    “There is a lot of background I guess you haven’t seen.”

    Yes indeed I could not figure what you were getting at with your exponential. Most things relaxing to equilibrium do that asymptotically,exponentially. I didn’t mean to criticize your work! I was confused about what seemed to be the party line “it is all radiation”.

    “Radiation does not dominate but does do things, my view on that has appeared, I’m scathing. Clue: why did I recently mention a hard vacuum in relation to lunar dust as critical to heat flow? (a paragraph I added)”

    Oh yes! Thermal EMR can be difficult to understand when propagating through a dissipative medium, like Earth’s air! First thing to remember, “it is not heat, it does not work like heat”.
    The vacuum plus low gravity makes Lunar regolith have very low thermal conductivity. Higher conductivity and emissivity where the astronauts stepped on/in it. Just like snow.

    “I am accessing the only instruments which seem to be available, making the most of that.
    A while ago I showed conclusively the Kipp & Zonen instruments are used in my view irresponsibly as recommended by the manufacturers. The static radiation field is added on and published as factual.”

    Kipp and Zonen, like AZO, are only resellers with their own marketing, to comply with the WMO nonsense. Try Huskeflux, who actually make the things, I find their engineers knowledgeable, and able to speak the correct language without BS. EKO also manufactures, I have had little contact.

    “The work with the green line involved a great deal of work creating the software. The collects published plots daily, parses the plots, includes dedicated OCR to read scales on autoscaled plots, produces the data and replots thumbnails you see. Originally the underlying data was not accessible, it is now but appears a month or so later. I haven’t written the software to handle this yet. It is over course more accurate.”

    I understand the effort. To help with understanding, Use radiance settle rather than Watts settle.
    For right y-axis instead of Watts/sqm, use apparent PI x (W/(sr x m^2)), or W/(sr x m^2)/PI. This changes it back from a fake scalar flux to a apparant vector radiance. If the WMO choses to ignore steradians and always use PI steradians, that is their geometry problem.

    “These radiation instruments are simple as in 19th century passive technology. The only modern thing are the filtering covers. The instruments turn heat flow into or out of the body via a surface into electricity. The body temperature is also sensed and externally in a computer the static thermal field is computed which is then added to the heat flow data, giving the published result.”

    The thermopile is old but very accurate, if thermal conductivity is constant. It measures sensible heat flux. The real problem is that they attempt to indicate the falsehood of black body radiance while measuring only 3-9 micron flux through a 60% transmissive silicon window. calibrated to a cavity blackbody. If they were to measure air radiance through 100 microns, where 75% of the exitance would be. The down looking radiance would be lower, and the downwelling, (up looking) radiance would be higher. Higher than the down looking in daylight. the difference would be much less or even negative.

    “The thumbnail with the green line contains two traces, actual radiative flow is the difference between the two and shown in the bottom right plot. To do this I have reverse computed from the published data using the air temperature as a measure of the instrument body temperature.
    Subsequent access to the actual data confirms the results. I’ll stop there.”

    What actual data of thermal EMR flux do you have access to?

    “In this instance we are dealing with nocturnal radiation (see the Angstrom articles) for which no filtering cover is needed.”

    Without a window and evacuated thermopile, you are measuring the combination of both radiative and convective flux with the convective always being much greater upward. If you made the surface of that “open” thermopile highly reflective, you would see little change in flux.

    I hope you “get” the contrived gross abandonment of science! ATTACK!!