Airborne thermometer to measure Arctic temperatures


Physicists find a way to measure precise ocean temperature from afar

Russian scientists from the National University of Science and Technology MISiS, MIPT, and Prokhorov General Physics Institute (GPI) of the Russian Academy of Sciences have compared the effectiveness of several techniques of remote water temperature detection based on laser spectroscopy and evaluated various approaches to spectral profile interpretation. The paper detailing the study was published in Optics Letters. The researchers examined four data processing techniques drawing on the relevant analyses in prior publications. The technique which the authors themselves previously proposed, developed and obtained a patent for proved to be precise up to 0.15 degrees Celsius.


Figure 1. Water temperature map of the Arctic region.

The research findings will support further development of sea surface temperature remote sensing solutions, enabling scientists to keep track of thermal energy flows in hard-to-reach areas such as the Arctic region, where average temperatures are rising approximately twice as fast as they are elsewhere on the planet.

In their study, the scientists focused on Raman spectroscopy, which is based on the phenomenon of Raman scattering discovered in the 1920s (Figure 2.). It involves the interaction of a medium with a light wave: The scattered light is modulated by the molecular vibrations of the medium, resulting in the wavelengths of some of the photons being shifted; in other words, some of the scattered light changes its color.


Figure 2. Raman scattering. CREDIT MIPT Press Office

Raman scattering and, by extension, the field of Raman spectroscopy were named after Sir C. V. Raman, an Indian physicist who was awarded a Nobel Prize for the discovery of this effect. Interestingly, Russian scientific literature tends to refer to the same phenomenon as “combination scattering,” a term coined to emphasize its independent discovery by Soviet researchers.

“With the climate changing so rapidly, remote sensing of water temperature is a priority, but the radiometry techniques currently in use are only precise up to about a half degree. Raman spectroscopy enables measurements with a much greater precision,” claims Mikhail Grishin, one of the authors of the study, a Ph.D. student at MIPT, and a researcher at the Laser Spectroscopy Laboratory of the Wave Research Center at GPI.

The experiment carried out by the scientists involved probing water with a pulsed laser and using a spectrometer to analyze the light that was scattered back. Depending on the temperature of the water, its characteristic OH stretching vibrations spectral band was variably transformed. The scientists needed to find out whether it is possible to establish a clear relationship between water temperature and one of the spectral band parameters.

The scientists examined the temperature dependence of several spectral band parameters (aka metrics), viz., certain parts of the area below the graph (see Fig. 3), differential spectra (the result of subtraction of two spectra), and the location of the peak of the curve fitting the band spectrum. Although it proved possible to establish a relationship between water temperature and each of the above mentioned metrics, the estimated temperature measurement accuracy of the respective techniques varied. Statistical analysis of experimental data showed that temperature dependence was most pronounced when the wavelength that corresponds to the peak of the curve fitting the band spectrum was used as a metric. The scientists were granted a patent for the corresponding approach to spectral profile interpretation by the Russian patent office.


Figure 3. Raman scattering spectrum of water OH stretching vibrations at two different temperatures (left); the two-color technique (right), one of the approaches to spectral data analysis in Raman spectroscopy. CREDIT MIPT Press Office

Seawater temperatures in the Arctic are currently monitored using a range of techniques including direct measurements made by weather buoys and merchant or research vessels. However, to track the temperature dynamics of sea surface water in real time and over vast areas, it is necessary to make aerial observations using sensing equipment installed on aircraft or satellites, which irradiates the water with a laser and collects the scattered light. A spatial resolution of less than one kilometer enables researchers to create very detailed temperature maps which can be used to monitor the transfer of heat by ocean currents, predict how fast Arctic ice is going to melt, and make a global climate change forecast. As unmanned aerial vehicles (UAVs) become better, remote sensing equipment should also be improved to be more precise, lightweight, compact, and energy-efficient. The scientists are developing both the software and the laser-detector system.

Vasily Lednev, one of the authors of the study, a leading expert at the Department of Certification and Analytical Control of NUST MISiS, told us how he sees the future of this research: “One of the main hurdles faced in remote sensing of the sea surface is the necessity to calibrate equipment and verify satellite measurement results against contact measurements of seawater parameters (temperature, chlorophyll concentration, etc.). The development and design of compact autonomous lidar (laser radar) systems which can be mounted on UAVs will enable us to obtain detailed sea charts featuring a range of water parameters. These lidar systems are also of immediate interest to the study of hard-to-reach and dangerous objects like icebergs or ice shelves.”

The average annual changes in the temperature of the world’s oceans tend to be very small. It is currently heating up by a mere tenth of a degree every ten years, whereas seasonal temperature variations can amount to several degrees. This means that an error of just half of a degree will cause a significant drop in precision of the overall picture of temperature dynamics that we get. In the case of seasonal measurements, the uncertainty can reach 20 percent or more, while long-term climate trends may remain unidentified due to the measurement error.

The remote-sensing thermometers currently in use operate in the microwave spectral range. Raman scattering spectrometry has a significant advantage over microwave radiometry in that the probing laser radiation falls into the visible (blue-green) part of the spectrum. Unlike microwave radiation, to which water is almost completely opaque, visible light can penetrate a layer of water that is 1-10 meters thick. With microwave sensing, the data is only available for the 30-micron-thick surface layer whose temperature is significantly affected by the cold Arctic winds. This gives rise to an error, which is almost entirely avoided in measurements based on Raman scattering. To correct errors of this kind, satellite-based microwave radiometers need to be calibrated against ground-based measurements. By contrast, Raman spectrometry does not face this obstacle and can produce useful data independently from contact observations.



67 thoughts on “Airborne thermometer to measure Arctic temperatures

  1. 5th paragraph starts –
    ““With the climate changing so rapidly, remote sensing of water temperature is a priority,”

  2. This answers a very basic question I’ve been asking myself for years about the records of past and ongoing global temperatures and their accuracy and reliability for estimating trends and forecast future temperature rises. If there is such an inherent instrument/method inaccuracy/uncertainty of 0.5 degree centigrade with any measurement, with the very many other factors such as differing distribution of measuring stations and their differing and separate heat island inputs affecting accuracy, different methods of measurement and measurement consistency and reliability, and many other varying effect factors etc. etc. etc. then how can anyone within the Green Brigade with any degree of certainty and credibility say that, for instance, average global temperatures are rising at 0.15 degree centigrade per decade leading to a 2 degrees centigrade rise by the year 2100 and as such this requires radical Green policies and projects?

    This ignores the additional considerations of where the heat needed for such temperature rises comes from – whether man-made and by what means as well as naturally and by what means, etc. etc. with or without CO2 inputs?

    From this, what has still amazed me for many years, as a retired professional engineer and project manager involved in many major design and build process and infrastructure projects worldwide from tendering through manufacture, installation, commissioning and operations – including all performance and cost considerations, is a quite baffling question: how can the world have spent, and is still spending, £billions of taxpayers’ and consumers’ funds implementing current Environmental and Infrastructure services and works projects based on man-made CO2 driven CAGW and Climate Change theories? Given proper investigation and application of scientific and technological knowledge of established and scientifically proven theories and normal project investment and risk analysis techniques for any such projects then such massive expenditure involved and uncertainties of effectiveness and outcomes being provided then these green solutions for such projects should never get past even the thinning out stage of system options available to meet said project functional requirements as carried out and leading up to the preliminary short list of options available but needing more investigation.

    • The UN IPCC approach is described in its “Detection and Attribution Analysis”:

      “The approaches used in detection and attribution research described above cannot fully account for all uncertainties, and thus ultimately expert judgement is required to give a calibrated assessment of whether
      a specific cause is responsible for a given climate change.”

      The IPCC answer to your quandary is to depend on climate experts whose brows are anointed by the holy oil of Climate Science to ferret out climate truths from statistical uncertainty.

      My CV would read much like yours, so the \sarc in this response will have to be pried from my keyboard. A colleague reports to me that there is a situation where one whose brow fairly drips with holy climate oil claims that it is possible to detect the fingerprint of man on the recent Sierra snowfall.

    • Sadly climate scientists, acting in a scientific manner only, don’t seem to see the obvious. You can use assumptions as part of the scientific method along with cursory data. You don’t have to have decent measurements all the time. Your paper just becomes one with a more hypothetical bent that may be used if your assumptions are confirmed to be reasonable.

      Trouble is that assumptions aside, measurement method dictates the resolution and accuracy. Or in simpler words you design your test equipment to be able to achieve a certain resolution and accuracy. You back this up with characterisation and regular calibration.

      You do this in a lab and in the outside environment. Tolerances will change.

      So it always makes me laugh when you see temperature siting stations that clearly were meant to give a general estimate to the nearest degree somehow being changed into 0.01 degree variations which are then declared national standards!

      Money is a hell of a drug.

    • Actually, that is a very interesting question. First, the basics. Water is a HUGE absorber in the IR. Operationally, that means that samples have to be rigorously dried, so as to avoid water contamination peaks all over your spectrum. You also like to keep your instruments in climate controlled labs. People go through great effort to protect their expensive instruments from water in all it’s forms. This leads to certain unexpected effects. Such as, if you ask your spectroscopist to put a sample of water into the instrument, your spectroscopist will usually screech, turn purple, and keel over. So you will likely not get your spectrum.
      That said:
      The peak they are looking at is the hydrogen bond peak, which is a great big wide featureless absorption band. When you freeze water, what you are doing is freezing the H-bonding in place. Without taking a series of spectra, I would expect two effects:
      1) The peak will narrow and sharpen to some extent, as the molecules are more constrained. This could cause problems with the peak maximum search algorithm. (probably minor)
      2) The peak maximum will shift upscale, as the H-bonds bonds are “frozen” is place and so appear stronger. (Note: IR peak wavelength (in cm-1) are a direct measure of molecular bond strength)
      BUT: this upscale shift is also interpreted, by their method, of a warmer sample. So without correction, ice might look warmer than water. But: as I mentioned above, ice might have a sharper, narrower peak, so peak width might provide a way to discriminate between phases. If so, then you could select the correct calibration curve.

      These are questions which make me wish I still had an IR spectrometer kicking around.

    • Ice and water spectra are very similar. Steam (vapor) is the outlier.

      Blue and red are water and ice, green is steam. Absorption coefficients.

      Liquid water absorbs poorly in the visible spectrum. This is why there is light to considerable depth when you scuba. Counterintuitively, visible light, the most intense part of the solar spectrum, warms the ocean as a “cavity” precisely because the energy gets in and is dissipated at depth.

      CO2, the erstwhile ocean radiator of ideologues, penetrates only a few microns. Counterintuitively, this skin warming, to whatever extent it takes place, cools the oceans.

      • gymnosperm,

        It should be noted that the absorption spectrum you have displayed is for vertical illumination, where the peak solar radiation has a reflectance of about 2%. As the angle of incidence increases, the reflectance increases, ultimately reaching 100% at 90 degrees. The spectrum of the reflected light begins to approach that of the incident (solar) light at angles over about 60 degrees and ultimately is identical at 90 degrees.

      • Thank you, and noted. The reflected light gets a second shot at atmospheric water (in all three phases) on its way out.

        While doing little to the surface, reflected light is potentially a significant energy source to the atmosphere. Water absorbs strongly in the near IR. Up to half an atmosphere in the mid latitudes there is a lot of water in the air, and according to your formula a lot of reflected light as well.

      • Hydrogen bonding is a key. If I am remembering it correctly, from 40 years ago, dihydrogen monoxide forms the strongest hydrogen bonds. Enough for water vapor to not be monomers, but a mixture of monomers, dimers, trimers, and higher agglomerations, maybe even up to numbers approaching condensed forms.

        Ah Raman IR spectroscopy. That also brings back 40 year old memories.

  3. Curious that it’s always those on the skeptical side of global warming who are more motivated to improve the acquisition of actual climate data. WUWT?

  4. perhaps, if we can see the temp in C or K. Publishing an anomaly chart without baseline attribution is problematic.

  5. Figure 3 shows a shift of 62 cm^-1 in peak maximum per 70 C in temperature change. Assuming the cm^-1/T shift is linear, they must be able to accurately detect a 0.89 cm^-1 Raman shift in order to see a 1 C difference in temperature.

    Look at the noise in the data, in Figure 3. It’s high.

    The article claims a precision of ±0.15 C. We all know that precision is not accuracy. An accuracy of ±0.15 C requires reliably detecting a Raman shift of 0.13 cm^-1.

    A precision of ±0.15 C just means their Gaussian fit to repeated measurements of the same temperature reproduces itself to within ±0.15 C. That says nothing about the accuracy of the measurements.

    Given the measurement noise in Figure 3, there doesn’t seem to be much hope in accurately detecting ±1 C, much less ±0.15 C.

    The method may be better at seeing below the surface, but the calibrated accuracy of the satellite radiometers is about ±0.3 C. It doesn’t look like the Raman method will exceed that accuracy, unless the S/N is radically improved.

    • Also, re: accuracy, the trend of any warming is the key to AGW, so, why does the study make a big deal out of this:

      With microwave sensing, the data is only available for the 30-micron-thick surface layer whose temperature is significantly affected by the cold Arctic winds. …


      And why do they only talk about the Arctic?

      Isn’t the Antarctic of any interest to them?

      • Hi Janice:
        As far as the 30 micron layer goes, I think it is probably better to get a bit farther into the bulk of the material, to avoid the boundary effects. If you are measuring too close to the boundary, you end up measuring evaporative cooling due to wind speed as much as temperature. Then your colleagues laugh at you.
        As far as Russian interest in the arctic goes, remember they have a huge border on the arctic basin, stretching from Finland all the way to the Bering Straight. While the West is fretting about CAGW, the Russian Bear is more concerned with geo-politics and strategic concerns in that corner of the world.

      • Here is a proxy for both Arctic and Antarctic combined, sea ice area, both at record (satellite) lows:

      • Hi, Guys! :)

        Thank you, Ol’ Seadog — yes, indeed….

        Thank you, Professor Tony, for helping me understand. And also, thanks for the intriguing mini-lecture about H bonds up above.

        You are so very welcome, GaryH.

      • Same source data Janice, just showing all the smaller scale oscillations. If mines junk…
        If you don’t like what it reveals when presented like that, instead of a knee-jerk “junk” perhaps ask yourself why.

      • “Isn’t the Antarctic of any interest to them?”

        No, for a variety of reasons.
        (1) It’s a long way away from everywhere.
        (2) Most of it is Australian.
        (3) You can run out of vodka.
        (4) You can run out of banana milkshakes.
        (5) It’s damned cold!

      • @ Roha
        At least your Milkshakes wont melt (though they might solidify making them impossible to enjoy through a straw)
        @Tony Mc
        That chart is looks alot like it is using Bad Data sources

        Similar to this

        or this

        The Drop Off is simply too dramatic to be trusted. Do you know the source of the data that was used to create it?

      • Bryan A
        “That chart is looks a lot like it is using Bad Data sources”

        You’re not the first to make that suggestion Bryan and I agree it looks so anomalous as to be suspect. But it is not, the data is from here:

        The data is “fine” and can be verified here

        The data source is the same as for Janet’s graph above – you know the “junk” one.

        I’ve posted this graph several times. Each time it gets questioned, then, when I point out it is accurate

        A few days ago it was Monkton “oi, that can’t be right”. Then crickets.

        If it accurate Bryan do you, as appears form that plot, think the cryoshere is under-going a significant change?

      • Tony,
        Thanks for the informative links to the Data sources.
        So it appears that the Data is sourced from 5 satellite platforms using a myriad of different sensors.
        It also indicates that, through the intervening years, technological development has increased the available sensors resolution.

        Given this, I would say that the Data does indicate high latitude changes are occurring.

        The 5 platforms
        Nimbus 7 which operated from 1978 to 1994
        DMSP 5D/2 F8 which operated from 1987 – current SSM/I sensor
        DMSP 5D/2 F11 which operated from 1991 until it exploded in orbit in 2004
        DMSP 5D/2 F13 which operated from 1995 until it exploded in orbit in Feb 2015
        DMSP 5D/3 F 17 which was launched in 2006 with the newest sensor SSMI/S

        As to the “Significance” of the changes observed, it certainly is within the spectrum of the current satellite era observations though it still strongly resembles a possible artifact of potential aging sensor error.

    • Actually, only about 42% of Antarctica is Australian, but that is still too much for people of discernment.

    • “though it still strongly resembles a possible artifact of potential aging sensor error.”

      Yes I agree Bryan, it is such an outlier that at first glance “error” is the obvious explanation.

      Except…it isn’t isn’t an error, it can be cross-verified with separate independent data (cited above). There is no error. What there is is a very unusual downward trending outlier. Which for real amateur science buffs should pique their interest. Like wuwt?

      But no, no one here is willing to comment about it, other than to say “That junk graph of T. Mcl. CAN’T be right.”, which is disappointing given that so many armchair ‘scientists’ commenting here are so critical of what they perceive as “junk” science.

      “Junk” is, apparently, anything that fails to accord with their internal biases. Ironically demonstrated by Janice’s failure to notice that ‘her’ graph is based on the same data as ‘my’ “junk” one.

      So I’ll keep posting it. And this one:

      Which, given the current poor state of the sea ice, makes it pretty clear that without a massively cold outlier of a spring/summer weather the Arctic is going to be transformed into something quite different this summer of next.

  6. Pat Frank,
    I have observed that Raman scattering spectra from opaque minerals is much noisier than from transparent minerals. The reason apparently being that one only gets surface scattering from opaque minerals, but obtains bulk volume scattering from transparent minerals. I’m going to make an extrapolation here and suggest that the SNR is related to the extinction coefficient of the complex index of refraction of water, which varies with wavelength, temperature, salinity, and suspended particulates. So, one thing they could do is use a blue laser instead of a blue-green, which might be affected green reflections from plankton, and would have better penetration (lower extinction coefficient) in clear water. I think that the technique has some promise, but it probably isn’t just an issue of instrumentation. They will have to take into consideration the physical environment.

  7. How much warming will these measurements cause?

    We will focus and direct microwave radiation and lasers at the Earth’s surface in an effort to determine if incidental reflected radiation from itself is causing excessive heating.

    It’s akin to drilling tens of thousands of ice cores from the ice that we so desperately want to preserve.

  8. why the need for this when we know ocean heat content to the zettajoule ? (i hope the sarc tag is not required).
    in all seriousness significant improvement is needed in this field , from personal experience i have a hard time agreeing with stated sea surface temps in areas of the north east atlantic/north sea .

  9. “…areas such as the Arctic region, where average temperatures are rising approximately twice as fast as they are elsewhere on the planet.”

    Do we have enough Arctic temperature readings over a long enough time to make such a claim? Or are they estimating the temperature or basing it on climate models? If global warming is global, why aren’t Antarctic temperatures also rising twice as fast as elsewhere? Something doesn’t add up.

    • There are perfectly logical answers to all of your legitimate questions Louis. Jump on google and find them out. Start with: why is Antarctica colder than the Arctic?

    • UAH NoPol has good reading this century before the 2015/16 El Nino spike.

      And the end of last century before the 1998 El Nino

      • Try plotting the UAH v5.6 version.
        Ie, before it was “adjusted to this v6.0.

        Here are the regional differences between the two version.

        The Arctic trend has been halved.
        Now wot would you say if it had been doubled?
        Or a surface series trend had been warmed that much post “adjustments”?

      • Gee Toneb that’s an easy one, If the adjustments had doubled the warming trend, it would have been referred to as being Karlized (ala Thomas Karl) but since the adjustments seem to remove a warming trend, Warmers, like yourself, seem to think the adjustments are improper.
        Ajdustments = Warmer = good
        Adjustments = Cooler = Bad???

        That thuroughly explains the AGW mindset

    • … the Arctic region, where average temperatures are rising approximately twice as fast…


      Bristol-Myers Company (Bristol) petitions for review of an order of the Federal Trade Commission (the Commission or FTC) made in respect to the advertising by Bristol of its well-known analgesics, Bufferin and Excedrin. … We have considered each of Bristol’s claims as to the remedial order and deny the petition for review and grant enforcement. …

      In concluding that Bristol and its advertising agencies had deceptively advertised Excedrin and Bufferin, the Commission found that Bristol had misrepresented that the analgesic superiority of Excedrin and Bufferin over competing products was scientifically proved, or “established,” by the artful use of certain phrases such as “scientific tests” and “medically endorsed,” as well as by the use of visual images. ***

      Bristol-myers Company, Petitioner, v. Federal Trade Commission, Respondent, 738 F.2d 554 (2d Cir. 1984)

      (Source: )

  10. “The technique which the authors themselves previously proposed, developed and obtained a patent for proved to be precise up to 0.15 degrees Celsius.”

    In a laboratory setting perhaps. Heisenberg’s uncertainty principle hampers measurement precision of any wave-like system:

    A precise location must be defined before a precise value can be measured. In the absence of precise location the value precision is lost. This is due to the inherent inhomogeneity of the measured system, irrespective of the instrumentation used.

    Now all ‘man-made this and that’ debates share one thing in common – the measured location something like ‘atmosphere’, ‘ice in the Antarctica’, ‘water in the Arctic’ etc. That’s when the debate moves from science to politics. There is no point debating the precision of politics. So why accept an exception to ‘man-made this and that’?

  11. Unlike microwave radiation, to which water is almost completely opaque, visible light can penetrate a layer of water that is 1-10 meters thick. *
    With microwave sensing, the data is only available for the 30-micron-thick surface layer whose temperature is significantly affected by the cold Arctic winds.

    * Good News!

Comments are closed.