Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

What is meant by 'Earth's surface temperature'?

  1. Mar 18, 2015 #1
    Seeking clarification here. As a climate change sceptic, I loosely follow several blogs that offer critical views of the science behind the idea of global warming, but I am not especially across the breadth of the science. One of the things I struggle to understand is the idea of 'surface temperature'. Or more exactly, what is meant by this.

    When we talk of the surface temperature of the moon, we are talking about the temperature of the moon's actual surface (or at least, I presume we are). I assume this is inferred from measuring the moon's spectrum of radiation? Without an atmosphere, it is easy enough to establish an actual, and an average, temperature.

    The earth's temperature however, as considered by the various temperature indices, is measured by thermometers. These are generally measuring the air temperature several feet above the surface. They are not measuring any surface at all. This is easy to see by noting the differences in temperature depending on whether the thermometer device is in shade, in direct sun, near a heat source, or if a hot or cold wind is blowing.

    I *think* that the sea temps, as measured by water sampling in times past, or Argo floats etc now, are measuring the water temperature itself, hence it is measuring a surface. But surely ocean temps so derived are not directly comparable with air temps over land?

    My questions then, are:

    1. Is the use of the term 'surface temperature' for the various indices sloppy nomenclature?
    2. Should ocean and air temps be combined (if in fact they are)?
    3. Is the earth's average temperature as expressed in terms of the greenhouse effect (ie 14C rather than -18C) derived from these indices, or is it derived from measurements from space of earth's radiating temperature (as is done for the moon)?
     
  2. jcsd
  3. Mar 18, 2015 #2

    DrClaude

    User Avatar

    Staff: Mentor

    You are confusing different measurements. Land surface temperature is measured by satellites. Air temperature is what is measured by thermometers.
    Source: http://earthobservatory.nasa.gov/GlobalMaps/view.php?d1=MOD11C1_M_LSTDA

    The figure of -18 °C is what you get when you consider the amount of solar radiation hitting the Earth, and consider the Earth to be an imperfect black body (reflection of 30%). At equilibrium, you then get a temperature of 255 K. The fact that this is obviously not what is observed means that there is an additional source of heat, which happens to be the atmosphere trapping radiation that is sent back to the ground, making the temperature higher.
     
  4. Mar 18, 2015 #3
    Thanks for that. I understood most of the major indices to be constructed from thermometers. Satellite measurements have only been in place since the 70s I think? So when we do backwards comparisons, are we not using thermometer derived temperatures? How can these be compared to satellite temps? As I asked above, that would be an apples and oranges comparison wouldn't it? When we look at all the anomaly graphs, we are not using the land surface temperatures described at your link, surely?

    In terms of the greenhouse effect, the 255K is the theoretical temperature of an 'imperfect blackbody' at equilibrium with 30% incoming radiation reflected. The average temperature of ~ 288K measured, is what is measured by satellites? Can you show me a source for that? It seems unlikely to me that the global average thermometer measurement would be the same as the satellite measured temperature.

    I'm not trying to find fault with anything, just trying to get my head around the meaning of all the temperature indices.
     
  5. Mar 18, 2015 #4

    DrClaude

    User Avatar

    Staff: Mentor

    I'm not a climate scientist, so I can't give any detailed answer. My understanding is that a lot of work goes into trying to harmonize the data, coming from such diverse sources, from satellites to tree ring data.

    For the point of the exercise, it doesn't matter which temperature you use as a comparison. For instance, both land surface temperature and air temperature are much higher than the black body back-of-the-envelope calculation.
     
  6. Mar 18, 2015 #5
    I think the Nasa GISS Says quite a bit about it, but the answers do not inspire confidence.
    http://data.giss.nasa.gov/gistemp/abs_temp.html
     
  7. Mar 18, 2015 #6

    Svein

    User Avatar
    Science Advisor

    There we go, doing "average" again. Way back, when developing a program to calculate the need for heating/cooling of an office building, I got hold of one year of official meteorological temperature data for one full year in Oslo Norway. Not the "average" temperature during a day, but actual thermometer readings taken four times a day at fixed hours. Running standard statistics on those gave slightly different "averages" over a year at each time-of-day, but in every case the standard deviation was slightly more than 10°C.

    With that background, what do you think about the statement "the average temperature of the earth has increased by 0.1°C in the last 100 years".
     
  8. Mar 18, 2015 #7
    My world is mostly digital now, but used to be very analog. (I know digital is a special case of analog.)
    There was something called Signal to Noise ratio. I suspect those same formulas would apply.
     
  9. Mar 18, 2015 #8
    Personally I am somewhat skeptical that we can say much about historical temperature trends considering both the likely daily variation in temperature at any single location and the extent of changes to the record via homogenisation etc.

    However my question was more about what 'surface temperature' means. I think DrClaude above is right in claiming that the average temperature of 14C that is usually offered as evidence of the greenhouse warming effect is a satellite measured land surface temperature, but I have never seen a source that indicates that. I'd be interested to see someone cite a legitimate source for that.

    Most discussion on blogs talk about surface temperature, but then refer to things like the GHCN or GISS or HADCRUT indices which as far as I know are derived from thermometers. While it is reasonable to measure the atmospheric temps and then offer some idea of global trends, it seems to me that we shouldn't claim that to be the 'surface temperature'. But that's just my thinking on it and as I said I only have a pretty basic grasp of the science.

    So refining my questions earlier:

    1. Is the 14C average earth surface temperature that illustrates the greenhouse effect derived from satellite measured SURFACE temperature, or is it from thermometer records?
    2. Do we have an historical record of global temperature anomalies derived from the actual land surface temperature measurements which illustrate a trend in same? And how does that compare to the trends shown by the thermometer records? I realise we can only have satellite records from some 30 or 40 years, but that would be a useful comparison just the same.
     
  10. Mar 18, 2015 #9
    The GISS has their averages for the last 134 years,
    http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
    But as the included note states,
    "sources: GHCN-v3 1880-02/2015 + SST: ERSST 1880-02/2015
    using elimination of outliers and homogeneity adjustment
    Notes: 1950 DJF = Dec 1949 - Feb 1950 ;"
    and
    "Best estimate for absolute global mean for 1951-1980 is 14.0 deg-C or 57.2 deg-F,
    so add that to the temperature change if you want to use an absolute scale
    (this note applies to global annual means only, J-D and D-N !)"
     
  11. Mar 18, 2015 #10

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    1400 uncorrelated measurements with a standard deviation of 10°C allow to calculate the average with an uncertainty of about 0.3°C. But with your fixed measurements, it is even better - you can define this measurement procedure to be standard, and repeat it every year, completely removing the effects of daily and yearly temperature variations if you compare two different values.
     
  12. Mar 18, 2015 #11
    1. Question 1

    The average surface temperate of the earth (globally) for a year, is a calculated statistic, that relies upon these measurements and calculations:
    1. An average of daily local temperate at a given place (e.g. a minimum and a maximum temperature in 24 hours).
    2. An average of those daily temperature averages at that place for a calendar year (say 365 days), for that given place.
    3. An average of these calendary year average temperatures for locations distributed over many places on the surface of the earth, at different longitude and latitude.
    In addition, there will be some local adjustment of temperature data if it is considered to be affected by local factors that are not of interest to the global statistic being measured.

    The above approach is the one commonly adopted by meteorological stations, which are in the business of collecting this sort of data and so that tends to get used. The daily average temperature is based on the minimum and maximum temperatures that they routinely collect. The various research institutes will have an affiliation with these sort of data collectors, then the information is collated into databases for further analysis and use.

    Temperature anywhere on the earth varies with altitude, place and time of day. So on a city street, in summer, during the day, the ground itself might be say 70C, and the air above it 35C. At the same time, the ground temperature in a nearby grass park might be only say 30-35C, or 25C in the shade, and the interior of a car might be closer to 60-70C. As you go higher in the atmosphere, the air temperature will drop until it is below freezing temperature for water, which is where we will find clouds and ice if there is moisture in the air.

    The convention is that surface air temperature (measured by thermometers inside Stevenson screens at places like met stations and airports) is measured about 1.5 to 2m above the ground surface. The temperature statistics and the averages that are gathered only make sense based on that assumption. You do not have daily data collection saying that the maximum temperature today was 70C, even though the ground actually did reach that temperature. The daily statistics, used for daily weather reports and predictions are just the air temperature, which is what most people are concerned with for comfort and daily activities.

    Due to the averaging process, and the fact that temperatures are often latitudinally distributed due to the rotation of the earth and the angle of incidence of the sun (i.e. warmer near equator, cooler toward the poles, and reasonably consistent across latitudes), the global average temperature statistic that is produced can also be viewed as a kind of representative number for the average daily (air) temperature that might be expected in the mid-latitudes, averaged over the course of a year. A number of about 15C (288K) suggests that the air temperature is quite pleasant during the day in mid-latitudes, assuming that the daily range of temperatures is not too great (e.g. if it is 7 to 10C variation from that average either way).

    2. Question 2

    It depends. In practice, they seem to be by the research organisations (e.g. the GISS global averages). See this URL : http://data.giss.nasa.gov/gistemp/ and here : http://data.giss.nasa.gov/gistemp/graphs_v3/

    Should they? If you are measuring air temperatures, probably not a good idea to combine them. If you are measuring temperature at a particular altitude (e.g. sea level), you could if you wanted to be consistent.

    3. Question 3.

    The earth's global average temperature in the vicinity of 14C or 15C is derived from observations near the surface of the earth. These could be from land based or sea-based observations. For example, in Svente Arrhenius' 1896 paper : "On the influence of carbonic acid in the air upon the temperature of the ground.", he used temperature observations of others to calculate the average global temperature statistic for himself, based on specific latitudes and seasons. This is an example of Steps 2 and 3 I have referred to in answer to question 1.

    The "greenhouse effect", as it is known, is a statement of the issue that presents itself when physical scientists try and predict what constant temperature the earth should be from a general theory about the incoming solar radiation. So the -18C (or 255K) that is referred to, is the temperature that scientists have predicted the earth should obtain, independently of observations.

    The 255K temperature is based on the assumption that the earth receives all of the sun's directional energy (e.g. 1367 Watts per metre squared on average, for the cross-sectional shape of the earth i.e. a circular disk), and then warms uniformally to a temperature that allows the earth to emit all of this energy outward from four times the surface area. Due the physical uncertainty of how this would happen, sometimes it is said that the earth simply redistributes this energy instantaneously, or is a "perfect thermal conductor". The theory assumes there is some solid surface to work with. It doesn't distinguish between an atmosphere, ocean or solid ground.

    Scientists have tried to explain the difference between 255K (predicted) and 287/288K (calculated, based on observations) not so much by examining the initial assumptions again, but by suggesting that there is a period where energy builds up and remains in the system long enough to warm the lower levels of the atmosphere, and then later settles down into a cycle where energy is released to space from the higher altitudes in the atmosphere at a lower rate due to the lower temperature.

    So to summarise:
    (a) in making the initial predictions that there was some temperature difference between theory and observation to explain, the 255K was a prediction about the surface temperature at the ground surface (in theory, but not observed)
    (b) in explaining the current situation, the 255K is used as a theoretical emission temperature, but using a "surface" that is higher in the atmosphere.
    (c) observations will not actually find a uniform 255K temperature in the atmosphere above the earth since temperature is not uniform at a given altitude; rather it will again need to be calculated as an average from different places and times.
     
  13. Mar 19, 2015 #12

    Svein

    User Avatar
    Science Advisor

    upload_2015-3-19_7-55-37.png Here is the daily noon temperature over a year (Oslo, Norway, 1969)
     
  14. Mar 19, 2015 #13

    Svein

    User Avatar
    Science Advisor

    I did not say or imply "1400 uncorrelated measurements with a standard deviation of 10°C". What I have is, for each day of the year, the temperature at several fixed time points. Due to the format of the data (think a massive FORTRAN FORMAT statement). it is easy to calculate the average and the standard deviation for each time point, but really hard to calculate the total average and standard deviation. Thus, what I have calculated is the average and the standard deviation for each time point during a year. Most of these standard deviations (of 365 temperatures) are between 10 and 11 degrees*. Using the rule for compound standard deviation, the standard deviation of the aggregate is also between 10 and 11 degrees.

    *The sole exception is the measurements taken at 7 o'clock in the morning, where the standard deviation was between 8 and 9 degrees.
     
  15. Mar 19, 2015 #14

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    It does not make sense to talk about the standard deviation of a single measured temperature value (unless your temperature sensor is really bad). Do you mean the standard deviation in the set of measurements for a specific time of the day?
    The number of 10 degrees is the width of your distribution - you can determine the mean much better than this width. And if you pool the different times of the day for a combined average, you get an even better estimate for the mean.
     
  16. Mar 19, 2015 #15

    Bystander

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2016 Award

    See Berry, Bollay, and Beers, Handbook of Meteorology, pp. 540-2, for state of the art circa 1945.
     
  17. Mar 19, 2015 #16
    Thanks CraigAU for your explanations, however I am not certain you quite tackled where I am coming from. I freely admit I might misunderstand the matter so much that my question is just silly. Perhaps it is better if I posed this in the Physics forum as I can see we are getting more into that field.

    I will give that a go.
     
  18. Mar 19, 2015 #17

    Bystander

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2016 Award

    Not really --- this is more a matter of there being no operational definition consistent with the available meteorological methods and records.
     
  19. Mar 20, 2015 #18

    Svein

    User Avatar
    Science Advisor

    Yes. See post #13.
    The mean of what?
     
  20. Mar 20, 2015 #19

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    The mean of all temperature values.
     
  21. Mar 20, 2015 #20

    Svein

    User Avatar
    Science Advisor

    The mean of all temperatures is about 4°C with a standard deviation of about 10°C.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: What is meant by 'Earth's surface temperature'?
Loading...