How not to measure temperature, part 4, at the Royal Observatory


This is a Google Earth satellite photo of the Royal Observatory in Edinburgh. The weather station, at 55°55′22.71″N 3°11′17.69″W is the white box in the middle of the grass circle. Of concern is not just the nearby roads, but also the buildings. It appears the station is almost completely encircled by tall buildings.

This means that the heat from the buildings will significantly bias the temperature, and reduce wind which adds further bias.

You’d think top scientists would know better?


8 thoughts on “How not to measure temperature, part 4, at the Royal Observatory

  1. Your expose of inherent measurement biases in long-term earth temperature is great. I do have one niggling questions, though:

    Isn’t the fact that some measurement stations are surrounded by asphalt, buildings, etc, one of the points of global warming: all of the man-made structures are helping cause global-warming. By measuring in these places we are actually measuring the real phenomenon and not an error. The concrete, asphalt, buildings, air conditioners, etc. are systemic to the fact that the population is growing and the effects of this growth (more buildings, roads, etc.) are felt in the temperature. The main point against that line of reasoning, that I can see, is that the volume of atmosphere is many orders of magnitude larger than the localized effects of the infrastructure creep: pouring a 1000 acre concrete parking lot won’t affect the temperature 1000 feet up and 50 miles away. Therefore, the concrete parking lot won’t affect the weather-patterns , etc. Or put another way, an actual localized temperature increase of 1 or 2 degrees, while real at the location, cannot change the temperature of a location even 500 feet away because the volume of air that must change is too large. Therefore to use these “bad” stations as an indicator of continuous temperature change across the large area between stations is wrong.

    Anyway, just a thought, that the placement of some of the measurement boxes might not be as egregious as thought. This could be totally off base for the reason given above, among others. In addition, it does not let the measurement folks off the hook for such violations as: different box paint, lights/heaters in the boxes themselves, and differing instruments across time, among others.

  2. Ever notice how anthropogenic global warming advocates always quote so called “top scientists”?

    I guess any scientist who questions the methods and data or has a heterodox point of view is, necessarily, a “bottom feeder”.

    I would expect better from “top scientists”. This is absolutely ridiculous.

  3. Dave, Interesting comments.

    There are the 4000 stations that are currently used in the GISTEMP data set used for climate modeling. 1221 of them are in the USA.

    These weather stations only sample the micro-climate surrounding the sensor. Assume an area around the sensor of, on average about 25 square meters…5m X 5m….with wind effects, maybe that is too small? OK how about 100 square meters but that is broad since the temperature sensor bulb really only measures the volume of air immediately around the sensor, which is quite small, but wind does carry and mix heat from nearby sources, so eventually it evens out over an area.

    So then, with 100 square meters of area around a sensor being the representative area sample in the micro climate, multiplied by 4000 sensors worldwide, we can agree that roughly 400,000 meters squared or 400 square kilometers squared of earths surface is actually represented in the meteorological station data used in computer modeling of temperature; compare this to the actual area of earth’s surface 510,065,600 square kilometers.

    This demonstrates the weakness of using this data set to measure or predict a global mean temperature. The area being covered by weather station data is only .00008% of the earths total surface.

    Global decisions are being rendered on that .00008%

  4. Dave, you said some things that make sense but need a further analysis that will make your point stronger: man-made structures (cities, etc) are helping to store heat not helping “cause” global warming. Global warming means the whole world –and cities, paved roads, etc, are not more than 2% of the Earth’s surface. Think: about 75% is ocean, the remaining 25% are mountains, jungles, forests, pampas, deserts, and open country.

    You are forgot to explain a little better one fundamental factor (that is advanced by a later post by Anthony): weight of the station in the overall record. One example: a weather station like the one in the picture really records temperature in an area of say, 1 hectare, or 0.01 km2. A rural station surrounded by 1000 km2 shows the temperature of a region 100.000 times bigger. But both stations are given the same importance or weight in the raw data, and the small paved area has the same meaning in the “global warming scenario” as one lonely station in Siberia that represents a region of perhaps a million square kilometers.

    Presumably, GISS or Hadley make corrections for giving each station its proportional importance in the overall record, but I have serious doubts about the way records are kept and managed. The record from a station close from where I live, (a rural one in Argentina central area) was represented by GISS in the graph appearing in GISS Station Data showing a warming for 2006 that was spurious. The record lacked the two last months in the year, but GISS made up the yearly average just averaging prior years (as they told me). That was happening in February. Apparently the Argentinea weather service didn’t prove the records on time.

    Actually, both November and December were much cooler than previous years, as have been January, Februray, March, April and May 2007 –you surely have seen the news in TV about our recent cold spell that set the country in a energetic crisis and 40+ dead.

    GISS graph showed 2006 as a warming year, when actually it was a cooling one. The period from January 1st to May 31st has been the coldest in 36 years, and broke many records: snow in mid summer on Bariloche (!) so skiers could practice their sport for two days; the earliest snow in Cordoba for April –since 1880; and the average temperature for the summer and autumn has been consistently 5º C below normal. A strange event, for sure, that shows that the climate is changing in the Southern Hemisphere, but towards cooling, a trend in our huge region of central Argentina that started back in 1987.

    Just food for thought. And yes, my job is in the meteorology and climate activity.

  5. Thank you so much for this research, I have thought the whole magilla was a crock of crap from the start. It has to do with money flowing into the pockets of the “scientific” community and the likes of algore, the idiot child of another corrupt politician.

    The Watchman, AKA Harley Hassler.

  6. This “research” is a giant red herring being used by Global Warming deniers.

    1) the CWOP (Citizens Weather Observing Program) has many tools already embedded in it, to warn users and account for, poorly sited or otherwise skewed data.

    Here is my site:

    Scroll towards the bottom and you will see the MADIS ratings.. currently, because it has been cloudy lately, my rating has been good. When there is a stretch of sunny weather, though, because the sensor is on my roof it gets heated by the roof… however…

    if you click on the “32 Weeks” link to get the long analysis, then MADIS reports this:

    “It appears that your station is sited in a location where cool air collects at night. This is often near the bottom of a valley. The cool nighttime readings are probably correct, and the analysis values are not taking into account the local topography.”

    Lo and Behold, I am indeed at the bottom of the Alberni Valley, so while, instantaneously, and on sunny days only, my readings may be 1-2 degrees higher than they should be… over the year, and taken on a wider geographical scale, they are actually lower, because of the much larger effect that Valleys have on local temperatures.

    (my sensor is also right beside my bathroom vent, which obviously skewes both the humidity and temperature values), however, because the fan is only on for less than 20 minutes… and not necessarily after having a shower, the effect on the averages over even the day is negligeable… and over the year, not at all, again as you can see from MADIS.)

    Point being… you’re “analysis” is overblown. These “problems” have been dealt with and accounted for for many years using the most sophisticated techniques available.

Comments are closed.