I believe this is a truly important piece of work. I hope Dr. Spencer will submit it to a journal. I’m grateful to Dr. Spencer for his email suggesting I post it here. Consider this early peer review. Beat it up, find any errors, and point out flaws, so that he can make it better. – Anthony
by Roy W. Spencer, Ph. D.
UPDATED (12:30 p.m. CST, March 3): Appended new discussion & plots showing importance of how low-population density stations are handled.
ABSTRACT
Global hourly surface temperature observations and 1 km resolution population density data for the year 2000 are used together to quantify the average urban heat island (UHI) effect. While the rate of warming with population increase is the greatest at the lowest population densities, some warming continues with population increases even for densely populated cities. Statistics like those presented here could be used to correct the surface temperature record for spurious warming caused by the UHI effect, providing better estimates of temperature trends.
METHOD
Using NOAA’s International Surface Hourly (ISH) weather data from around the world during 2000, I computed daily, monthly, and then 1-year average temperatures for each weather station. For a station to be used, a daily average temperature computation required the 4 synoptic temperature observations at 00, 06, 12, and 18 UTC; a monthly average required at least 20 good days per month; and a yearly average required all 12 months.
For each of those weather station locations I also stored the average population density from the 1 km gridded global population density data archived at the Socioeconomic Data and Applications Center (SEDAC).
All station pairs within 150 km of each other had their 1-year average difference in temperature related to their difference in population. Averaging of these station pairs’ results was done in 10 population bins each for Station1 and Station2, with bin boundaries at 0, 20, 50, 100, 200, 400, 800, 1600, 3200, 6400, and 50000 persons per sq. km.
Because some stations are located next to large water bodies, I used an old USAF 1/6 deg lat/lon percent water coverage dataset to ensure that there was no more than a 20% difference in the percent water coverage between the two stations in each match-up. (I believe this water coverage dataset is no longer publicly available).
Elevation effects were estimated by regressing station pair temperature differences against station elevation differences, which yielded a cooling rate of 5.4 deg. C per km increase in station elevation. Then, all station temperatures were adjusted to sea level (0 km elevation) with this relationship.
After all screening, a total of 10,307 unique station pairs were accepted for analysis from 2000.
RESULTS & DISCUSSION
The following graph shows the average rate of warming with population density increase (vertical axis), as a function of the average populations of the station pairs. Each data point represents a population bin average for the intersection of a higher population station with its lower-population station mate.
Using the data in the above graph, we can now compute average cumulative warming from a population density of zero, the results of which are shown in the next graph. [Note that this step would be unnecessary if every populated station location had a zero-population station nearby. In that case, it would be much easier to compute the average warming associated with a population density increase.]
This graph shows that the most rapid rate of warming with population increase is at the lowest population densities. The non-linear relationship is not a new discovery, as it has been noted by previous researchers who found an approximate logarithmic dependence of warming on population.
Significantly, this means that monitoring long-term warming at more rural stations could have greater spurious warming than monitoring in the cities. For instance, a population increase from 0 to 20 people per sq. km gives a warming of +0.22 deg C, but for a densely populated location having 1,000 people per sq. km, it takes an additional 1,500 people (to 2,500 people per sq. km) to get the same 0.22 deg. C warming. (Of course, if one can find stations whose environment has not changed at all, that would be the preferred situation.)
Since this analysis used only 1 year of data, other years could be examined to see how robust the above relationship is. Also, since there are gridded population data for 1990, 2000, and 2010 (estimated), one could examine whether there is any indication of the temperature-population relationship changing over time.
This is the type of information which I can envision being used to adjust station temperatures throughout the historical record, even as stations come, go, and move. As mentioned above, the elevation adjustment for individual stations can be done fairly easily, and the population adjustments could then be done without having to inter-calibrate stations.
Such adjustments help to maximize the number of stations used in temperature trend analysis, rather than simply throwing the data out. Note that the philosophy here is not to provide the best adjustments for each station individually, but to do adjustments for spurious effects which, when averaged over all stations, will remove the effect when averaged over all stations. This ensures simplicity and reproducibility of the analysis.
UPDATE:
The above results are quite sensitive to how the stations with very low population densities are handled. I’ve recomputed the above results by adding a single data point representing 724 more station pairs where BOTH stations are within the lowest population density category: 0 to 20 people per sq. km. This increases the signal of warming at low population densities, from the previously mentioned +0.22 deg C warming from zero to 20 people per sq. km, to +0.77 deg. C of warming.
This is over a factor of 3 more warming from 0 to 20 persons per sq. km with the additional data. This is important because most weather observation sites have relatively low population densities: in my dataset, I find that one-half of all stations have population densities below 100 persons per sq. km. The following plot zooms in on the lower left corner of the previous plot so you can better see the warming at the lowest population densities.
Clearly, any UHI adjustments to past thermometer data will depend upon how the UHI effect is quantified at these very low population densities.
Also, since I didn’t mention it earlier, I should clarify that population density is just an accessible index that is presumed to be related to how much the environment around the thermometer site has been modified over time, by replacing vegetation with manmade structures. Population density is not expected to always be a good index of this modification — for instance, population densities at large airports can be expected to be low, but the surrounding runway surfaces and airplane traffic can be expected to cause considerable spurious warming, much more than would be expected for their population density.
Nice. A couple of comments, FWIW:
1. You don’t define “station pairs.” I could guess what they are, but explicitly defining them and how they are selected seems important to your paper.
2. Is it possible to extend this work to “energy density?” Or does population density pick that up?
Thanks Dr Spencer, a very useful piece of work that convincingly confirms the UHI effect and documents its magnitude.
A couple of comments:
Can you make the USAF water coverage data public, bearing in mind all the calls for transparency and data availability?
I guess that population density is shorthand for heat energy release per unit area. I wonder if some third world areas of high density population might have a lower heat production per unit area compared with first world. Would adjusting for per capita income help the analysis?
1. More information must be provided on how the station pairing was accomplished. What, in addition to the 150 Km proximity, was considered in the pairing?
2. The two-station pair average graph would be more informative if the horizontal axis were logarithmic. Or perhaps simply non-linear, with the bin values equally distributed on the axis.
3. Is Dr. Spencer certain that he obtained raw data from the ISH? I am quite tired of learning, after the fact, that alleged raw data has been manipulated in some way. See Darwin Zero for an example.
4. Somehow 1.6 deg. C does not correspond, in my memory, to the UHI corrections in the IPC report. The chronicled Alaska report, for example, showed a much greater UHI effect.
5. The elevation corrections appear quite reasonable.
A marvelous piece of work!
UHI is very variable depending on time of day, snow cover, wind speed, cloudiness, elevation (cold air sinks) etc. On a windy day, UHI may be undetectable. Along the Front Range, locations closer to the mountains (i.e. higher elevation) are usually warmer than lower elevations further from the mountains.
Perhaps a blind comparison of a large number of stations (like Dr. Spencer is doing) averages out all of micro-scale effects?
I would venture that the first 1 degree on the curve (up to population density of 100/km^2) is not sufficiently accurate to be useful. The heat sources are too thinly spread for their impact to be related to the weather station which might be in their back garden, or 2 fields away. Focus instead on how much can be deduced about the 1-2 degree range. Rural is a very difficult judgement to make, but how valuable is the ability to correct for a change in population density from 1000 to 7000?
I’m not suggesting that the 1st degree of population impact is over-estimated or not important – just that it’s use is much harder to justify as ‘not exagerated’.
Spencer: “this means that monitoring long-term warming at more rural stations could have greater spurious warming than monitoring in the cities.”
One would then reasonably expect to see much higher trend at rural stations than at urban ones, ok? However, the USA data show exactly the opposite – just 0.1 deg C of warming in 20th century, but 0.6-0.7 degrees C in the cities. How that can be, if the rural stations should have a larger warming bias?
Another interesting analysis on the “non-existent” UHI problem.
It would be interesting to see how the analysis changed with all the stations pairs at different altitudes removed.
The correction factor is the “average” but in practice – depending on the terrain – air can be easily pushed up or down changing it from its “potential temperature”. Perhaps this all averages out or perhaps not.
The Japan UHI paper showed the huge variation in UHI vs population density from 561 stations, with a strong time of day bias (night time in cities is where the big effect takes place). UHI is strongest at night, almost non-existent during the midday to mid-afternoon time frame.
Although UHI studies have taken a life of its own, maybe we should just stop thinking that precise measurements of a light ephemeral substance 6ft off the ground in a few thousand random locations around the world have much chance of being an accurate indicator of what’s happening to climate.
Even if we adjust for population maybe someone then does another study on rural temperature stations and finds that there is a vegetation growth effect due to temperature so we have to then correct the temperature in stations by a factor relating to temperature! and proximity of nearby vegetation..
Which is Why Global Mean Surface Temperature Should be Relegated
Chris H (11:27:22) :
I wonder if some third world areas of high density population might have a lower heat production per unit area compared with first world. Would adjusting for per capita income help the analysis?
Hard working third world people irradiate more IR, while for sure first world climate modellers irradiate almost nothing, and their per capita income is many thousand times than real workers, though becoming each passing day of a more imaginary value.
Do you mean “hourly surface DATA”?
REPLY: yes, that was an editing error. My original title had both “data” and “area” in it but was too wordy, so I scaled it back. Somehow I deleted data and left area, fixed now thanks. – Anthony
“0.1 degrees C” was for RURAL stations in my previous comment….
I always find Dr Spencer’s work to be imaginative, instructive and valuable. However, I can’t help but think that the time for using global temperatures to calculate changes in global heat is past. Temperature is a proxy measure for heat energy, so why use a proxy when we can measure heat energy directly using satellites? Climate science has become bogged-down with theoretical global temperatures and anomalies, using temperature data from weather stations that were intended to monitor local weather conditions, not global climate, which the Earth does not have!
So, does this involve making a large artificial correction to the data?
A most interesting piece of work! I am no expert here, but would it be possible to argue that the UHI effect should be larger in colder places like Alaska!?
Recommend using logarithmic or geometric horizontal axis –
reiterating mathman’s recommendation above:
“This graph shows that the most rapid rate of warming with population increase is at the lowest population densities.”
But if this is a discussion of UHI effects, I find it hard to categorize those low densities as “urban”.
“Significantly, this means that monitoring long-term warming at more rural stations could have greater spurious warming than monitoring in the cities.”
Why is this warming spurious? Again, these rural stations should not be considered urban? Is it really logical to accept that increasing density from zero to 20 people per sq. km will cause +0.77 deg. C of warming locally, without ths sq. km being inside a bubble?
Last but not least…though really last: When we die we suddenly lose heat.
After all screening, a total of 10,307 unique station pairs were accepted for analysis from 2000.
20,000?
Regards Michael Oakes
The layman reader might be interested (I am) in examples of various weather station locations with their population densities. Further, where does the typical (average?) station fit on the UHI graph? Has Dr. Spencer done a back of the envelope calculation to see what the warming bias has been in the official temperatures data sets, as suggested by this analysis?
After all screening, a total of 10,307 unique station pairs were accepted for analysis from 2000.
Sorry, re-read it. meaning from year 2000, I was expecting to be told from how many stations.
If possible please withdraw these posts.
Regards Michael Oakes
Two questions on an important piece of work
(1) I guess the UHI effect has increased over time, so that a town of 10,000 in the year 2000 has more UHIE than a town (same density) of 10,000 in the year 1900. How do you propose to adjust for this?
(2) The maximum UHI you show is ~2ºC. How do you account for Anthony’s urban transect at Reno which showed a UHIE of ~6ºC IIRC?
It would seem to me that population density could be one factor, while another factor could be surrounding area. For example, if there is one square kilometer with 200 people living in it, and surrounding that square kilometer there is natural wilderness, that would seem different than if there is one square kilometer with 200 people living in it, and surrounding it is a thousand more square kilometers with 200 people per square kilometer. If a station is sited on the edge of a large metropolis, the station itself may be in a low population density square kilometer, yet could be effected by the neighboring metropolis. (Maybe you’ve already accounted for this?)
Ivan, if a station is Rural, and still has 100 years of data, it hasn’t been doing much growing. In fact, when you get down to “small towns,” probably half, or more, are smaller than they were 50 years ago.
Just curious, but were there any regional variations in the relationship? The reason I ask, is that there was a discussion here recently where someone was talking about electrical consumption in NYC as a potential forcing. It was an interesting thought and one I hadn’t really given proper consideration to before. There is a drastic difference, for example, between world average (2 kW/capita) and those wasteful (joking!) Canadians (11 kW/capita) *
There’s also the long term up trend in total electric consumption from ~5 TW in 1965 to 15 TW in 2005 *
To get to the point: Is there was any signal from heat generation due to electrical consumption and usage patterns that could be ascertained, or if the UHI effect was effectively consistent from region to region as a direct relation to population?
“Each data point represents a population bin average for the intersection of a higher population station with its lower-population station mate.”
This doesn’t make any sense. What is a “population bin”? How do stations “intersect”? I hesitate to venture a guess.
Also on graph 1, you mention warming per pop density increase. What time period? On the x-axis, is this the population density before or after the increase in population density?
Also, somewhere you need a rationale for why you’re trying the methodology. Even if you’re just looking for correlation, you need to give people a rational basis for why this reveals some causation. Otherwise, it’s correlation of pirates to global warming all over again.
Also, how are you assigning population density to a point? Some cities are larger than others but the population density where the station is located may not be representative. Have you ensured that the stations are located in the center of what would be expected to be a heat island?
Define “warm bias”. How is it calculated? What is zero warm bias?
” I used an old USAF 1/6 deg lat/lon percent water coverage dataset to ensure that there was no more than a 20% difference in the percent water coverage between the two stations in each match-up.”
What was done if there was more than a 20% difference? The pair was dropped?
10,307 unique pairs = how many unique stations?
How many of these from the US?
How many US stations met the NASA citing criteria?