by Roy W. Spencer, Ph. D.
This is an update to my previous posts [here and here on WUWT] describing a new technique for estimating the average amount of urban heat island (UHI) warming accompanying an increase in population density. The analysis is based upon 4x per day temperature observations in the NOAA International Surface Hourly (ISH) dataset, and on 1 km population density data for the year 2000.
I’m providing a couple of charts with new results, below. The first chart shows the global yearly average warming-vs-population density increase from each year from 2000 to 2009. They all show clear evidence of UHI warming, even for small population density increases at very low population density. A population density of only 100 persons per sq. km exhibits average warming of about 0.8 deg. C compared to a nearby unpopulated temperature monitoring location.
In this analysis, the number of independent temperature monitoring stations having at least 1 neighboring station with a lower population density within 150 km of it, increased from 2,183 in 2000, to 4,290 in 2009…an increase by a factor of 2 in ten years. The number of all resulting station pairs increased from 9,832 in 2000 to 30,761 in 2009, an increase of 3X.
The next chart shows how the results for the U.S. differ from non-US stations. In order to beat down the noise for the US-only results, I included all ten years (2000 thru 2009) in the analysis. The US results are obviously different from the non-US stations, with much less warming with an increase in population density, and even evidence of an actual slight cooling for the lowest population categories.
The cooling signal appeared in 5 of the 10 years, not all of them, a fact I am mentioning just in case someone asks whether it existed in all 10 years. I don’t know the reason for this, but I suspect that a little thought from Anthony Watts, Joe D’Aleo & others will help figure it out.
John Christy has agreed to co-author a paper on this new technique, since he has some experience publishing in this area of research (UHI & land use change effects on thermometer data) than me. We have not yet decided what journal to submit to.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
So even a population of only 100 per Sq km shows an average warming of 0.8C over a nearby location with little development.
But what about the recovery from the Little ice age, or more active sun, surely if we take this into account we may have had a slight cooling effect over the last 150 years?
Afterall we now have Phil Jones of cru fame telling us that there has been no statistical warming over the last 15 years and the MWP may have been as warm as today, I mean where does it end?
cooling?
raw data or fudged data?
You may have found one more way to prove data corruption.
Tim L
A nice population site http://sedac.ciesin.columbia.edu/gpw/index.jsp , boy do India and China have issues!
James Sexton (12:08:01) :
“The US results are obviously different from the non-US stations, with much less warming with an increase in population density, and even evidence of an actual slight cooling for the lowest population categories.”
Am I reading this right? Did I misinterpret your original analysis? Doesn’t the above statement run counter to the original work? I thought we saw greater warming in rural areas with an increase in population density?
This might not be as counterintuitive as it seems… albedo analysis dictates that you should see cooling if/when forest is cleared. So if development in the US, say more than other areas, begins by cutting down trees then it could explain this effect. I’d be curious if there is a difference in this effect, say, between Oregon and Arizona or New Mexico.
We might have stumbled onto a completely novel variable in the equation – give it time
This is a really insightful analysis, and I can’t wait to see it published.
That said, I’d very much like to see this analysis done with latitude, altitude, and proximity to (significant) water included as potentially explanatory variables. It seems to me that low-density areas that are close to high-density areas tend to be located in “getaway” areas – mountains, oceans, or just areas preferentially to the north of the big city. It could also be that population density tends to drop as you go toward the poles, creating a bias for “low density” to be associated with “cooler” via latitude.
I doubt that accounts for the basic findings, but it seems prudent to take the extra step prior to publication (it might also resolve the oddball US data). And if data like “distance from water” is not readily available, it might be a good opportunity for crowdsourcing the work – to, perhaps, readers of this blog.
Still confused regarding rest of world vs US response with regard to population density and UHI effect. Does the US have better sited thermometers, thus minimizing the UHI impact versuse rest of world (maybe)? Does the US have more open space between urban/suburban areas that minimize UHI effect (more likely)? Also, does re-forestation of many rural areas in the US (plus having wealth to afford more green areas than rest of world) also causing the discrepancy (even more likely)? Thus, maybe we should look at more than two lines (i.e., US vs rest of world), such as 5 lines of stations at different population densities at DIFFERENT GDP levels.
ot/ How long before the hatchet mob arrives do you think 🙂 http://en.wikipedia.org/wiki/Wattsupwiththat
Robin Edwards: “To me this means that if a series of actual temperature measurements is made at Town, and then reported in the literature as 10.0, 11.0, 9.4 and 7.1C, I, the user, should put them into my computer as 8.6, 9.6, 8.0 and 5.7C . Is this wrong?”
I don’t think it is necessarily as simple as that. One factor that has a big influence on the amount of urban warming is the wind – when there is a howling gale blowing, the temperature difference between a city and the surrounding countryside will be very small. The biggest difference in temperature will occur in still conditions, especially on frosty cloud-free winter nights.
As we know only too well in the UK, the amount of wind can vary a lot between years. Some of the very mild winters that the UK received in the early 2000’s were largely a result of strong westerlies dominating the winter weather patterns, but more recent and cooler winters have had a lot less wind. As a result, UHI effects would be quite different between those years.
Roy Spencer (12:21:40) :
“James…my original results were based upon the year 2000 only, and (as I recall) global data…not US-only.
Also, I have changed to log-plots, since (as you mentioned) warming in the global average IS much stronger at low population densities.”
Dr. Spencer……thanks for clearing that up for me, for some reason, I was under the impression the first analysis was U.S. only. I went back and read it again, yep, I was wrong.
So, naturally, it comes to a question, (and perhaps tangential), did you notice any trends regarding the development of the nation or area, such as 3rd or 2nd world nations? (Those terms are probably too ambiguous and subjective.) I would think other similar developed nations such as the U.S. would trend in a similar manner as the U.S.
Dirk,
The UHI data shown (y axis) is not a time series (one year compared against another year). Instead, it’s the delta between two stations within 150 km of each other of differing population densities. Thus, gdp growth rates between 2000 and 2009 are irrelevent. However, average gdp for 2000-2009 for each station pair is relevent (and should be compared against one another and reported, as noted by me in an earlier post).
My understanding of UHI was is that higher overnight or minimum temperatures are a hallmark of the UHI. Lower minimum temeratures than flow through to higher mean temperatures in the commonly cited data sets.
It would be interesting to see the night time (or minimum) temeratures compared with daytime (or maximum temperatures for these 4hrly temperature datasets.
In deveoping a “general theory” of UHI one would have to contemplate heat generation by urban communities of differing density, which might have complex or unexpected relationship to population density, and different seasonal characteristics. Economic factors might also drive UHI.
Dr. Spencer,
Sorry to go OT, but I’ve been pondering a point you raised in your post analyzing station dropout (your raw ISH analysis vs. CRU), and I think I might have gotten the importance of the 36% difference in variability… would that impact the statistical significance of the temperature trend?
I wonder if you need more detailed information on urban land use, especially as pertains to US vs. non US. I suspect that would account for some difference. Take two cases:
1. A shantytown covering all the land and leaving no vegetation in Logos, Nigeria.
2. A High rise, housing the same number of peoplebut surrounded by vegatation, in Munich or New York.
I know its intuitive, but I would expect to see a difference.
The preceding post wasn’t there when I submitted. So I’ll add that it seem Nick B. and I are having the same thought.
DirkH (13:17:56) :
“For the fact that the U.S. behaves differently than the rest of the world, i suspect it might be because the U.S. already had a high per capita energy consumption in 2000 which might not have risen as much during the last 10 years as per capita consumption in less developed countries. It should be reflected in the high GDP growth rates of, say, the BRIC countries. Just a guess.”
Following your thought: it seems this might point to the U.S. actually cutting their influence during the 2000 decade (most U.S. citizens have had a huge guilt trip laid on them so they are cutting big-time) while the rest of the world goes along it’s merry way with city construction, big-time. For instance, I run my AC one-third what I use to in the 70-90’s, I heat my house to 60F instead of 74F like I used to, get the drift? Most American homes have cut way back, also added insulation, planted trees, putting a lighter colored roof on, etc. Seems this would have a negative influence when comparing U.S. to global UHI spreads.
Hello. Just wondering if any difference was noted by longitude. The reason I ask is because by using zulu time at 6 hour intervals you may miss out on the Max daily temps. I don’t know if that would make a difference or not. Thanks
“Chris (14:13:10) :
Dirk,
The UHI data shown (y axis) is not a time series (one year compared ”
I know. But over the examined decade the growth rates of developing countries were on average higher than in the US – they were constantly more dynamic than the US. How this can creep into Dr. Roy’s observation is not entirely clear to me, wait… reading the article…
Dr. Spencer uses 1-year-warming averages against population density increase. That’s where the effect of growth over time creeps in or the dynamism of an economy/population.
My PhD is in math, not statistics or the physical sciences, but one look at your first spaghetti graph convinces me that the argument you’re making is compelling. While the hypothesis and basic empirical facts are old hat, the analysis itself appears to be quite new. It’s a wonder it hasn’t been already done in this level detail, but I guess the billions spent on studying and massaging climate data didn’t have a few dollars left over for this, which evidently can be done with minimal resources. Well done Roy. In my view this should be submitted for publication with all due haste.
On behalf of those interested in the US-only dip in low populations I would request that at some time you make clear which years display this anomaly, and predominantly at which locations. Immediately I wonder, is it always the same locations? Was it a cluster of years, or random? Perhaps Anthony knows of some move to refit rural instruments in certain years that may have had some effect. Also, someone should check that the U.S. data is truly raw source, untouched, homogenized, diced or sliced, which by itself could explain anomalous behavior like this. I’ve no doubt this sort of answer will pop out in the mix, given enough detailed information.
I’m sure Zeke Hausfather on Lucia’s blog will pull another magic graph from his high hat….
Shouldn’t the ‘Global Average’ be almost on top of the ‘non-US average’ ? The US is a tiny fraction of the globe, right? I assume these averages are area weighted (?)
“…new technique for estimating the average amount of urban heat island (UHI) warming…”
what is the proportion of global UHI (joules) to the total amount of energy stored on Earth. If it is statiscially significant, this supports the view that man is warming the planet.
“wayne (14:29:24) :
[…]
Following your thought: it seems this might point to the U.S. actually cutting their influence during the 2000 decade (most U.S. citizens have had a huge guilt trip laid on them so they are cutting big-time) while the rest of the world goes along it’s merry way with city construction, big-time. For instance, I run my AC one-third what I use to in the 70-90’s, I heat my house to 60F instead of 74F like I used to, get the drift?”
I can relate. I’m German, we’re taxed to the hilt and thrift is our way to survive. Energy-wise, we’re both pretty much stagnant economies, i think. (The renewables hype doesn’t really add substantial amounts of energy for now)
BTW, re the guilt-trap: At least you can walk away when your house is under water. In Germany, the bank will auction off your house and use the money to reduce your debt. You still owe them whatever remains – so when you’re not paying your mortgage you lose the house and remain indebted to your bank. So don’t complain, you don’t know how good you have it over in the USA. Buying a house in Germany is plain madness when you have to take out a mortgage in my opinion and i don’t understand people who do it.
You might have covered this, but a population density of 10 per sq km means scattered people avergage 300 m apart in a 1 sq km cell and there is plenty of room for a station with no UHI effect.
However, an adjacent 1 sq km cell might have a people density of 1000 per sq km and so if the wind blows the right way, it might affect the zero UHI cell.
It is valid to recalculate, by selecting a weather station and calculating the people density for the larger cell of (say) 100 sq km surrounding it, rather than a cell of 1 sq km? Then compare pairs of cells of area 100 sq km, keeping the rest of the methodolgy as is? Yes, it produces other complications such as land:water ratios etc, but I worry that a 1 sq km cell is too small a unit. It is the main way that I can envisage for the graphed strong rise in delta UHI with low population density pairs.
Tom G(ologist) (12:32:39) :
So-called green rooves would also make a difference in urban settings. These are things which are easily implemented and don’t disrupt civilization.
For some reason the term “boiling the ocean” comes to mind, although in this case we’d be talking about “cooling the globe”. I think a more reasoned approach would be to look at matching roof albedo with the prevailing temps for an area. Areas that where buildings have a majority heat load, go for black roofs… areas with mostly cooling loads go with white roofs… areas in between go with “green” roofs (thermal mass) and you might see some significant changes in power consumption for building HVAC loads. No sense putting highly reflective roofs where it’s always cold outside and you’re running the heat all the time. Actions taken to increase energy efficiency are all around win-wins as long as they’re economically viable (or at least reasonable).
Dirk, Chris
GDP vs. Power Consumption is highly variable per capita between countries. Ref – so level of GDP as a direct relation to UHI might not be very useful… but I think what you guys are getting at is the change, or perhaps rate of change in GDP for specific areas could be a good variable to use. Very interesting!
/starts digging in garage for dusty old econometrics textbook
Dr. Spencer, do you think you could do a quick ‘how we derive temperature’ from the satellite radiometers?
I can’t for the life of me work out how you can convert the ‘IR heat’ signal into temperature.