Urban heat confuses deniers at WUWT

Today Anthony Watts has copied and pasted a press release about urban heat (archived here.) A team led by Lahouari Bounoua of NASA's Biospheric Sciences Laboratory has completed an analysis of the impact of urbanisation on surface temperatures across the USA. Anthony said it was "obvious", implying there was no value to the research. He's wrong. The research quantified what to Anthony might have seemed "obvious", and came up with some less obvious findings as well.

The bottom line is, if you live in a city that gets hot, do what you can to get plenty of trees and shrubs planted. The more the better.

This work isn't to be confused with factoring in the impact of urban heat on the mean global surface temperature anomaly. What it is about is how urban centres are mostly hotter than the surrounding rural centres. It says nothing about differences in the rate of warming, or trends over time.

The paper is published in Environmental Research Letters and is open access.  (It's taken a while for the research to be completed from the look of things. Some of the authors presented early findings at the AGU Fall Meeting in 2009.)

Vegetation, tar and cement - and surface temperature

The research pointed to vegetation as the main factor determining how much difference there was between the temperature in cities and in surrounding rural areas. There's not a gradual effect, there are plateaus and leaps. From the press release at ScienceDaily.com:

Impervious surfaces' biggest effect is causing a difference in surface temperature between an urban area and surrounding vegetation. The researchers, who used multiple satellites' observations of urban areas and their surroundings combined into a model, found that averaged over the continental United States, areas covered in part by impervious surfaces, be they downtowns, suburbs, or interstate roads, had a summer temperature 1.9°C higher than surrounding rural areas. In winter, the temperature difference was 1.5 °C higher in urban areas....

...The urban heat island effect occurs primarily during the day when urban impervious surfaces absorb more solar radiation than the surrounding vegetated areas, resulting in a few degrees temperature difference. The urban area has also lost the trees and vegetation that naturally cool the air. As a by-product of photosynthesis, leaves release water back into to the atmosphere in a process called evapotranspiration, which cools the local surface temperature the same way that sweat evaporating off a person's skin cools them off. Trees with broad leaves, like those found in many deciduous forests on the East coast, have more pores to exchange water than trees with needles, and so have more of a cooling effect.

Impervious surface and vegetation data from NASA/U.S. Geologic Survey's Landsat 7 Enhanced Thematic Mapper Plus (EMT+) sensor and NASA's Moderate Resolution Imaging Spectroradiometer (MODIS) sensors on the Terra and Aqua satellites were combined with NASA's Simple Biosphere model to recreate the interaction between vegetation, urbanization and the atmosphere at five-kilometer resolution and at half-hour time steps across the continental United States for the year 2001. The temperatures associated with urban heat islands range within a couple degrees, even within a city, with temperatures peaking in the central, often tree-free downtown and tapering out over tree-rich neighborhoods often found in the suburbs.

The northeast I-95 corridor, Baltimore-Washington, Atlanta and the I-85 corridor in the southeast, and the major cities and roads of the Midwest and West Coast show the highest urban temperatures relative to their surrounding rural areas. Smaller cities have less pronounced increases in temperature compared to the surrounding areas. In cities like Phoenix built in the desert, the urban area actually has a cooling effect because of irrigated lawns and trees that wouldn't be there without the city.

"Anywhere in the U.S. small cities generate less heat than mega-cities," said Bounoua. The reason is the effect vegetation has on keeping a lid on rising temperatures.

Bounoua and his colleagues used the model environment to simulate what the temperature would be for a city if all the impervious surfaces were replaced with vegetation. Then slowly they began reintroducing the urban impervious surfaces one percentage point at a time, to see how the temperature rose as vegetation decreased and impervious surfaces expanded.

What they found was unexpected. When the impervious surfaces were at one percent the corresponding rise in temperature was about 1.3°C. That temperature difference then held steady at about 1.3°C as impervious surfaces increased to 35 percent. As soon as the urban impervious surfaces surpassed 35 percent of the city's land area, then temperature began increasing as the area of urban surfaces increased, reaching 1.6°C warmer by 65 percent urbanization.

At the human level, a rise of 1°C can raise energy demands for air conditioning in the summer from 5 to 20 percent in the United States, according the Environmental Protection Agency. So even though 0.3°C may seem like a small difference, it still may have impact on energy use, said Bounoua, especially when urban heat island effects are exacerbated by global temperature rises due to climate change.

Below is Figure 2 from the paper, showing how Phoenix differs from Chicago and Kansas City. Unlike elsewhere, Phoenix is cooler than the surrounding rural areas. For each of three cities, the charts shows the mean day and night temperatures over summer (June, July and August) for the co-existing urban class, vegetation class with the highest fraction, and the weighted average temperature of all vegetation classes (excluding urban).

Figure 2: Summer (June, July and August) mean land surface (skin) temperature diurnal cycle for the urban class, the vegetation class with the highest fraction, and the weighted average temperature of all vegetation classes (excluding urban) co-existing within the climate modeling grid for (a) Chicago, (b) Kansas City, and (c) Phoenix. Source: Bounoua15

Convection vs evapotranspiration

This new paper contrasts with another bit of research I wrote about last year, which was published in Nature (Zhao14).  The results are similar in that both found Phoenix to be cooler during the day than its rural surrounds, but other cities were hotter. However the interpretation of why that is the case was different. In the earlier study, what Lei Zhao and colleagues argue that it wasn't just evaporation or evapotranspiration, it was the nature of the surfaces that made a difference via convection. They determined that in wetter areas, the smooth surfaces of buildings and roads etc are much less conducive to heat diffusion than the vegetation around the cities, so the cities are hotter than the surrounds. In dry areas by contrast, the surrounding scrubby vegetation is less effective at diffusing heat than the surfaces in the cities, so the cities are cooler than the surrounds.

I'd say that despite the differences in interpretation, both studies would suggest that planting urban areas with lush vegetation would be expected to lower the temperature.

From the WUWT comments

As you'd expect, many of the people at WUWT couldn't distinguish between urban heat effect as a straight comparison between city and surrounds, and the difference in trend of heating, which would happen as urban centres grew. In other words, they don't understand the concept of anomalies from a baseline. Once cities stopped growing, the anomalies over time would be less or non-existent. That is, while a city might be hotter than its surrounds, once a city has stopped growing or changing, the increase in temperature from one decade to the next would probably be the same as it is for the rural surrounds.

Mark from the Midwest is wrong. It's not as obvious as that, is it.

August 25, 2015 at 5:46 pm
“Everybody thinks, ‘urban heat island, things heat up.’ But it’s not as simple as that
B.S. it is as simple as that, sure it’s a complex mess to disentangle the amount of heat, and given the poor statistical controls that could exist in any model it’s really tough, but it’s as simple as that. Cut down trees and create roads, then build houses and things heat up.

Mike Jonas has more than one thing wrong in his short comment. First of all, global mean surface temperature trends are anomalies, so whether something is "extended" or not makes no difference, except if the regions are changing over time (urban areas or vegetation in rural areas). Secondly, satellites don't measure surface temperature. Their instruments can be used to calculate atmospheric temperature - and even that is in slices, not at a point, so it's a bit rough. (The satellite instruments don't measure temperature directly, unlike thermometers on the ground.)

August 25, 2015 at 6:49 pm
True. Physically, the urban area is a small proportion of the whole, but they have a high proportion of the thermometers and are extrapolated to areas without few or no thermometers, so their impact on global temperature is out of proportion to their area. For global temperature, it’s now best to use the satellite measures.

JohnWho didn't bother reading the paper, which looked at vegetation cover.

August 25, 2015 at 6:39 pm
“Anywhere in the U.S. small cities generate less heat than mega-cities,” said Bounoua. The reason is the effect vegetation has on keeping a lid on rising temperatures.”
Gee, wouldn’t the reason be that the smaller cities simply aren’t as large as the “mega-cities”?

Alf, like many WUWT-ers, confused the UHI effect on trends with the difference in temperature between cities and their rural surrounds.

August 25, 2015 at 7:04 pm
Check Berkeley Earth; they claim that uhi is between 0 and .14 degrees. Who to believe?

So did M Seward, who wrote:

August 25, 2015 at 8:19 pm
“The study, published this month in Environmental Research Letters, also quantifies how plants within existing urban areas, along roads, in parks and in wooded neighborhoods, for example, regulate the urban heat effect.”
It also likely quantifies the effect on the so called global warming of the UHI corrupting the non UHI ‘signal’. Maybe the ‘climate scientists’ are just getting round to realising the magnitude of the data corruption problem and some are now running interference by distraction, turning merde into manure so to speak. 

Similarly, tomwys1 hasn't grasped the concept of anomalies. Nor does he have the first clue about the distribution of weather stations. When people try to make fun of something they don't understand, it makes them look dumb, not clever.

August 25, 2015 at 8:43 pm
Elsewhere on the planet, vast uninhabited areas without temperature stations are awarded the temperature designations of the two Urban Heat Islands (averaged, of course) at either end of their expanse.
I think that NASA and its reputation have been hijacked!!!
The “Space” in the NA”S”A acronym used to refer to where you might find the Moon, Planets, and Stars, not zones between Earthbound temperature sensors. 

References and further reading

Lahouari Bounoua, Ping Zhang, Georgy Mostovoy, Kurtis Thome, Jeffrey Masek, Marc Imhoff, Marshall Shepherd, Dale Quattrochi, Joseph Santanello, Julie Silva, Robert Wolfe, Ally Mounirou Toure. "Impact of urbanization on US surface climate". Environmental Research Letters, 2015; 10 (8): 084010 DOI: 10.1088/1748-9326/10/8/084010 (open access)

• Vegetation essential for limiting city warming effects - press release at ScienceDaily.com

Lei Zhao, Xuhui Lee, Ronald B. Smith, Keith Oleson. "Strong contributions of local background climate to urban heat islands". Nature, 2014; 511 (7508): 216 DOI: 10.1038/nature13462 (pdf here)

• Urban heat: Not a myth, and worst where it's wet - press release at ScienceDaily.com
• Urban heating and what drives it - HotWhopper article about the above