Scientific Congress on Climate Change
International Alliance of Research Universities
Copenhagen, Denmark, March 2009
The Factors of Urban Morphology in Greenhouse Gas Emissions:
A Research Summary
Michael Mehaffy, Presenter (With co-authors Stuart Cowan, Diana Urge-Vorsatz)
(Oral version as presented - Download PowerPoint Presentation as read-only here.)
(S) Denotes slide change
I’ll pick up on the previous points about building energy use by extending the discussion to urban systems. And I want to report to you compelling evidence that urban morphology – that is, the form of our settlements – is now and will continue to be one of the most important drivers of energy use and of greenhouse gas emissions per capita, in the critical decades ahead. Yet because of its complexity and pace, it may also be one of the most under-appreciated.
We can think of our cities and suburbs as energy-using systems in their own right, with their own systemic designs, and their own characteristic efficiencies and inefficiencies. They are, to be sure, very complex systems, and they involve many interacting variables, notably the particularly challenging one of human behaviour. Nonetheless we are now able to draw some fairly solid conclusions from this research about the importance of urban form, and its components. So this presentation will summarise my co-authors’ and my paper, which will give an overview of this maturing and, we think, very important topic.
First off, (S) we know that urban density – that is, the number of people per hectare - is very powerfully associated with lower energy use and greenhouse gas emissions. In fact density as a single factor is associated with such a dramatic reduction in energy and carbon, that it can by itself exceed the effects of some building technologies and building rating schemes, like the Energy Star program in the US. Here is a chart of urban density (S) and potential energy savings, (S) as measured in research by Allen and colleagues – other researchers have similar data. As you can see, going from 6(S) to 12(S) units to the acre, or about 2.5 to 5 to the hectare, gets us a doubling of savings. Going all the way (S) up to 100 persons per acre (about 40 to the hectare) gets you another doubling – though as you see the curve starts to flatten out at these higher densities. But as you see (S) that greatly exceeds the Energy Star savings.
But we know that density is not a factor all by itself, but contains other co-factors that are associated with it. So we need to tease out those factors and begin to understand how they all work within the urban system.
Let me start with an instructive example of two strikingly different morphologies within one region, the San Francisco Bay area of the USA. (S) This is a helpful example because it eliminates major factors such as climate, regional economy, regional culture, income disparity and other factors that often cloud the picture. This is a map of density in the region, with red as the highest and blue as the lowest. And as you see the highest density is in the northeast part of San Francisco, with much lower density in the East Bay and other peripheral areas.
(S) So here’s another map, this time showing daily averages of CO2 emissions per household just as a result of transportation, including car, rail, bus and other modes. This is from research by the Bay Area Metropolitan Transportation Commission, conducted in 2007. This is showing a range from less than 17.6 pounds per household, in the dark green, to over 53 pounds per household, in the red – that is a threefold range, or a tripling from the lowest to the highest.
(S) You can see very clearly that the highest density areas are in fact the lowest emissions areas, and vice versa. (S-S-S-S-S-S). But it is very instructive that we can in fact see from this example, what other variations there are than just urban density.
We know, for example, (S) that the San Francisco neighborhoods are well-served by public transit, they have a well-connected street grid, are walkable, they have a diverse mix of uses, with a range of daily needs within close range. And they have other advantages besides transportation: they tend to have smaller homes in attached dwellings, they tend to have shorter lengths of infrastructure such as lengths of transmission line per person, they tend to have a relatively small quantity of pavement per person, and so on. And we’ll get into some of these factors in more detail.
By comparison, (S) the urban morphology in the East Bay area with its significantly higher emissions is strikingly different. In addition to being lower density, It’s much more fragmented, unwalkable, poorly served by transit. Most trips must be taken by car, and the distance of such trips is likely to be much greater on average. The building typology tends to rely more on single family detached dwellings that tend to be larger and have more consumption per person.
And of course on a per capita basis, there is much higher infrastructure, much higher land use, much greater impact on ecosystems and ecosystems services like water purification, and so on. There is also likely to be a systemic effect on behaviour, as people increasingly turn to all the amenities offered by an auto-dependent lifestyle: the convenience retail, drive-through facilities and so on. As we know, many of these are also very high-energy and high-emission facilities.
I’ll just mention here a key point from a policy point of view, that the true external costs of this urban pattern are not being paid by these consumers. So in effect, these costs are being passed on to others elsewhere or in the future, and there is in effect a subsidy for this kind of high-consumption development pattern. The result of that is, that this appears to be an appealing, cost-effective choice for development around the world.
By the way, let me just add that when the chickens come home to roost for these costs, there are major economic consequences – and that is precisely what has happened with the current banking crisis. It began with a wave of mortgage defaults in exactly this kind of sprawling, car-dependent neighborhood, and when rising fuel prices combined with adjusting mortgage rates and a normal cyclical downtorn, it created a perfect economic storm. So there is indeed a link between these long-term environmental issues and the health of our economies – and in this crisis I suggest we’re already seeing the first clear warning signs of that.
But if we look at the developing world, we can see that this kind of high-consumption, auto-dominated infrastructure is being built right across the world, in (S-S-S-S) Brazil, China, India and Romania. Because such infrastructure is locked in for many years, the consequences for long-term emissions are grave.
This is essentially an American model of development, but like many American exports, it is being adopted by the rest of the developing world very quickly. To illustrate the problem I’ll show an example of drive-through McDonalds restaurants, again in (S-S-S-S) Brazil, India, China and Romania. I don’t mean to pick on McDonald’s here, but it’s a good proxy for other very similar development around the world. And this represents a huge problem when we are talking about trying to get reductions in energy use and in emissions.
(S) And indeed, looking at motor fuel, the difference between the American levels of energy use and that of other regions is striking. As you can see, the level of use in Europe (S) is almost one-fourth that of the US per capita, and yet their cities have a comparable or even standard of living.
What is the likely impact of this kind of development trend over time? We can begin to get a sense of it (S) by looking at the IMF’s projection of global car fleet size, from about 650 million today, to three billion by 2050 – a staggering five-fold increase. Clearly, given the magnitude of automobile energy as a percentage of the total, this is a huge potential movement in the wrong direction, with regard to energy and emissions.
(S) But we must remember that when it comes to overall impact of all these factors, so-called tailpipe emissions are only the beginning of the calculation. I’ll start with the transportation sector first, and then talk about the others next. So we need to include the embodied energy of the production and transport of fuel; the embodied energy of the street infrastructure; the manufacture and maintenance of the vehicles themselves; and the leakage of air conditioning coolant. Taken together, these factors are fully half again the tailpipe emissions and a full one-third of the total energy picture for private automobile transport.
(S) So transportation efficiency is a major factor in energy use and emissions. But aside from the efficiency of the vehicles themselves, what are the elements of the urban form? (S) One of those is the efficient distribution of daily needs and activities within a close average spacing. We want to be able to keep our daily trips short, to work, to shopping and so on. (S) We know that another factor is the availability of effective, safe and convenient transit. (S) This graph shows that not only residential density, but the zonal density of transit, is a factor, particularly at lower residential densities.
We can begin to tease out other additional factors relating to urban form, that don’t relate strictly to personal automobile transport. (S) For example, walkability is very important, (S) as is bikability -- perhaps the lowest-emission and lowest-energy form of transport, even when fuelled by meat. (And we have to consider this, after all, if we’re looking at whole systems.) Related to that (S) is the integrity of the urban network: whether it is well-connected in a dense and efficient network, or whether it is hierarchical and fragmented, which lengthens average trips. And (S) whether the network is (S) friendly to pedestrians and bicyclists, or not. Obviously, these things matter!
Then (S) there is the very important topic of infrastructure – its embodied energy in construction and maintenance, (S) its operating energy, for things like pumping, lights, signals and so on, and (S) its transmission losses. Obviously the longer these have to be, and the lower the density, the more burden carried per capita. And (S) there is the very important topic of district energy and cogeneration, which is feasible at higher densities, and can produce very significant savings per capita.
Then (S) another important category is the association between urban morphology and building morphology. (S) Virtually by definition a denser urban morphology will have fewer single-family detached dwellings, and more attached and multi-family dwellings. These are inherently more energy-efficient in their number of common walls, (S) in their typical embodied energy, and in the fact that they tend to be more compact. (S) Why they are more compact is an important question, which seems to have to do with the economic desirability of urban locations and the greater cost as a result.
I’ll jut mention here an important caveat in this question of density of building type. It does not necessarily follow that if a density of x is beneficial, a density of 2x is twice as good. Indeed, we already noted a levelling off of the benefits over about 100 person per hectare, and we see a number of negative effects start to kick in, particularly with tall buildings. One of those is in the egress requirements of these structures, and the resulting loss of floorplate efficiency. Another is in the embodied energy of the materials required for their construction, which tend to use a much greater amount on average of steel and concrete per unit of construction. There are exposure issues for the typical curtain wall assemblies that these buildings employ for exterior enclosures. And there are heat island effects as well.
But this topic deserves more research and is really beyond the scope of this paper – except to note that density is somewhat like aspirin: if two are good, ten are not necessarily better.
The next category we will discuss here (S) is in so-called “externalities” – such factors as ecosystem services, like water purification and CO2 removal by vegetative cover, which may be lost when low-density development destroys those ecosystems. (S) Loss of croplands is another external cost that may have impacts down the road. And finally (S) albedo effects, which increase the greenhouse phenomenon, and heat island effects, which raise ambient temperatures and increase demand on cooling systems. Bear in mind again, we are looking at this per capita, so a dispersed urban pattern may not be as conspicuous as a dense urban area but may in fact be worse overall because it’s worse per capita.
The last category, (S) and one of the most difficult to assess – but still perhaps one of the most important - is what we might call “indirect systemic factors”. (S) These include behavioural factors such as the tendency to use more of something that’s already available. A related phenomenon (S) is that of “induced demand” – as systems get more efficient, they tend to get less expensive, and that only tends to increase consumption right back to where we were. Or highways that make it easier to drive, and then encourage more drivers, and then increase traffic right back to the congestion levels that the highways were built to alleviate. There is very good literature on these effects, but they are often not taken into account sufficiently when assessing conservation strategies – in large part because they can be difficult to predict.
And we will only mention what is a potentially very significant subject, (S) performance over time, and so-called “resilience” – ability to adapt and deal with change and stress on the system. If we are relying on very delicate technological solutions that become quickly obsolete, or that can’t be easily repaired, then we may be in trouble over time. If our buildings and our urban environments don’t weather well over time, or don’t accommodate change and variety, then we are going to have a problem. Obviously if we are too dependent on single sources of fuels, like oil, and those sources become unavailable, we don’t have a very resilient system. Conversely, if we have very sturdy, easily repairable, robust buildings and urban environments, that can still work reasonably well with a variety of power sources, or even without a lot of power, then we are likely to have a more resilient city.
Another aspect of this discussion (S) is the question of the livability of an urban environment, and its appeal at relatively high densities – particularly the optimum range that seems to be at about 100 persons per hectare. If we want people to live in these environments, by and large they will need to want to live in these environments - not only because they are economically vital, but because they are desirable neighbourhoods. That means they will have to follow the principles of attractiveness and beauty and appeal that human beings respond to. There is a very interesting area of biological science, so-called “biophilia”, that is revealing very interesting things about the most successful and responsive environments from a human point of view – their qualities of natural structure, plants, light, air, and so on. And related to this, there is the topic of “evidence-based design,” coming from the field of patient care, which suggests that we can learn to adapt our human environments much more successfully to well-being and health.
So these are all promising areas for conservation, and for putting ourselves on a much lower-carbon footing globally. (S) If – and it’s a very big if – if we could replace the American sprawl model with, for example, something more like a European (S) urban morphology, the savings could be very, very significant, with reason to suspect the economic vitality would not be compromised, and may even be enhanced long-term.
(S) As a very rough idea of that possible magnitude, here is an assessment, based on the research that’s presented in the paper. I should qualify that this is a very rough assessment, and we really need more work to quantify this more accurately. And this may be a maximum target that will be hard to achieve, at least for some time into the future. (S-S-S-S-S-S) We’ve gone through this exercise in the paper looking at the research in each case, though I’m afraid I don’t have time to go through it in detail here,
But nonetheless, as you see we could be talking about as much as one-third of all energy use. That’s not just the energy use associated with urban activities, but all energy use – manufacturing, consumer goods, industry, losses from generation, and so on. So this is a very, very significant magnitude.
As for strategies to achieve these reductions, there is some very good work on this as well, and again I only have time to summarise here what is a more thorough discussion in the paper.
(Read slides)
So let me conclude by saying that as we look to conserve energy and mitigate the effects of greenhouse gas emissions, we need to be very careful not to look only at individual users of energy, but to look also at the urban systems that shape the patterns of use. Though this is a more difficult area to investigate and to manage, I think you can begin to see here that it is a crucially important one. Though urban systems operate slowly, it is for precisely this reason that their cumulative potential over time is enormous. We can think of urban systems as our “operating systems” for the use of energy, and for the growth of energy-using systems. If we can change the underlying operating system, we can change the global pattern of growth in a very effective way. Thank you.