When we think of cities, and especially of their future, a recent South China Morning Post headline sums up well what typically first comes to mind: "the rise and rise of the Asian megacity." And for good reason: most of the world's largest urban areas are now located in the global South. That map also indicates the other important piece of the story, which is that the urban population in those same areas is, as a percentage of the whole, still the lowest in the world -- leaving plenty of room for growth that is progressing at a rate twice as fast as it did in the U.S. A number of metropolitan areas around the globe now sprawl over more than 5,000 square km. Smaller cities and towns, though, are large enough to produce a whole network of microclimates worthy of study as well. Even in the absence of climate change, cities have substantial impacts on conditions in their vicinity (e.g. pollutant levels, temperatures, precipitation, and winds). On the other hand, globally driven but locally amplified climate features like sea-level rise, coastal storms, and extreme heat make cities vulnerable — often more so, in fact, than nearby rural areas. In certain regions, the list expands to include drought, tornadoes, floods, and extreme cold. Each of these types of features has a characteristic ratio of global-local interplay, making long-term projections hard and comparisons between features even harder.
The environmental challenges faced by cities can be sorted into two kinds that historically have been reacted to differently. When the damage is perceived as short-term or exceptional, cities are typically rebuilt despite the devastation, whether it came in the form of a flood, fire, earthquake, hurricane, or tornado. On the other hand, when conditions get too cold, hot, dry, wet, or otherwise inhospitable, and this is believed to represent a fundamental long-term shift that makes rebuilding futile, then people have tended to get themselves and their property out of harm's way. The rational difference in behavior comes from whether the climate is believed to have 'stationarity', i.e., if the probability of certain extreme events is not changing. Models suggest that where the historical data give confidence on the 20th-century probability distribution for certain kinds of extreme events, their 21st-century distribution will shift; some events, like hurricanes, are difficult to draw conclusions for due to a combination of poor records and model uncertainties.
For large modern cities, with their huge immobile investments in real estate and infrastructure, not to mention their sentimental and symbolic importance, what should the ratio be between mitigation and avoidance? Total mitigation would call for constructing dikes and flood gates in places like New Orleans and Miami, or new desalination plants and water pipelines into Arizona, no matter how large the cost or how small the gain; total avoidance would dictate size and geographic limits on populations in such places. Either requires a very long-term approach, one that is not typically taken for political reasons. Further, there is the dilemma of how much discounting should occur when facing outcomes that are not guaranteed, and that will take place at an imprecisely known future date.
This is the gateway to many interesting questions about impacts which, clearly, have yet to be resolved with any satisfaction. Climate-modeling studies can help provide answers in two ways: 1. how is the probability of certain extreme events changing within certain urban areas going into the future? and 2. what role do the physical layouts and structures of those areas play in affecting those probabilities?
A Primer on Urban Climate
Temperatures of surfaces on a sunny day in Arizona. Source: ASU via LBL Heat Island Group.
Aerial view of São Paulo. Source: Skyscraper City.
An infrared gun reveals the complexity of the temperature structure in a still-life anthropogenic environment (left). Much of the work of urban climatology is focused on determining how these micro-variations interact in a network with thousands such scenes, animated by moving vehicles and people and intermingled with buildings of varying heights and materials (as at right) — with the whole ecosystem subject to the mercurial dictates of the overlying weather pattern.
Among the ways in which the presence of cities influences the climates of surrounding regions, the urban-heat-island [UHI] effect is probably the most familiar. It is easy to observe and experience, and is the main environmental influence on urban morbidity (sickness), energy usage, and comfort. The basic conceptual picture: because of the low albedo and high heat capacity of streets and buildings, as well as their impermeability to water, incoming radiation builds up heat during the day which is then radiated through the night. The build-up is accentuated by buildings that physically block winds (limiting the dispersal of heat by air movement) and reduce the "sky-view factor", thus limiting how much heat can escape radiatively into the night sky (the same effect that contributes to forests remaining warmer at night than open fields). As a consequence, the UHI effect is typically strongest in the evening. In dense cities on clear nights, these factors can combine to produce especially stark contrasts, as in the below chart (left) from the Hong Kong Observatory, where green data are from a rural area 35 km away but at a similar elevation. At right is a figure illustrating how cities, in interaction with their surrounding environments, develop distinctive urban boundary layers [UBLs, a.k.a. UHIs] within which air is mixed due to turbulence, with properties strongly influenced by the surface underneath. The urban canopy layer [UCL] exists amid the obstructions where winds cannot easily mix. By direct shading, albedo increase, and reduction of sensible (i.e. measurable with a thermometer) heating, not to mention pollution absorption, vegetation is an extremely effective counter to the UHI; one study from the 1990s estimated that in the arid summertime climate of southern California, filling in street trees wherever possible would save $80 million [2015 dollars] per year in cooling costs alone, due to a roughly 25% reduction in air-conditioning usage.
Urban boundary and canopy layers in an idealized cross-section. Source: The British Geographer
Observed precipitation in the area surrounding Houston, TX (within central oval), showing enhancement within and downwind of the urban area. Source: http://dx.doi.org/10.1175/1087-3562(2003)007<0001:DOUIRA>2.0.CO;2
To produce precipitation, air really needs only two qualities: to be rising andsaturated. Generally, the faster the rise or the moister the air, the higher the precipitation amounts that are possible. There are two main theories regarding the effect of urban areas on extreme precipitation, and they have opposing conclusions: the first posits that because cities produce more pollution, the higher number of aerosols allows water to condense onto them and form small raindrops, rather than grow into larger ones big enough to fall out of the cloud. [It also means that urban areas tend to be cloudier, especially if smog is taken into account.] According to the second, the UHI leads to more-buoyant air that rises higher and releases more energy as latent heat upon condensation, forming stronger storms. These storms are also driven by the fact that urban evapotranspiration tends to occur faster because most moisture remains on or near the surface — instead of penetrating deep into the soil as it does in rural areas — and also that air is often physically forced upwards when it encounters tall buildings and/or the UBL. Modeling studies indicate that both effects are at play, but that the overall result is an 'island' characterized by an increase in heavy precipitation in and downwind of urban areas (also see figure at right), accompanied by a modest reduction in light-precipitation events.
Among the multitude of other variables of interest, wind and pollution levels are two of the most prominent. Winds have long been known to be lower inside urban areas relative to outside. However, it often does not seem that way when standing on the street, because of the Venturi effect — a fancy name for the fact that flows move faster through constricted passages, like between tall buildings. Of course, winds are also affected by precipitation and especially temperature variations, on as small a scale as from one side of a street to the other. In street canyons, winds can essentially be described as the interaction of channeling along the street and a weaker vortex perpendicular to it, roughly the height of the canyon, that enables air to move between streets along rooftops. This dynamic means that, overall, within-canyon winds are usually constrained to be parallel to the street orientation.
As anthropogenic phenomena, the various urban 'islands' beat to the rhythm of human life, giving them clear temporal patterns; for instance, because pollution varies over the course of a week,precipitation in the southeast U.S. has a midweek peak as well. Each urban area has its own idiosyncratic distribution of emissions sources, of building heights, of flow-constricting mountains, and of breeze-generating lakes and oceans, which all combine to influence temperature and moisture across space and time.This means that regional effects may vary considerably from the typical global picture. In desert climates, for example, cities are often cooler than the surrounding (unvegetated) area because of street trees that regulate temperatures by shading and evaporative cooling. In Arctic and sub-Arctic regions, extreme heat and hurricanes may not be a concern, but melting sea ice and permafrost are. Where buildings are tall, the UCL limits mixing, the UHI effect tends to be stronger, and pollution is higher; where they are not, conditions depend on (among other things) whether the low-profile land use is leafy neighborhoods or low-albedo asphalt.
Extremes in Urban Areas and their Impacts
Historical U.S. temperature-mortality relationship, showing the number of deaths at various temperatures relative to the number at 20 deg C. Dashes indicate confidence intervals at the 95th percentile. Source: doi:10.1038/nclimate1902.
Air conditioners, of course, have made the average Western resident less sensitive to the vicissitudes of outdoor temperatures, simply by reducing the time that they are exposed to untreated air. Even within countries, access varies significantly by socioeconomic status: although the average Bostonian does not truly 'experience' heat waves as measured in a recent study using portable thermometers, those who are older and poorer often do. Mortality among the elderly (75+) increases about 25% in the strongest heat waves. On the other hand, a series of studies in the mid-20th century showed that uncomfortable temperatures have a measurable impact on productivity — in fact, when factories first became air-conditioned, summer worker productivity improved about 25%. Needless to say, indoor heating also improves quality of life a lot, though as nearly the entire population has had access to it for as long as such records have been kept, there is presumably little that can be done to reduce the exposure that still occurs.
But these benefits come at a cost both monetary and energetic. In 2004, the U.S. devoted approximately 3.29 quadrillion BTU's of energy to air conditioning, or roughly 8% of the total national usage. At the same time, sub-Saharan Africa combined (excepting South Africa) consumed 3.04 quadrillion BTU's total — for a population more than twice as large. We know that in the interconnected physical environment, no action can occur without an accompanying reaction (with the complicating details of feedback effects, spatial movements, and time delays). Without any consideration of the effects of aerosols and greenhouse gases, this affects summertime weather in areas where population density is high — the figure below illustrates that the waste heat from A/C's is itself substantial enough to raise urban temperatures by several deg F. To be fair, urban areas accrue many economic and social benefits to their residents, and also environmental benefits due to the energy efficiencies of public transit and large building size. At this point in time, comparing climatic and socioeconomic factors is a task too complex to have been tackled with any success; in the future, perhaps, such complete analyses will become possible.
A/C waste heat and the UHI in Paris: "real" temperatures are observations, while "dry A/C" scenarios assume all waste heat is sensible. Source: doi:10.1016/j.apenergy.2012.02.015
Separately from the effects of waste heat, recent papers have found with fairly high confidence that UHI effects are magnified during heat waves. While UHI magnitude in the Baltimore area is not nearly as large as in the nearly-ideal Hong Kong example above (dry conditions, clear skies, a light north wind), it is on the order of 1 deg C (1.8 deg F), increasing to 2 or 3 deg C during the studied four-day heat wave. For that event, with temperatures around 35 deg C, the temperature-mortality relationship curve predicts a mortality increase of about 15% outside the UHI vs 20% inside it-- a difference amounting to an expected 2.75 more deaths in Baltimore city. As with a lot of climate-related figures, that may not sound like much; but once you begin multiplying by population and integrating over time — not to mention tallying up all the non-fatal impacts -- the numbers get large very fast, with this effect alone likely causing thousands of extra deaths per year worldwide.
Extreme precipitation is figuratively not as flashy as temperature, but literally it can be. In most regions of the world, where climate change is expected to increase available moisture, urban precipitation is consequently generally expected to become more extreme. But as alluded to in the previous section, the story is more complex than for temperature. Observations show the precipitation distribution is shifting toward fewer but more extreme events across the U.S., whereas in a global sampling of urban areas, precipitation extremes were not seen to be increasing (though they are for temperature). Many researchers expect precipitation variability to increase over the 21st century, though so far this has not been widely observed. Sea-level rise, on the other hand, is indisputable; and with about 2/3 of the world's largest cities partially within 30 feet of sea level, but only about 10% of the total population, the problem takes on a particularly urban aspect. The situation in Bangladesh prefigures an impending challenge for parts of the world that are better-prepared to handle it — though as discussed above, the cost (and associated temporal) point at which prudence becomes folly, and mitigation abandoned in favor of avoidance, is an unsettled matter that, because of its importance, should be subject to fierce but well-informed debate.