RESEARCH AND ANALYSIS The Changing Metabolism of Cities · The Changing Metabolism of Cities Christopher Kennedy, John Cuddihy, and Joshua Engel-Yan Summary Data from urban metabolism

RESEARCH AND ANALYSIS

The Changing Metabolismof CitiesChristopher Kennedy, John Cuddihy, and Joshua Engel-Yan

Summary

Data from urban metabolism studies from eight metropoli-tan regions across five continents, conducted in various yearssince 1965, are assembled in consistent units and compared.Together with studies of water, materials, energy, and nutrientflows from additional cities, the comparison provides insightsinto the changing metabolism of cities. Most cities studied ex-hibit increasing per capita metabolism with respect to water,wastewater, energy, and materials, although one city showedincreasing efficiency for energy and water over the 1990s.Changes in solid waste streams and air pollutant emissions aremixed.

The review also identifies metabolic processes thatthreaten the sustainability of cities. These include alteredground water levels, exhaustion of local materials, accumu-lation of toxic materials, summer heat islands, and irregular ac-cumulation of nutrients. Beyond concerns over the sheer mag-nitudes of resource flows into cities, an understanding of theseaccumulation or storage processes in the urban metabolism iscritical. Growth, which is inherently part of metabolism, causeschanges in water stored in urban aquifers, materials in thebuilding stock, heat stored in the urban canopy layer, and po-tentially useful nutrients in urban waste dumps.

Practical reasons exist for understanding urban metabolism.The vitality of cities depends on spatial relationships withsurrounding hinterlands and global resource webs. Increas-ing metabolism implies greater loss of farmland, forests, andspecies diversity; plus more traffic and more pollution. Urbanpolicy makers should consider to what extent their nearest re-sources are close to exhaustion and, if necessary, appropriatestrategies to slow exploitation. It is apparent from this reviewthat metabolism data have been established for only a fewcities worldwide, and interpretation issues exist due to lack ofcommon conventions. Further urban metabolism studies arerequired.

Keywords

global citiesindustrial ecologymaterials flow analysis (MFA)sustainable citiesurban environmenturban metabolism

e-supplement available on the JIEWeb site

Address correspondence to:Prof. Christopher KennedyDepartment of Civil EngineeringUniversity of Toronto35 St. George StreetToronto, Ontario M5S 1A4, Canada<[email protected]><www.civil.engineering.utoronto.ca>

© 2007 by the Massachusetts Institute ofTechnology and Yale University

Volume 11, Number 2

www.mitpressjournals.org/jie Journal of Industrial Ecology 43

R E S E A R C H A N D A N A LYS I S

Introduction

Cities grow in complex ways due to theirsheer size, social structures, economic systems,and geopolitical settings, and the evolution oftechnology (Hall 1998). Responding to waves ofnew technology, cities have grown outward fromsmall dense cores, first as linear transit cities andthen as sprawling automobile cities. Changes inindustry have also been important; obsolete fac-tories have often closed down, leaving behindcontaminated soil and groundwater. Risks asso-ciated with rebuilding on such brownfield siteshave encouraged developers to pursue greenfieldsites on the edges of cities. Ever-growing urbanpopulations also fuel the expansion of cities. Yeteven cities showing no change or decreasing pop-ulations, such as some older industrial cities, arestill growing outward.

Forty years ago, in the wake of rapid urbanexpansion, Abel Wolman published a pioneer-ing article on the metabolism of cities. Wolman(1965) developed the urban metabolism conceptin response to deteriorating air and water qualityin American cities—issues still recognized todayas threatening sustainable urban development.Wolman analyzed the metabolism of a hypothet-ical American city, quantifying the overall fluxesof energy, water, materials, and wastes into andout of an urban region of 1 million people.

The metabolism of an ecosystem has been de-fined by ecologists as the production (via pho-tosynthesis) and consumption (by respiration) oforganic matter; it is typically expressed in termsof energy (Odum 1971). Although a few studieshave focused on quantifying the embodied energyin cities (Zucchetto 1975; Huang 1998), otherurban metabolism studies have more broadly in-cluded fluxes of nutrients and materials and theurban hydrologic cycle (Baccini and Brunner1991). In this broader context, urban metabolismmight be defined as the sum total of the technical andsocioeconomic processes that occur in cities, resultingin growth, production of energy, and elimination ofwaste.

Since Wolman’s work, a handful of urbanmetabolism studies have been conducted in urbanregions around the globe. By reviewing thesestudies, this article describes how the urbanmetabolism of cities is changing. It also demon-

strates how understanding of accumulation pro-cesses in the urban metabolism is essential to thesustainable development of cities.

Sustainable development can be understoodas development without increases in the through-put of materials and energy beyond the bio-sphere’s capacity for regeneration and waste as-similation (Goodland and Daly 1996). Given thisdefinition, a sustainable city implies an urban re-gion for which the inflows of materials and en-ergy and the disposal of wastes do not exceed thecapacity of its hinterlands. As discussed in thisarticle, this definition presents difficulties in thecontext of cities dependent on global markets;nevertheless, it provides a relative bar againstwhich progress may be measured. In quantify-ing material and energy fluxes, urban metabolismstudies are valuable for assessing the direction ofa city’s development.

The objectives of this article are twofold.The first objective is to review previously pub-lished metabolism studies to elucidate what weknow about how urban metabolism is changing.A previous review article on energy and mate-rial flows to cities has been presented by Deckerand colleagues (2000), but it made only minorreference to the metabolism concept; it did notinclude reference to many of the studies con-sidered here; and it did not specifically ask howurban metabolism is changing. The number ofmetabolism studies is not sufficient to apply anyform of statistical analysis, and therefore, somemight argue, to identify any generalizable trends.Nevertheless, the balance of evidence generallypoints to increasing urban metabolism.

The second objective is to identify critical pro-cesses in the urban metabolism that threaten thesustainable development of cities. It has been ar-gued that high levels of urban resource consump-tion and waste production are not issues aboutsustaining cities per se, but reflect concerns overthe role of cities in global sustainable develop-ment (Satterthwaite 1997). Although partiallyagreeing, we aim to show that there are also pro-cesses within the urban metabolism that threatenurban sustainability itself. In particular we high-light storage processes—water in urban aquifers,heat stored in urban canopy layers, toxic mate-rials in the building stock, and nutrients withinurban waste dumps—all of which require careful

44 Journal of Industrial Ecology


long-term management. Understanding thechanges to such storage terms in some cities maybe as important as reducing the sheer magnitudesof inputs and outputs.

The cities studied are metropolitan regions,that is, similar to standard metropolitan statisti-cal areas (SMSAs) in the United States, whichoften encompass several politically defined cities,that is, cities under the jurisdiction of a local mu-nicipal government. These metropolitan regionscan generally be regarded as commutersheds, al-though the basis for definition may differ betweencountries.

The metabolism of cities will be analyzed interms of four fundamental flows or cycles—thoseof water, materials, energy, and nutrients. Dif-ferences in the cycles may be expected betweencities due to age, stage of development (i.e., avail-able technologies), and cultural factors. Otherdifferences, particularly in energy flows, may beassociated with climate or with urban populationdensity (table 1). Each of the two objectives willbe addressed sequentially within a discussion ofthe four cycles.

Previous Metabolism Studies

Insights into the changing metabolism ofcities can be gained from a few studies fromaround the world, conducted over severaldecades. These studies are typically of greatermetropolitan areas, including rural or agriculturalfringes around urban centers. One of the earli-est and most comprehensive studies was that ofBrussels, Belgium by the ecologists Duvigneaudand Denaeyer-De Smet (1977), which includedquantification of urban biomass and even organicdischarges from cats and dogs (figure 1)! In study-ing the city of Hong Kong, Newcombe and col-leagues (1978) were able to determine inflowsand outflows of construction materials and fin-ished goods. An update of the Hong Kong studyby Warren-Rhodes and Koenig (2001) showedthat per capita food, water, and materials con-sumption had increased by 20%, 40%, and 149%,respectively—from 1971 to 1997. Upward trendsin per capita resource inputs and waste outputswere also reported in a study of Sydney, Australia(Newman 1999). Studying a North Americancity, Sahely and colleagues (2003) found that al-

though most inputs to Toronto’s metabolism wereconstant or increasing, some per capita outputs,notably residential solid waste, had decreasedbetween 1987 and 1999. Metabolism studies ofother cities include those for Tokyo (Hanya andAmbe 1976), Vienna (Hendriks et al. 2000),Greater London (Chartered Institute of WastesManagement 2002), Cape Town (Gasson 2002),and part of the Swiss Lowlands (Baccini 1997).Some studies have quantified urban metabolismless comprehensively than others. Bohle (1994)considers the urban metabolism perspective onurban food systems in developing countries. Afew studies of nutrient flows in urban systems havebeen undertaken (Nilson 1995; Bjorklund et al.1999, Baker et al. 2001), whereas others havespecifically investigated metals or plastics in theurban metabolism. Collectively these metabolismstudies provide a quantitative appraisal of the dif-ferent ways that cities are changing worldwide.

Related to the urban metabolism concept isthe application of the ecological footprint tech-nique to cities (Wackernagel and Rees 1995).The ecological footprint of a city is the amountof biologically productive area required to provideits natural resources and to assimilate its wastes.Published studies include those for Vancouver(Wackernagel and Rees 1995), Santiago deChile (Wackernagel 1998), Cardiff (Collinset al. 2006), and cities of the Baltic region ofEurope (Folke et al. 1997). The equivalent areasof ecosystems for sustaining cities are typicallyone or two orders of magnitude greater than theareas of the cities themselves.

Water

In terms of sheer mass, water is by farthe largest component of urban metabolism.Wolman’s calculations for the 1960s for a one-million-person U.S. city estimated the input ofwater at 625,000 tonnes1 per day compared tojust 9,500 tonnes of fuel and 2,000 tonnes of food.Most of this inflow is discharged as wastewater,with the remainder being lost by activities suchas the watering of lawns. Data from the citiesin table 1 show that wastewater represents be-tween 75% and 100% of the mass of water inflow(figure 2a). The six studies since 1990 typicallyhave higher per capita water inputs than the four

Kennedy, Cuddihy, and Engel-Yan, The Changing Metabolism of Cities 45


Tabl

e1

Cha

ract

erist

ics

ofci

ties

(met

ropo

litan

area

s)w

here

urba

nm

etab

olism

stud

ies

have

been

cond

ucte

d

Av.

tem

pera

ture

(◦ C)

Popu

latio

nU

rban

ized

Loca

tion

City

Ref

eren

ceYe

ar(m

illio

n)de

nsity

(cap

/km

2 )(g

eogr

aphi

cco

ordi

nate

s)A

ltitu

de(m

)Su

mm

erW

inte

r

U.S

.typ

ical

Wol

man

1965

1965

1.0

——

——

—

Bru

ssel

sD

uvig

neau

dan

d19

70s

1.07

56,

640

50◦

N76

16.4

4.2

Den

aeye

r-D

eSm

et19

774◦

ET

okyo

Han

yaan

dA

mbe

1976

1970

11.5

136,

632

35◦

N17

24.3

6.9

139◦

EH

ong

Kon

gN

ewco

mbe

and

colle

ague

s197

819

713.

940

22◦

N88

28.0

17.0

War

ren-

Rho

desa

ndK

oeni

g20

0119

977.

058

,000

114◦

ESy

dney

New

man

1999

1970

2.79

33◦

S0

22.3

1319

903.

657

2,00

015

1◦E

Tor

onto

Sahe

lyan

dco

lleag

ues2

003

1987

4.03

843

◦N

105

19.9

−1.8

1999

5.07

12,

920

79◦

WV

ienn

aH

endr

iksa

ndco

lleag

ues2

000

1990

s1.

53,

710

48◦

N17

018

.51.

816

◦E

Lond

onC

hart

ered

Inst

itut

eof

2000

7.4

6,73

051

◦N

1416

.74.

9W

aste

sMan

agem

ent2

002

Whi

te20

0320

000◦

WC

ape

Tow

nG

asso

n20

0220

003.

03,

900

34◦

S6

20.8

13.5

18◦

E

Not

e:ca

p=

capi

ta.O

nesq

uare

kilo

met

er(k

m2 ,S

I)=

100

hect

ares

(ha)

≈0.

386

squa

rem

iles≈

247

acre

s.O

nem

eter

(m,S

I)≈

3.28

feet

(ft)

.(C

elsi

uste

mpe

ratu

re[◦

C]∗

9/5)

+32

=Fa

hren

heit

tem

pera

ture

[◦F]

.



Figure 1 The urban metabolism of Brussels, Belgium in the early 1970s. Source: Duvigneaud andDenaeyer-De Smet 1977.

from the early 1970s, although Tokyo in 1970 andLondon and Cape Town in 2000 are significantexceptions. In the case of Toronto, per capitawater use slightly declined over the 1990s, duelargely to a reduction in industrial consumption.Nevertheless, none of the cities has quite reachedthe level of water use of the average U.S. cityreported by Wolman.

The impacts of water on the sustainability ofcities have further dimensions, beyond the cru-cial need for inhabitants to have a safe, reliablewater supply. For those that rely on ground wa-ter, the long-term relationship between cities andtheir ground water environments is illustrative.As cities evolve from small settlements they typi-cally go through several stages of developmentin which the relationship with an underlyingaquifer changes (Foster et al. 1998). In the earlysettlement stage, water supply is obtained fromshallow urban wells and boreholes, and waste-water and drainage waters are discharged to theground or to a watercourse (figure 3a). As thesettlement grows into a small city the underly-

ing water table falls with increased extraction,so deeper wells are drilled. Overexploitation ofground water often occurs. Moreover, as a resultof urban activities, often including continued dis-charge of wastewater to the ground, the shallowground water in the city center becomes polluted.Subsidence may begin to occur, depending onsoil characteristics. The paving of land surfacesand the growth of drainage systems also begin tohave a discernible impact on the ground watersystem (figure 3b). As the city expands furtherand matures, a turnaround can sometimes occur.With widespread contamination of the aquiferbelow the city, or a movement of heavy, water-extracting industry away from city centers, pump-ing of the urban aquifer ceases. The city now re-lies on periurban well fields or expensive waterimports from distant sources. With the cessationof pumping, the water table below the city beginsto rise. Because of changes in the surface recharge,the water table may rise above that under virginconditions, potentially causing flooding or dam-age to infrastructure (figure 3c). In this evolving



0

50

100

150

200

250

av.

US

. ci

ty (

1965

)

Bru

ssel

s (1

970s

)

Syd

ney

(197

0)

Tok

yo (

1970

)

Hon

g K

ong

(197

1)

Tor

onto

(19

87)

Syd

ney

(199

0)

Vie

nna

(199

0s)

Hon

g K

ong

(199

7)

Tor

onto

(19

99)

Lond

on (

2000

)

Cap

e T

own

(200

0)

t/ca

p/y

r

Water Supply

Wastewater

(a)

(b)

00.5

11.5

22.5

33.5

44.5

5

av.

US

. ci

ty (

1965

)

Bru

ssel

s (1

970s

)

Syd

ney

(197

0)

Tok

yo (

1970

)

Hon

g K

ong

(197

1)

Tor

onto

(19

87)

Syd

ney

(199

0)

Vie

nna

(199

0s)

Hon

g K

ong

(199

7)

Tor

onto

(19

99)

Lond

on (

2000

)

Cap

e T

own

(200

0)

t/ca

p/y

r

Residential

Total

Figure 2 Metabolism parameters from selected cities. (a) Fresh water inputs and wastewater releases.(b) Solid waste disposal. (c) Energy inputs. (d) Contaminant emissions. Missing bars indicate data unavailable.Date refers to the date of measurement; t refers to tonnes. For sources see table 1. The data behind thisfigure are available in an electronic supplement (e-supplement) on the JIE Web site. One tonne (t) =103 kilograms (kg, SI) ≈ 1.102 short tons. One gigajoule (GJ) = 109 joules (J, SI) ≈ 2.39 × 105 kilocalories(kcal) ≈ 9.48 × 105 British Thermal Units (BTU). SO2 = sulfur dioxide; NOx = nitrogen oxides; VOC =volatile organic compounds.

process cities can go from exploiting ground waterresources to potentially being flooded by them.

Problems of overexploiting ground waterhave occurred in many cities. One example isBeijing, China, where the water table dropped by45 m between 1950 and 1990 (Chang 1998). In

coastal cities, overpumping has caused salt waterintrusion, threatening ground water supplies; ex-amples include Gothenburg, Perth, Manila, andJakarta (Volker and Henry 1988). Decades oflowering of water tables in Mexico City havecaused land subsidence of 7.5 m in the center of



020406080

100120140160

av.

U.S

. ci

ty (

1965

)

Bru

ssel

s (1

970s

)

Syd

ney

(197

0)

Tok

yo (

1970

)

Hon

g K

ong

(197

1)

Tor

onto

(19

87)

Syd

ney

(199

0)

Hon

g K

ong

(199

7)

Tor

onto

(19

99)

Lond

on (

2000

)

Cap

e T

own

(200

0)

GJ/

cap

/yr

Primary

Direct

(c)

(d)

00.010.020.030.040.050.060.07

av.

US

. ci

ty (

1965

)

Bru

ssel

s (1

970s

)

Syd

ney

(197

0)

Hon

g K

ong

(197

1)

Syd

ney

(199

0)

Lond

on (

1995

)

Hon

g K

ong

(199

7)

Tor

onto

(19

95)

Cap

e T

own

(200

0)

t/ca

p/y

r SO2

NOx

VOC

Particulates

Figure 2 Continued

the city, altering surface drainage and damaginginfrastructure (Ortega-Guerrero et al. 1993).

Concerns over the potential impact of risingwater tables applies to at least one of the studycities. Because industrial withdrawals decreasedin London during the 1960s, the water table in anunderlying chalk aquifer has been rising at a rateof 1 to 2.5 m per year under central London (Cox1994; Castro and Swyngendow 2000). Moreover,leakage from London’s water distribution systems,estimated to make up 28% of the total watersupply, may be adding to this rise in the watertable. Where the water table will reach equilib-rium is an open question, especially given the

change in surface features and local climate overthe past decades.

Rising water tables have also been recorded forseveral cities in the Middle East: Riyadh, Jeddah,Damman, Kuwait, Al-Ajman, Beirut, Cairo, andKarachi (Abu-Rizaiza 1999). In many cases thecause is subterranean discharge of wastewaterflows, where there is no surface water outlet. Suchrising of ground water levels threatens urban in-frastructure, including basements, foundations,tunnels, and other subsurface pipes and cables.Resulting cracks in columns, beams, and wallswere reported in Riyadh, Saudi Arabia. In manyrespects the physical integrity of cities depends



Water Table

(a) Early settlement

Water Supply from Peri-urban Wellfields

Water Supply from Peri-urban Wellfields

Imports of Water from Distant Surface or Ground Water Sources

Water Table

(b) Town grows into a city

(c) Mature city

Figure 3 Evolution of water supply and wastewater disposal for a typical city underlain by a shallowaquifer. The arrows indicate water flows; the line with dots indicates the water table. Source: Adapted fromFigure 1.2 of Foster et al. 1998.

on achieving equilibrium in the ground watercomponent of the urban metabolism.

Materials

Material inputs to cities are generally lesswell quantified than water inputs, despite theirsignificance to urban infrastructure. Detailedanalyses, however, have been conducted forHong Kong and Vienna. Material flow analyseshave also been conducted for Hamburg and a few

other European cities (Hammer et al. 2003). In1997, Hong Kong’s daily material inputs included363 tonnes of glass; 3,390 tonnes of plastics;9,822 tonnes of cement; 2,095 tonnes of wood;7,240 tonnes of iron and steel; and 2,768 tonnesof paper. Relative to 1971, inputs of plastics hadrisen by 400%; iron and steel increased by closeto 300%; cement and paper inputs were both upby 275%; and wood inputs had more than dou-bled. In comparison, the population only rose by78%. The historical trend in the construction



material input to Vienna is also upward. In mostdecades prior to the 1970s, the per capita use ofconstruction materials was 0.1 m3/yr (the 1930sbeing a significant exception at 1.4 m3/cap/yr).2

The volumes of input in the 1970s, 1980s, and1990s, however, were 2.4, 3.3, and 4.3 m3/cap/yr,respectively (Brunner and Rechberger 2004).Upward per capita trends, for the period 1992and 2001, were also estimated for direct materialinput and domestic material consumption for theCity of Hamburg (Hammer et al. 2003).. This ev-idence suggests that cities are becoming increas-ingly material-intensive.

Measures of solid waste outputs are availablefor most of the cities in table 1, although thesedata need careful interpretation. Where citieshave introduced significant recycling schemes,the disposal of residential household waste maybe declining in per capita terms. For example,in Toronto, per capita residential waste declinedby 27% between 1987 and 1999. If other wastestreams such as the commercial and industrialare included, however, the overall trend maywell be upward; three of the cities studied in the1990s have total waste outputs greater than 1.5t/cap/yr, whereas all cities studied in the 1970s,except Tokyo, had outputs below 1 t/cap/yr (fig-ure 2b). Some caution has to be taken in in-terpreting material outputs. Construction wasteis often the largest solid waste component (e.g.,57% for Tokyo), but because much of it is con-sidered inert and can be recycled or used as fill,it may not appear in calculations of total wasteoutputs.

In Vienna, where some of the most extensivematerial flow studies have been conducted, theproduction of “holes” from mining of aggregatesby far exceeds the city’s output of waste mate-rials. Construction materials, of course, go intothe building stock and typically remain withinthe urban fabric for many decades. In Viennathe stock of materials in the buildings and in-frastructure is estimated to be 350 t/cap. Thisstock is clearly growing: the input of constructionmaterials and consumer products is on the orderof 12 to 18 t/cap/yr, whereas solid waste is only3 t/cap/yr. Taking a slightly different perspective,it is evident that the production of holes dueto the excavation of construction materials by farexceeds the rate at which they are backfilled. The

cumulative hole in the vicinity of Vienna due toexcavation of materials since 1880 is estimatedto be over 200 million m3, and growing rapidly(Brunner and Rechberger 2004).

A dimension of material flows that impactsthe sustainability of a city is the distance overwhich materials are transported. As cities growand transportation infrastructure develops, rawmaterials seemingly travel increasingly long dis-tances into cities. During the first half of thetwentieth century, mineral aggregates used bythe construction industry in Toronto came fromquarries within a few kilometers of the city. By1970, the aggregates were traveling an average of160 km from various counties around the Torontoregion (figure 4), the former sources having beenexhausted or consumed within the urban area.Douglas and colleagues (2002) describe similareffects during the growth of Manchester, Eng-land, since the industrial revolution. In some re-spects the city is like a plant stretching its rootsout further and further until its resource needsare met—a concept that parallels the ecologicalfootprint. One aspect of this growth is that citiesrequire greater expenditure of transportation en-ergy as materials travel from increasing distances.

Energy

The quantification of urban energy fluxes asa component of urban metabolism has variedin depth and breadth between studies. One ofthe most comprehensive was that of Brussels, forwhich both natural and anthropogenic energysources were quantified. Most other studies havefocused directly on anthropogenic sources, ne-glecting net all-wave radiation and heat transferdue to evapotranspiration, local advection, soilconduction, and mass water transfer. The impor-tance of urban heat islands, discussed below, how-ever, indicates the significance of incorporatinganthropogenic sources into the natural surfaceenergy balance of cities.

Comparisons of the anthropogenic energy in-puts for the cities of table 1 need to distin-guish between the direct energy consumed andthe primary energy consumption, which includesenergy losses in the production of electricity(figure 2c). With its cold winters, and fairly warmsummers, Toronto has the highest per capita



Figure 4 The distribution of mineral aggregate sources for Metropolitan Toronto in 1972. MetropolitanToronto, in the center of the figure, is the destination of all flow shown. Source: Ontario Ministry of NaturalResources.

energy use of the cities. London and Sydney usesignificantly more energy per capita than HongKong, potentially due to its warm winters andhigher urban density. A further interesting con-trast is that between London in 2000 and Brusselsin the early 1970s. Both cities lie at approxi-mately the same latitude and have similar cli-mates and almost identical population densities,yet per capita consumption in modern-day Lon-don is an order of magnitude higher. Upwardtrends in energy consumption since the 1970sare also evident for Sydney and Hong Kong.A minor reduction has been experienced inToronto since 1987, perhaps due to changing in-dustrial demands and increasing efficiency. Withthe exception of Toronto, none of the other citieshave reached the energy consumption levels ofWolman’s average U.S. city.

The relationship between transportation en-ergy demand and urban form has been widelystudied, but with various conclusions. Trans-

portation accounts for a significant proportionof urban anthropogenic energy. Among the mostsignificant and highly debated findings are thoseof Newman and Kenworthy (1991), which illus-trate per capita transportation energy consump-tion decreasing as population density increases(figure 5). All of the cities from table 1 exceptCape Town were included in the study, whichused data from the 1980s. Hong Kong is theprime example of a dense city with low trans-portation energy; the European trio Brussels, Lon-don, and Vienna are less energy-intensive thanthe newer cities of Sydney and Toronto. Al-though the premise that urban form influencestransportation energy has been criticized, mostnotably by Gordon and Richardson (1989), stud-ies have demonstrated that although populationdensity may not in itself contribute to explainingtransportation demand, distance from the cen-tral business district and other employment cen-ters does (Miller and Ibrahim 1998). As such,



Figure 5 Variation in annual transportation energy consumption and population density among severalglobal cities during the 1980s. ’000 MJ = one thousand megajoules = one gigajoule (GJ) = 109 joules(J, SI) ≈ 2.39 × 105 kilocalories (kcal) ≈ 9.48 × 105 British Thermal Units (BTU). One hectare (ha) =0.01 square kilometers (km2 , SI) ≈ 0.00386 square miles ≈ 2.47 acres. Source: Newman and Kenworthy1991. Reprinted with permission.

the hypothesis linking transportation demand,energy consumption, and urban spatial structureis valid. The higher per capita energy consump-tion reflected in the urban metabolism of NorthAmerican cities may thus be explained at least inpart by their expansive urban form.

A further aspect of the urban energy bal-ance influencing sustainability is the urban heatisland. Most mid- and high-latitude cities ex-hibit high urban air temperatures relative totheir rural surroundings, particularly during theevening (Taha 1997). On a calm clear night af-ter a sunny windless day, elevated temperaturesat the canopy layer can be as much as 10◦C dif-ferent in large cities. This urban heat island hasbeen identified as a consistent phenomenon byurban climatologists (Landsberg 1981; Oke et al.1991; Oke 1995). Much of the heat island ef-fect is due to the built form distorting the naturalenergy balance—rooftops and paved surfaces ab-sorb heat and reduce evaporation, whereas streetcanyons and other microscale effects can act asheat traps. An interesting issue, however, is therelative importance of anthropogenic energy in-

puts in elevating temperatures. All of the en-ergy that is pumped into cities will eventuallyturn into heat. Estimated anthropogenic con-tributions range from 16 W/m2 for St. Louis to159 W/m2 for Manhattan (Taha 1997).3 Santa-mouris (2001) summarizes a collection of studies,some of which present contradictory findings, orat least highlight the importance of city-specificfeatures.

Increases in temperature directly impact sum-mer cooling loads, thus introducing a potentiallycyclic effect on energy demand. For U.S. citieswith populations greater than 100,000, peak elec-tricity loads increase by about 1% for everydegree Celsius increase in temperature (Santa-mouris 2001). For high ambient temperatures,utility loads in Los Angeles have demonstrated anet rate of increase of 167 megawatts (MW)4 per◦C. In Toronto, a 1◦C increase on summer dayscorresponds to roughly a 1.6% increase in peakelectricity demand. Hot summer days are criticalin Toronto in that they cause the highest electric-ity demand throughout the year. Such demandsare often met through increased coal-generated



power, elevating emissions of greenhouse gasesand other air pollutants. Thus to the extent thatanthropogenic energy contributes to the summer-time heat island, there is a small positive feedbackloop raising both temperature and contaminantemissions.

As air contaminants are largely associatedwith energy production or utilization, it is perhapsappropriate to analyze them with the energy cy-cle. Unlike air quality, which can be directly mea-sured, contaminant emissions can be difficult toquantify due to their wide range of sources withinan urban region. With the exception of some un-published values for London in 1995, per capitasulfur dioxide (SO2) and nitrogen oxides (NOx )emissions in the group of cities (appearing infigure 2d) are all lower than for Wolman’s U.S.city. The data from Hong Kong and Sydneyshow that although SO2 emissions are decreasing,NOx has increased since the 1970s. Emissionsof volatile organic compounds (VOCs) are fairlysimilar for Toronto in 1997, Sydney in 1990,Brussels in the early 1970s, and the average U.S.city in 1965; the highest values seen were thosefor Sydney in 1970, whereas Hong Kong has hadconsistently low emissions of VOCs. Particulateemissions reported in the past two decades aresignificantly lower than the 0.05 t/cap for the av-erage 1965 U.S. city.

As urban air quality is a concern, it may bemore appropriate to express emission intensitiesin kg per unit area, rather than kg per capita, as infigure 2d. (The data behind figure 2 are includedin an e-supplement for interested readers.5) Notethat cities can also be subject to external sourcesof air pollution.

Greenhouse gases are a further form of air pol-lutant of concern for global sustainability. Car-bon dioxide (CO2) emissions from the group ofcities correspond quite closely with energy in-puts. Emissions from Toronto are estimated tobe around 14 t/cap/yr, followed by Sydney atclose to 9 t/cap/yr. Yet there is still quite ahigh level of uncertainty in urban greenhouse gasemissions; for example, CO2 emissions reportedfor London in the late 1990s range from 5.5 to8.5 t/cap/yr. Overall, the emissions of contami-nants remain quite closely tied to anthropogenicenergy emissions; but this link could weaken ascities learn to exploit renewable energy sources,

for example, using photovoltaics, solar waterheating, or energy recovery from wastewater.

Nutrients

Understanding the flow of nutrients throughthe urban system is vital to successful nutri-ent management strategies and urban sustainabil-ity. Consequences of improper management in-clude eutrophication of water bodies, release ofheavy metals onto agricultural lands, acid rain,and groundwater pollution (Nilson 1995; Bakeret al. 2001). The Hong Kong metabolism study(Warren-Rhodes and Koenig 2001) included ananalysis of key nutrients: nitrogen and phospho-rus. A nitrogen balance for the Central Arizona–Phoenix (CAP) ecosystem (Baker et al. 2001), aphosphorus budget for the Swedish municipalityof Gavle (Nilson 1995), and a study of Bangkok(Faerge et al. 2001) also provide insight into theflow of nutrients through urban systems.

The studies reveal the extent to which naturalnutrient flows are altered in human-dominatedecosystems. In the CAP and Gavle regions, ap-proximately 90% of nitrogen and phosphorus in-puts are human-mediated. Whereas the major-ity of phosphorus fluxes are related to humanfood production, import, and consumption (agri-cultural production and food import), nitrogenfluxes in urban systems are becoming increas-ingly linked to combustion processes (Bjorklundet al. 1999, Baker et al. 2001). In the CAP region(which includes agricultural land and desert inaddition to urban Phoenix), fixation from NOx

emissions from combustion is the single largestnitrogen input, greater than nitrogen inputs fromcommercial fertilizers applied to both crops andlandscapes! In Hong Kong, 42% of the nitrogenoutput is NOx from combustion, which is equiv-alent to nitrogen outputs from wastewater. Thus,reductions in nitrogen inputs and outputs mightbe most readily achieved by reducing NOx emis-sions (Baker et al. 2001).

In addition to managing nutrient inputs andoutputs of the urban system, nutrient storagemust also be considered. Nutrients may remainin the urban system through accumulation in soilor groundwater by inadvertent losses or directdisposal or through nutrient recycling, such asthe use of food waste for agricultural fertilizer.



Accumulation often results in negative environ-mental consequences, such as groundwater pol-lution. All study areas exhibit high levels of nu-trient storage. Approximately 60% of all nitro-gen and phosphorus inputs to Hong Kong, 60%of phosphorus inputs to the Gavle municipality,20% of nitrogen inputs to the CAP region, and51% of phosphorus (but only 3% of nitrogen) inBangkok do not leave the system. Although asmall amount of these nutrients are recycled, themajority are accumulated in municipal and indus-trial waste sites, agricultural soils, and groundwa-ter pools (Nilson 1995; Baker et al. 2001; Warren-Rhodes and Koenig 2001).

The relatively low levels of nutrient recyclingpracticed in these study areas highlight the lackof synergy that exists between urban centers andtheir hinterlands. Girardet (1992) suggests thatfor cities to be sustainable from a nutrient per-spective, they must practice fertility exchange, inwhich the nutrients in urban sewage are returnedto local farmland. This relationship between thecity and its hinterlands requires adequate sewagetreatment and an appropriate means of sewagetransportation, as well as a good supply of localagriculture. The process of urban–rural nutrientrecycling was prevalent in the United States dur-ing the nineteenth century, until the develop-ment of the modern fertilizer industry (Wines1985). Similar processes existed more recently inChina, where 14 of the country’s 15 largest citieswere largely self-sufficient in food, supplying amajority of their food requirements from agri-cultural suburbs, which were kept fertile usingtreated human waste (Girardet 1992); such prac-tices have since changed. In contrast, Hong Kongproduces only 5% of its food needs locally, andhuman food wastes, which used to be recycled asbone meal fertilizers, are now disposed of in land-fills (Warren-Rhodes and Koenig 2001). Giventhat the Hong Kong model is more representa-tive of most modern cities than the former Chi-nese case, what can be done to improve urbansustainability from a nutrient perspective?

It is not clear that the fertility exchange be-tween a city and its hinterlands described byGirardet (1992) is an appropriate goal for themodern city. Traditionally, there was a symbi-otic relationship between cities and their sur-rounding rural areas, but this relationship has

been significantly weakened by modern changesin transportation technology and access to globalmarkets. As cities grow and sewage treatmentbecomes more centralized, the costs of transport-ing sewage sludge to local agricultural lands be-come more prohibitive. Moreover, pharmaceuti-cals and other toxics present in wastewater maymake sludge recycling hazardous to health with-out expensive treatment. These challenges maymake other sewage resource recovery alterna-tives, such as energy and water reclamation, moreattractive for implementation.

The extent to which a city should be depen-dent on its hinterlands to maintain a sustainablefood supply is a difficult question. Many citiesobtain their food from continental or global net-works of suppliers. For example, 81% of London’s6.9 million tonnes of annual food is importedfrom outside the United Kingdom (CharteredInstitute of Wastes Management 2002). Clearly,cities have grown away from dependence on thesurrounding landscape. The relationship of mu-tual dependence between the city and its hin-terlands described by von Thunen’s (1966 [origi-nally published 1826]) “isolated state” no longerholds. Peet (1969) shows how the average dis-tance of British agricultural imports increasedfrom 1,820 miles (from London) in 1831–1835to 5,880 miles by 1909–1913; he argues that acontinental-scale von Thunen-type agriculturalsystem has operated since the nineteenth cen-tury. Being part of a continental food networkmay have economic advantages—and the net-work as a whole may be more resilient than a sin-gle city. But the sustainability of entire continen-tal food webs is itself in question. In the UnitedStates for example, agriculture is heavily relianton fossil fuels; 7.3 units of energy are consumedto produce one unit of food energy; depletionof topsoil exceeds regeneration; and groundwaterwithdrawals exceed recharge in major agriculturalregions (Heller and Keolian 2003). Further re-search is needed to determine whether more lo-cally grown food production, either within urbanareas or periurban surroundings, offers a more sus-tainable agricultural system.

Even though cities have grown away from de-pendence on the surrounding landscape, theycannot function in isolation from their natu-ral environment. The most effective nutrient



management strategies are those that are specif-ically tailored to individual ecosystems (Bakeret al. 2001). The creation of such strategies re-quires an understanding of nutrient input, ac-cumulation, and output, which can be acquiredfrom detailed nutrient balances within an urbanmetabolism study.

Conclusions

With data from metabolism studies in eightcities, spread over five continents and severaldecades, observation of strong trends would notbe expected from this review. Moreover, thereare concerns over the commensurability of datafrom different cities, especially for waste streams.Nevertheless, with three different cities hav-ing been analyzed at different times, and othersupporting evidence, some sense of how themetabolism of cities is changing can be gleaned.Many of the data suggest that the metabolism ofcities is increasing: water and wastewater flowswere typically greater for studies in the 1990sthan those in the early 1970s; Hamburg, HongKong, and Vienna have become more materials-intensive; and energy inputs to Hong Kong andSydney have increased. However, some signspoint to increasing efficiency, for example, percapita energy and water inputs leveling off inToronto over the 1990s. Other changes aremixed; for example, cities that have implementedlarge-scale recycling have seen reductions in res-idential waste disposal in absolute terms, butother waste streams—such as commercial andindustrial—may well be on the increase. Simi-larly, emissions of SO2 and particulates may havedecreased in several cities, whereas other air pol-lutants such as NOx have increased. Changesin urban metabolism are quite varied betweencities.

Future studies might attempt to identify dif-ferent classes of urban metabolism. These areperhaps apparent from the transportation energydata, where old European, dense Asian, and NewWorld cities are distinctive. Climate likely hasan impact on the type of metabolism; cities withinterior continental climates would be expectedto expend more energy on winter heating andsummer cooling. The cost of energy may also in-fluence consumption. The age of a city, or its

stage of development, could be a further factor inthe type of urban metabolism.

Beyond concerns over the sheer magnitudesof resource flows into cities, there are more subtleimbalances and feedbacks that threaten sustain-able urban development. Access to large quan-tities of fresh water is clearly vital to sustain-ing a city, but the difference between the inputfrom and output to surface waters may be as im-portant as the sheer volume of supply. Where acity imports large volumes of water, but releaseswater into underlying aquifers, whether throughleaking water pipes, septic tanks, or other means,changing water tables may threaten the integrityof urban infrastructure. Whether it is water inan urban aquifer, construction materials used asfill, heat stored in rooftops and pavements, or nu-trients accumulated in soils or waste sites, theseaccumulation processes should be understood sothat resources can be used appropriately. Furtherexamination of how resources are used and storedwithin a city may yield some surprises. For ex-ample, Brunner and Rechberger (2001) suggestthat “Modern cities are material hot spots con-taining more hazardous materials than most haz-ardous landfills. . . .” Moreover, growth, which isinherently part of a city’s metabolism, implies achange in storage. Thus, in addition to quantify-ing inputs and outputs, future studies should aimto characterize the storage processes in the urbanmetabolism.

Beyond scientific interest in the nature of citygrowth, there are practical reasons for studyingurban metabolism. The implications of increasingmetabolism, at least with current predominanttechnologies, are greater loss of farmland, forests,and species diversity, plus more traffic and morepollution. In short, cities will have even largerecological footprints.

The vitality of cities depends on spatial re-lationships with surrounding hinterlands andglobal resource webs. As cities have grown andtransportation technologies have changed, re-sources have traveled greater distances to reachcities. For heavier materials, which are more ex-pensive to transport, the exhaustion of the near-est, most accessible resources may at some pointbecome a constraint on the growth of cities.For many goods, including food, modern citiesno longer rely on their hinterlands; rather, they



participate in continental and global trading net-works. Thus, full evaluation of urban sustainabil-ity requires a broad scope of analysis.

Urban policy makers should be encouraged tounderstand the urban metabolism of their cities.It is practical for them to know if they are usingwater, energy, materials, and nutrients efficiently,and how this efficiency compares to that of othercities. They must consider to what extent theirnearest resources are close to exhaustion and,if necessary, appropriate strategies to slow ex-ploitation. It is apparent from this review thatmetabolism data have been established for onlya few cities worldwide and there are interpreta-tion issues due to lack of common conventions;there is much more work to be done. Resourceaccounting and management are typically under-taken at national levels, but such practices mayarguably be too broad and miss understanding ofthe urban driving processes.

Acknowledgment

This research was supported by the NaturalSciences and Engineering Research Council ofCanada.

Notes

1. One tonne (t) = 103 kilograms (kg, SI) ≈ 1.102short tons.

2. One cubic meter (m3, SI) ≈ 1.31 cubic yards (yd3).3. One watt (W, SI) ≈ 3.412 British Thermal Units

(BTU)/hour ≈ 1.341 × 10−3 horsepower (HP).One square meter (m2, SI) ≈ 1.20 square yards(yd2).

4. One megawatt (MW) = 106 watts (W, SI) = 1megajoule/second (MJ/s) ≈ 56.91 × 103 BritishThermal Units (BTU)/minute.

5. Available at the JIE Web site.

References

Abu-Rizaiza, Q. S. 1999. Threats from ground watertable rise in urban areas in developing countries.Water International 24(1): 46–52.

Baccini, P. 1997. A city’s metabolism: Towards thesustainable development of urban systems. Journalof Urban Technology 4(2): 27–39.

Baccini, P. and P. H. Brunner. 1991. Metabolism of theanthroposphere. Berlin: Springer-Verlag.

Baker, L. A., D. Hope, Y. Xu, J. Edmonds, and L.Lauver. 2001. Nitrogen balance for the CentralArizona–Phoenix (CAP) ecosystem. Ecosystems4: 582–602.

Bjorklund, A., C. Bjuggren, M. Dalemo, and U. Sones-son. 1999. Planning biodegradable waste manage-ment in Stockholm. Journal of Industrial Ecology3(4): 43–58.

Bohle, H. G. 1994. Metropolitan food systems indeveloping countries: The perspective of urbanmetabolism. GeoJournal 34(3): 245–251.

Brunner, P. H. and H. Rechberger. 2001. Anthro-pogenic metabolism and environmental legacies.In Encyclopedia of Global Environmental Change,Vol. 3, edited by T. Munn. Chichester, UK:Wiley.

Brunner, P. H. and H. Rechberger. 2004. Practical hand-book of material flow analysis. Boca Raton, FL: CRCPress.

Castro, E. and E. Swyngendow. 2000. “Case study: Lon-don,” project Metropolitan Cities and Sustain-able Use of Water (METRON), Environment andClimate Programme, Framework IV—programproject, DGXII. Brussels: European Commission.

Chang, S. D. 1998. Beijing: Perspectives on preser-vation, environment, and development. Cities15(1): 13–25.

Chartered Institute of Wastes Management. 2002. Aresource flow and ecological footprint analysis ofGreater London. London: Best Foot Forward.

Collins, A., A. Flynn, T. Weidmann, and J. Barrett.2006. The environmental impacts of consump-tion at a subnational level: The ecological foot-print of Cardiff. Journal of Industrial Ecology 10(3):9–24.

Cox, D. W. 1994. The effects of changing ground waterlevels on construction in the City of London. InGround water problems in urban areas, edited byW. B. Wilkinson. London: Thomas Telford.

Decker, H., S. Elliott, F. A. Smith, D. R. Blake, andF. Sherwood Rowland. 2000. Energy and materialflow through the urban ecosystem. Annual Reviewof Energy and the Environment 25: 685–740.

Douglas, I., R. Hodgson, and N. Lawson. 2002. Indus-try, environment and health through 200 years inManchester. Ecological Economics 41(2): 235–255.

Duvigneaud, P. and S. Denaeyer-De Smet. 1977.L’ecosysteme urbain bruxellois. [The Brussels urbanecosystem.] In Productivite en Belgique, edited byP. Duvigneaud and P. Kestemont. Traveaux de laSection Belge du Programme Biologique Interna-tional, Brussels. Paris: Edition Duculot.



Faerge, J., J. Magid, and F. W. T. Penning de Vries.2001. Urban nutrient balance for Bangkok. Eco-logical Modelling 139: 63–74.

Folke, C., A. Jansson, J. Larsson, and R. Costanza. 1997.Ecosystem appropriation by cities. Ambio 26(3):167–172.

Foster, S. S., A. R. Lawrence, and B. L. Morris.1998. Ground water in urban development. WorldBank Technical Paper no. 390. Washington, DC:World Bank.

Gasson, B. 2002. The ecological footprint of CapeTown: Unsustainable resource use and plan-ning implications. Paper presented at SAPI In-ternational Conference: Planning Africa, 17–20September, Durban, South Africa.

Girardet, H. 1992. The Gaia atlas of cities. London: GaiaBooks Limited.

Goodland, R. and H. Daly. 1996. Environmental sus-tainability: Universal and non-negotiable. Ecolog-ical Applications 6: 1002–1017.

Gordon, P. and H. W. Richardson. 1989. Gasoline con-sumption and cities: A reply. Journal of the Amer-ican Planning Association 55(3): 342–345.

Hall, P. 1998. Cities in civilization. New York: Pantheon.Hammer, M., S. Giljum, and F. Hinterberger. 2003.

Material flow analysis of the City of Hamburg.Paper presented at the workshop Quo vadis MFA?Material Flow Analysis—Where Do We Go? Is-sues, Trends and Perspectives of Research for Sus-tainable Resource Use, 9–10 October, Wuppertal.

Hanya, T. and Y. Ambe. 1976. A study on themetabolism of cities. In Science for a better environ-ment. Tokyo: HESC, Science Council of Japan.

Heller, M. C. and G. A. Keolian. 2003. Assessing thesustainability of the US food system: A life cycleperspective. Agricultural Systems 76: 1007–1041.

Hendriks, C., R. Obernosterer, D. Muller, S. Kytzia,P. Baccini, and P. Brunner. 2000. Material flowanalysis: A tool to support environmental pol-icy decision making. Case studies on the city ofVienna and the Swiss lowlands. Local Environment5: 311–328.

Huang, S. 1998. Urban ecosystems, energetic hi-erarchies, and ecological economics of Taipeimetropolis. Journal of Environmental Management52: 39–51.

Landsberg, H. E. 1981. The urban climate. New York:Academic Press.

Miller, E. J. and A. Ibrahim. 1998. Urban form andvehicular travel: Some empirical findings. Trans-portation Research Record 1617: 18–27.

Newcombe, K., J. D. Kalina, and A. R. Aston. 1978.The metabolism of a city: The case of Hong Kong.Ambio 7: 3–15.

Newman, P. W. G. 1999. Sustainability and cities:Extending the metabolism model. Landscape andUrban Planning 44: 219–226.

Newman, P. and J. Kenworthy. 1991. Cities and au-tomobile dependence: An international source book.Aldershot, UK: Avebury.

Nilson, J. 1995. A phosphorus budget for a Swedish mu-nicipality. Journal of Environmental Management45: 243–253.

Odum, E. P. 1971. Fundamentals of ecology. Philadel-phia: Saunders.

Oke, T. R. 1995. The heat island of the urban bound-ary layer: Characteristics, causes and effects. InWind climate in cities, edited by Jack E. Cermack.Dordrecht, the Netherlands: Kluwer Academic.

Oke, T. R., G. T. Johnson, D. G. Steyn, and I. D.Watson. 1991. Simulation of surface heat islandsunder ‘ideal’ conditions at night, Part 2: Diagno-sis of causation. Boundary-Layer Meteorology 56:339–358.

Ortega-Guerrero, A., J. A. Cherry, and D. L. Rudolph.1993. Large-scale aquitard consolidation nearMexico City. Ground Water 31(5): 708–718.

Peet, R. 1969. The spatial expansion of commercialagriculture in the nineteenth century: A vonThunen interpretation. Economic Geography 45:283–301.

Sahely, H. R., S. Dudding, and C. A. Kennedy. 2003.Estimating the urban metabolism of Canadiancities: GTA case study. Canadian Journal for CivilEngineering 30: 468–483.

Santamouris, M., ed. 2001. Energy and climate in theurban built environment. Athens: James & James.

Satterthwaite, D. 1997. Sustainable cities or citiesthat contribute to sustainable development.Urban Studies 34(10): 1667–1691.

Taha, H. 1997. Urban climates and heat islands:Albedo, evapotranspiration, and anthropogenicheat. Energy and Buildings 25: 99.

Volker, A. and J. C. Henry, eds. 1988. Side effects re-lated to groundwater management. In Side EffectsOf Water Resources Management. Institute of Hy-drology. Wallingford, UK: IAH Press. 123–204.

Von Thunen, J. H. 1966. The isolated state. 1826. Trans-lated by Carla M. Wartenberg; edited with an in-troduction by Peter Hall. Oxford, UK: Pergamon.

Wackernagel, M. 1998. The ecological footprint ofSantiago de Chile. Local Environment 3(1): 7–25.

Wackernagel, M. and W. E. Rees. 1995. Our ecolog-ical footprint: Reducing human impact on Earth.Philadelphia: New Society.

Warren-Rhodes, K. and A. Koenig. 2001. Escalatingtrends in the urban metabolism of Hong Kong:1971–1997. Ambio 30(7): 429–438.



White, R. 2003. Building the ecological city. Cambridge,UK: Woodhead.

Wines, R. 1985. Fertilizer in America: From waste recy-cling to resource exploitation. Philadelphia: TempleUniversity Press.

Wolman, A. 1965. The metabolism of cities. ScientificAmerican 213(3): 179–190.

Zucchetto, J. 1975. Energy, economic theory and math-ematical models for combining the systems of man

and nature. Case study, the urban region of Mi-ami. Ecological Modelling 1: 241–268.

About the Authors

Christopher Kennedy is an associate professor andJohn Cuddihy and Joshua Engel-Yan are former grad-uate students in the Department of Civil Engineeringat the University of Toronto in Toronto, Canada.


Documents

RESEARCH AND ANALYSIS The Changing Metabolism of Cities · The Changing Metabolism of Cities Christopher Kennedy, John Cuddihy, and Joshua Engel-Yan Summary Data from urban metabolism