Scholarly article on topic 'Feeding and housing the urban population: Environmental impacts at the peri-urban interface under different land-use scenarios'

Feeding and housing the urban population: Environmental impacts at the peri-urban interface under different land-use scenarios Academic research paper on "Social and economic geography"

CC BY
0
0
Share paper
Academic journal
Land Use Policy
OECD Field of science
Keywords
{"Life cycle assessment" / "Environmental impact" / "Multifunctional land use" / "Sustainable peri-urban development" / Urbanisation / Horticulture}

Abstract of research paper on Social and economic geography, author of scientific article — Alison Rothwell, Brad Ridoutt, Girija Page, William Bellotti

Abstract The environmental consequences of the decision to urbanise and displace peri-urban (PU) food production are not typically evaluated within a comprehensive, cross-sectoral approach. Using a novel application of life cycle assessment (LCA) within exploratory scenarios, a method for integrating housing and food production land uses in PU regions is proposed, based on relative environmental impacts. Using two housing types (greenfield and infill) and two types of food production (field and high-technology greenhouse (HTG) lettuce production), environmental impacts for five exploratory land-use scenarios are compared for PU land in a developed and growing city. Each scenario is able to house an equivalent residential population whilst delivering equal quantities of fresh food to a city market. The results clearly indicate that infill housing and food production has less environmental impact than greenfield development. The environmental impact categories of climate change, freshwater eutrophication, photochemical oxidant formation, particulate matter formation and human toxicity are reduced by 25–43 percent under infill scenarios. Sparing PU land through infill housing development combined with sustainable food intensification using HTG production, enabled multifunctional PU land-use including food production, housing and afforestation while delivering lower relative environmental impacts. Urban afforestation on PU land made available by these measures reduces the effect of climate change by up to 5 percent per hectare per year.

Academic research paper on topic "Feeding and housing the urban population: Environmental impacts at the peri-urban interface under different land-use scenarios"

Contents lists available at ScienceDirect

Land Use Policy

journal homepage: www.elsevier.com/locate/landusepol

Land Use Policy

Feeding and housing the urban population: Environmental impacts at the peri-urban interface under different land-use scenarios

Alison Rothwell3 *, Brad Ridouttb, Girija Page3, William Bellottia

a School of Science and Health, University of Western Sydney, Locked Bag 1797, Penrith, NSW2751, Australia b Commonwealth Scientific and Industrial Research Organisation (CSIRO), Private Bag 10, Clayton South, Victoria 3169, Australia

CrossMark

ARTICLE INFO

Article history:

Received 18 September 2014 Received in revised form 28 May 2015 Accepted 20 June 2015

Keywords:

Life cycle assessment Environmental impact Multifunctional land use Sustainable peri-urban development Urbanisation Horticulture

ABSTRACT

The environmental consequences of the decision to urbanise and displace peri-urban (PU) food production are not typically evaluated within a comprehensive, cross-sectoral approach. Using a novel application of life cycle assessment (LCA) within exploratory scenarios, a method for integrating housing and food production land uses in PU regions is proposed, based on relative environmental impacts. Using two housing types (greenfield and infill) and two types of food production (field and high-technology greenhouse (HTG) lettuce production), environmental impacts for five exploratory land-use scenarios are compared for PU land in a developed and growing city. Each scenario is able to house an equivalent residential population whilst delivering equal quantities of fresh food to a city market. The results clearly indicate that infill housing and food production has less environmental impact than greenfield development. The environmental impact categories of climate change, freshwater eutrophication, photochemical oxidant formation, particulate matter formation and human toxicity are reduced by 25-43 percent under infill scenarios. Sparing PU land through infill housing development combined with sustainable food intensification using HTG production, enabled multifunctional PU land-use including food production, housing and afforestation while delivering lower relative environmental impacts. Urban afforestation on PU land made available by these measures reduces the effect of climate change by up to 5 percent per hectare per year.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Urban areas have a substantial environmental impact in relation to their small absolute land area. Anthropogenic emissions from urban areas range upwards of 30 percent of the global total. Building-related greenhouse gas (GHG) emissions are predicted to rise a further 50-150 percent by the middle of the 21st century as urban populations expand (IPCC, 2014). The expansion of urban areas and the consequent direct changes in land use may contribute to environmental burdens in other sectors, including the agriculture sector — for example, when housing is allowed to extend onto peri-urban (PU) cropland, displacing food production to more remote locations (Low Choy and Buxton, 2013). Since the agriculture sector contributes a further 14 percent of total global anthropogenic emissions (United Nations Human Settlements Programme, 2011), the combined environmental impact of urban expansion replacing PU farms may

* Corresponding author. E-mail address: alison.rothwell@uws.edu.au (A. Rothwell).

http://dx.doi.org/10.1016/j.landusepol.2015.06.017 0264-8377/© 2015 Elsevier Ltd. All rights reserved.

be significant. Typically, however, these consequences are not evaluated comprehensively using a cross-sectoral approach. Few environmental studies have attempted to analyse the continued provision of fresh perishable food following displacement of PU agriculture.

Furthermore, the potential trade-offs between different environmental impacts that may occur have seldom been analysed. For example, reductions in GHG emissions may coincide with increases in other environmental effects associated with urban systems: increased particulate matter (Brochu et al., 2011); declining water quality (Tong and Chen, 2002); increased ozone concentration (Sicard et al., 2013); human toxicity; and water scarcity. Inhaled particulate matter may contribute to human ill health; declining water quality affects urban ecosystems. Increased ozone levels may adversely affect certain materials as well as human and plant health. Human toxicity results from the persistence of harmful chemicals leading to toxicity in the environment and in the food chain. Various chemicals are known to affect cognitive development, with research indicating that children may be more at risk in urban areas than in rural regions (Calderon-Garciduenas et al., 2008; Liu and Lewis, 2014). A growing awareness of the need

for efficient use of water resources in urban areas is reflected in measures to secure water supplies and water-scarcity pricing is beginning to appear on government agendas (Frontier Economics, 2011). Comparisons within a suite of environmental impacts permit trade-offs to be examined, together with a clearer appreciation of regional relevance.

The present study describes an approach for integrating housing and food-production land usage in PU regions, based on their relative environmental impacts. Using combinations of two housing types with two types of lettuce production, the environmental impacts of five exploratory PU land-use scenarios were analysed using a life cycle assessment (LCA) approach. Primary data was applied wherever possible. All scenarios accommodate equal residential populations and deliver equal quantities of fresh food to a city market. The possibility of further PU land use, such as afforestation, was also explored. By incorporating both housing and food system changes, the cross-sectoral approach adopted in the study has produced a novel and comprehensive assessment of the environmental consequences of urban expansion displacing PU food production.

1.1. Peri-urban landscapes and food provision

PU regions are transitional zones between urban and rural districts. Land in PU regions may be used for a multitude of purposes — housing, recreation, ecosystem preservation, commercial food production and other primary industries. Spatially, there is no consistent worldwide definition of how far from a metropolitan centre a PU region might extend. For example, they may range from tens to more than a hundred kilometres from a major centre. Population densities are defined in some contexts, as in the European PLUREL project (Piorr et al., 2011) but, more typically, densities remain undefined due to inter-country variation, with the focus continuing to be on the competition for resources in PU regions (FAO, 1999). PU regions are often highly contested. Conflict is manifested at the level of governance: land-use planning is typically based on an urban-rural dichotomy that exhibits poor integration of the demands of each. This may extend to the level of residents, with newer residents often possessing lifestyles that are opposed to production values. In Australia, as in other developed nations, PU agricultural regions have changed rapidly, historically succumbing to urban development and sprawl (Piorr et al., 2011; Millar and Roots, 2012; Low Choy and Buxton, 2013). It is paradoxical that the commercial fresh food production capacity to support an increasing urban population is therefore lost (Martin et al., 2008). In a developed and growing city such as Sydney, for example, expected population trajectories to 2061 will entail providing food for an additional 200 persons each day (ABS, 2013)1.

The opportunity to improve environmental outcomes for PU development by incorporating food provision into planning is often not recognised internationally (APA, 2007; Lovell, 2010; Zasada, 2011; van der Schans and Wiskerke, 2012; Pires and Burton, 2013; Russo et al., 2014). Conventional commercial PU agriculture is often missing from discussions about urban agriculture and its place in urban planning. Metropolitan strategies in Australia, for example, have yet to proactively consider the continuing provision of fresh food from surrounding agricultural land. This is in contrast to the progress evident in cities such as Chicago, London and Vancouver (Budge, 2013). Such inattention to commercial PU agriculture in urban planning is evident despite its significant contribution to local (and regional) markets. For example, in the USA, urban-influenced regions have been reported to produce most of the

fruit and vegetables, at 91 and 78 percent respectively (American Farmland Trust, 2013). In Sydney, PU vegetable production was recently valued at 27 percent of the value of production for the state of New South Wales. The contribution of PU agriculture is larger for specific crops, such as for lettuce at 53 percent (ABS, 2014)2. Commercial PU agriculture performs a more vital role than other forms of urban agriculture (e.g. community gardens and rooftop gardens) whose output quantity is typically dramatically lower in developed cities in the USA, for example (Brown, 2002). Direct-to-consumer urban agriculture in the San Francisco foodshed, such as community-supported agriculture, has been reported to contribute only 0.75 percent of overall production value (Thompson et al., 2008). Similarly, recent figures indicate that intra-urban vegetable production in the Sydney metropolitan region contributes only 0.6 percent of the value of PU vegetable production (ABS, 2014).

Planning approaches that consider only the immediate impact of urban development, yet ignore the displacement of commercial food producers, remain incomplete and potentially flawed. Comprehensive environmental assessment of any decision to urbanise PU horticultural lands requires that more than the urban system alone be taken into account. The consequences of such decisions cannot be fully understood unless the system is expanded to include the impact both of housing and of horticultural systems. Such an awareness must include consideration of food production displacement if the necessary food production to supply local retail markets is to continue. Informed decisions about long-term, sustainable urban development are possible only if the decision-makers have access to environmental data to complement economic and social considerations. (Economic and social factors are outside the scope of this study.)

1.2. Land-use integration in peri-urban settlements

If the interrelation of urban and agricultural systems is recognised, the opportunity then arises for PU regions to support environmental impact mitigation by integrating different land uses. Policy instruments that encourage such strategies have been recommended for minimising the lock-in risks associated with urban land use and infrastructure life spans (IPCC, 2014). Since PU regions are already typified by a wide range of land uses, it would appear reasonable to assess their ability to incorporate several functions, rather than limit them merely to the single function of conversion to greenfield housing. Such functions may include productive, cultural and ecological land use, which form three dimensions of landscape multifunctionality (Lovell and Taylor, 2013). Multifunctionality in urban landscapes is emerging as a necessity, since an obvious consequence of urbanisation is loss of agricultural and natural land. Benefits associated with effective multifunctional landscape planning include support and regulating services (e.g. carbon sequestration and afforestation) that flow on to increased biodiversity and reduced urban heat-island effects; however, despite the increasing application of the multifunctional landscape concept to agro-ecosystems, there are few examples in urban ecosystem planning (Lovell and Taylor, 2013).

Multifunctional combinations of land uses in PU regions will generate different environmental impacts to monofunctional housing use. The extent of any difference between such environmental impacts is not clear. Few studies of the environmental trade-offs for alternative PU land-use scenarios have been reported, in which housing, food production and co-benefits such as afforestation may be integrated. To fill this gap, several of the scenarios compared in the present study comprise the elements of housing, fresh food

1 Series B medium growth data for Greater Sydney.

2 Data representative of the Hawkesbury-Nepean region, a PU area of Greater Sydney.

and afforestation. These three PU land uses of housing, fresh food and afforestation contain three elements of landscape multifunc-tionality. Housing provides productive and cultural use, fresh food provision is a productive use and afforestation represents an ecological use. As urbanisation associated with population growth drives ever-greater land-use change in the PU fringe, it becomes increasingly urgent that we examine how to integrate options that will provide fresh food, housing and ecological benefits but with a lower impact on the environment.

1.3. Integrating environmental assessment

The lack of integrative planning for contested PU landscapes justifies moves to generate knowledge about alternative ways of incorporating food production into urban regions (Lovell, 2010; McDonald, 2014; Nilsson et al., 2014). It has been noted elsewhere that there is a trend toward cross-sectoral, scenario- and systems-based approaches to sustainable land management and adaptive planning (Thabrew et al., 2009; O'Farrell and Anderson, 2010; Scholz et al., 2012; Hussey et al., 2013). Huang et al. (2010) has suggested that, together with a cross-sectoral approach, bottom-up perspectives are important in urban planning for sustainability. LCA has been put forward as a systems-based, bottom-up decision support tool for assessing the environmental impacts of planned multifunctional urban landscapes (Lovell and Taylor, 2013). In sus-tainability science, Sala et al. (2013) emphasised the importance of life-cycle thinking when attempting to integrate transdisciplinary knowledge. LCA supports the environmental aspect of sustainabil-ity, since it provides evidence-based comparisons between the environmental impacts of particular systems.

LCA is a globally recognised methodology (ISO, 2006a,b) that is frequently used in both the building and food sectors. Approaches to environmental impact assessments in housing have traditionally been confined to studies of building materials or building life cycles (material production, construction, operation, maintenance, demolition and disposal) (Haapio and Viitaniemi, 2008). Horticultural and food environmental impact assessments also use LCA, where in this case the assessment may be restricted to cradle-to-farm-gate, or may include life-cycle impacts for the system post-farm (e.g. packaging, transport or consumer use and waste, depending on the study purpose). However, the environmental assessment of buildings is usually carried out in isolation, without consideration of land-use changes or other consequences of urbanisation. Similarly, horticultural environmental impact studies are typically conducted independent of alternative land-use considerations.

In the present study, LCA was used in hypothetical scenarios to integrate data from these traditionally disparate fields of housing and horticulture. Scenario development has been proposed as a way of informing future urban growth patterns with sustainable multifunctional landscapes and transitions (Wiek et al., 2006; Beardsley et al., 2009; O'Farrell and Anderson, 2010). The inclusion both of housing changes and the consequent food production changes is a novel application of LCA, whose capacity to examine environmental impacts other than emissions was exploited. It was appreciated that studies that compare alternative land-use should also consider potential environmental trade-offs.

In summary, in this paper a method of assessing the environmental impacts associated with the decision to urbanise horticultural land was presented. Scenarios combining two kinds of lettuce production and two types of housing were compared for one hectare (1 ha) of PU land in a developed and expanding city. All scenarios specified that equal quantities of fresh food and the housing of equal residential populations were to be provided. Multifunctional land-use options explored within several scenarios included those providing co-benefits such as urban afforestation. LCA was

used to generate a detailed inventory for each scenario as data for quantitative environmental impact analysis, which considered:

• The housing system, including construction, operation and transport.

• The food system and the associated change in the environmental impact as a result of displacement of PU food production.

• Alternative food production technologies that spare PU land by using high-productivity methods, enabling co-benefits such as afforestation.

• The environmental impact trade-offs, by investigating a range of impact indicators important to both PU and urban systems.

2. Methods

In all scenarios, 1 ha of PU land was required to house equal populations and produce the same quantity of fresh food. The food was required to be delivered to the city's central fruit and vegetable market (Sydney Markets®, Homebush, NSW). (The subsequent food supply chain, including consumption, was outside the scope of this study.)

A housing system and a food production system were assessed in each scenario. The housing system was represented by the two contrasting housing types of greenfield and infill. Greenfield housing development is a known cause of PU land-use change, due to its taking possession of agricultural land in expanding cities. Such developments are often termed greenfield sprawl'. Conversely, in urban planning the compact city is the generally accepted model for environmental sustainability (Neuman, 2005; Gleeson, 2012; Westerink et al., 2012). From an environmental impact perspective, compact growth usually produces better environmental outcomes (Beardsley et al., 2009; Shearer et al., 2009; Bierwagen et al., 2010; Schetke et al., 2012; Thorne et al., 2013). Compact city strategies promote urban consolidation, such as the development of apartments in existing urban centres (Randolph and Tice, 2013). Consequently, the two disparate housing types were compared: new greenfield houses in PU regions, and medium-density infill apartments in existing suburban centres. Infill apartments were required to house the same residential population as a greenfield development on 1 ha of PU land.

Within the food production system, field production and a high-technology greenhouse (HTG) were modelled for lettuce. Lettuce was selected as the crop because of the quantity produced and its year-round production in the PU region, its perishability and because it is a food staple (Hall et al., 2014). The two growing technologies yield very different quantities per unit area of land. The higher yield of modern lettuce production technologies, including a HTG, spares land from cultivation for other purposes. Afforestation was the chosen alternative PU land-use option in the study.

Existing PU land was considered to be horticultural. The geographical scope was represented by Sydney, Australia, where outer urban greenfield housing is characterised by urban sprawl, although a trend toward increased infill housing development has been reported (Rowley and Phibbs, 2012). Similar urban sprawl occurs in many developed and developing nations (Richardson and Bae, 2004; Siedentop and Fina, 2012), accompanied by the absorption of PU farmland (Boume et al., 2003; Imhoff et al., 2004; Simon, 2008).

Scenario narratives are described in Section 2.1. Following scenario development, LCA was used to generate detailed inventory and parameters in support of the narratives, in order for quantitative environmental impact analysis to occur (Section 2.2).

Table 1

Peri-urban (PU) land-use scenario narratives.

Scenario

Narrative

Housing system Horticultural system

Afforestation 1_F

Housing system

Horticultural system Afforestation 1_HTG_A Housing system

Horticultural system Afforestation 70I:30G_F Housing system

Horticultural system

Afforestation 70I:30G_HTG_A Housing system

Horticultural system Afforestation

1 ha of PU horticultural land is transformed to greenfield housing (detached single dwellings)

1 ha of PU field lettuce production is displaced by greenfield housing to a more remote field production location. The remote field farm is required to produce and transport the same quantity of produce to the city's central fruit and vegetable market as the displaced PU farms Nil

The same population as that housed in scenario G.F is housed in infill, inner suburban apartment-type housing. There is no change to the 1 ha of PU horticultural land as a result of housing

PU field lettuce production is maintained on 1 ha of existing PU horticultural land Nil

The same population as that housed in scenario G_F is housed in infill, inner suburban apartment-type housing. There is no change to the 1 ha of PU horticultural land as a result of housing PU field lettuce production has been replaced by a HTG

The reduced land area requirements of a higher-output HTG spares PU land for afforestation

An area ratio of 70% infill, inner suburban apartment-type housing to 30% detached greenfield single dwellings. Thus, 1 ha of existing PU horticultural land is converted to 0.3 ha of greenfield housing. There is no other change to the 1 ha of PU horticultural land. The balance of housing is infill in inner suburban regions

Field lettuce production is retained on the remaining 0.7 ha of existing PU horticultural land. The balance of lettuce production is

delivered to the city's central fruit and vegetable market from a more remote field production location

An area ratio of 70% infill, inner suburban apartment-type housing to 30% detached greenfield single dwellings. Thus, 1 ha of existing PU

horticultural land is converted to 0.3 ha of greenfield housing. There is no other change to the 1 ha of PU horticultural land. The balance of

housing is infill in inner suburban regions

PU field lettuce production has been replaced by a HTG

Any remaining PU land is used for afforestation

2.1. Land-use scenarios

Each of the five land-use scenarios for 1 ha of PU land incorporates a housing system and a horticultural system; several also include afforestation. The requirement to ensure that equal residential populations were housed and equal food quantities produced was central to the scenario narratives described in Table 1. The corresponding breakdown of the area of PU land required for housing, food production and afforestation in each scenario is given in Table 2. The different compositions of the land area was fundamental to the purpose of the study. One hectare of PU land (the functional unit of the study) was the benchmark for the life-cycle inventory (LCI) within the LCA.

Scenarios combining greenfield housing with field lettuce production involve either a complete (G_F) or partial (70I:30G_F) displacement of existing PU field lettuce production. Average annual yields for 1 ha of PU field lettuce production averaged 61 tonnes (the data for field lettuce production, both PU and remote, is derived empirically in Section 2.2.2). Whether complete or partial replacement of this quantity was needed, depended upon how much of the PU land was claimed for housing. Whereas the scenarios involving greenfield housing caused a displacement of PU food production, scenarios with infill housing for an equal population (LF, I_HTG_A, 70I:30G_HTG^A or 70I:30G_F) enabled some proportion of PU lettuce production to be retained.

Scenarios that included a PU HTG lettuce production technology needed a smaller land area to produce the same quantity of lettuce, enabling some proportion of the same area of PU land to be used for other purposes — afforestation, in these cases. The area of land made available for afforestation in this way depended upon the ratio of infill to greenfield housing and the HTG lettuce yield, the yield in this instance being 2661 per ha per year (I_HTG_A and 70I:30G_HTG_A).

All developed cities are marked by a combination of housing styles. In scenarios 70I:30G.HTG_A and 70I:30G.F, a ratio of 70 percent infill to 30 percent greenfield housing was adopted, based on

target ratios for new dwellings within the Metropolitan Plan for Sydney 2036 (NSW DOP, 2010). Although area is not a perfect proxy for a target by dwelling type, the aim was to show the comparison between areas of mixed housing types and areas of all greenfield and all infill types; the approach was therefore suitable for the purposes of illustration.

2.2. Life cycle assessment modelling of scenarios

Process-based LCA using SimaPro 8.0.4 software (PRe Consultants, 2014) was used to model the different scenarios on 1 ha of PU land. System boundaries and LCI for both the housing and horticultural systems are shown in Fig. 1. Within SimaPro, LCI were selected from either the Australasian database (Life Cycle Strategies, 2014) or EcoInvent database (Weidema et al., 2013). Sections 2.2.1 and 2.2.2 describe the detailed housing and horticultural system data used in the LCA. The environmental impacts characterised are defined in Section 2.2.3. The parameters representing greenfield housing (i.e., sizes of land parcel and house) are specific to Sydney, Australia; different values may be appropriate for other cities, but the analytical framework should be transferable.

2.2.1. Housing system

For greenfield housing development, the house modelled was a fully detached 300 m2 single storey brick veneer structure on a concrete slab foundation (Crawford, 2011) situated on a 500 m2 parcel of land. Fourteen such houses can be accommodated on 1 ha of land, after allowing for road area. Road area was provided by a land developer for a new greenfield housing precinct (Land-com, personal correspondence, 4 July 2012) and adjusted for 1 ha in proportion to the total developed area. The number of persons per hectare in a greenfield housing development was calculated by multiplying the number of houses per hectare by the average household of three persons, giving 42 persons per hectare. Household size was determined from spatially representative data

Table 2

Description of PU land-use demand for 1 hectare of PU land.

Scenario

Functional unit: 1 ha PU land

Housing system (per ha PU land)

Horticultural system (per ha PU land)

(per ha PU land)

Balance of food production

I_HTG_A

70I:30G_F

70I:30G_HTG_A

Greenfield

Infill

Infill

30% Greenfield

Infill (balance) 0

30% Greenfield

Infill (balance) 0

0.23 Field

Afforestation

Afforestation

Field production, 611 lettuce, remote location

Field production, 611 lettuce, PU location

HTG production (higher productivity than field, requires less land), 61 t lettuce, PU location

Field production, 611 lettuce, both PU and remote location to produce balance

HTG production (higher productivity than field, requires less land), 61 t lettuce, PU location

Lettuce at farm

Seeds Transplants Fertilisers Organic amendments Pesticides Herbicides Fungicides Wetting agents Irrigation water Fuel Electricity Capital equipment

Lettuce at city central market

LCI: Packaging Washing Capital equipment Electricity Transport Refrigeration

FOOD 61 t ha1 yr1

House construction

LCI: Insulations Ceramics

Brick Concrete Glass Gravel Wood Plastics Paints Plasterboard Steel Copper

House Operation and Transportation

Water Electricity Household travel (car, train, bus) Sensitivity to 50:50 solar PV:wind

HOUSING 42 persons

Fig. 1. Life cycle inventory captured within the system boundary for: (a) the food system; and (b) the housing system.

specific to locations where greenfield development has been occurring in north-western and south-western Sydney (ABS, 2011 )3.

Using the figure of 42 persons per hectare, the corresponding number of medium-density mid-rise (four- to six-storey) infill-style apartments (OECD, 1999) was calculated. Housing statistics for the inner west of Sydney indicate two persons per household for apartments four storeys and higher (ABS, 2011 )4. Thus a total of 21 infill apartments are required to house the same population as 1 ha of PU greenfield housing. Infill housing was assumed to be built on existing vacant urban land with no further implications for PU land use.

In addition to the inventory of construction materials, inventories for home operation and occupant transport were also assembled for each housing type. Household transport (household

3 Data aggregated at ABS statistical level 4.

4 Data aggregated at ABS statistical level 4.

travel kilometres) for the two housing types was determined from the Household Travel Survey (NSW Bureau of Transport Statistics, 2014). Household travel kilometres (for the same representative subregions as for household size) were extracted from household travel distance per average weekday and by mode (vehicle driver, train, bus). Annual travel kilometres by mode was calculated on the assumption of a five-day working week for 48 weeks of the year. LCI in SimaPro representing kilometres travelled by car, bus and train includes production and maintenance components. Housing operational energy was determined for greenfield and infill housing based on the same regions as above, using data from the Australian Energy Regulator (AER, undated). Operational energy includes heating and cooling, but does not include areas of community title (in the case of infill housing). Housing operational energy reflects the grid electricity mix for NSW (coal, oil, gas, hydro, wind), including transmission losses. Household water use was based on figures for average household consumption per housing type (IPART, 2004).

Fig. 2. Environmental impact per ha of PU land per year for the five housing and food land-use scenarios: (a) climate change (kg CO2-eq); (b) freshwater eutrophication (kg P-eq); (c) photochemical oxidant formation (kg NMVOC); (d) particulate matter (kg PM10-eq); (e) human toxicity (kg 1,4-DB-eq); (f) water scarcity (m3).

Sensitivity to using twenty percent renewable energy split 50:50 between solar photovoltaic (PV) and wind for operational energy within the housing system was conducted. Twenty percent renewable energy approximates targets established for overall energy mix within the Renewable Energy Directive for the year 2020 (EC, 2009).

2.2.2. Horticultural system

Each scenario was required to produce the same quantity of lettuce for delivery to the city's central fruit and vegetable market.

Data from two PU field lettuce farms and one remote field farm in the state ofVictoria, Australia, were obtained using a pre-structured questionnaire during interviews with farmers. An average yield of 61 tonnes per hectare per year of field-produced lettuce in the PU area was determined from the PU lettuce growers. Where field lettuce production was displaced by greenfield housing, displaced producers were assumed to move to a more remote location with the capacity to expand production. For this study, this location was interstate, approximately 900 km distant (in Victoria). Farms from this location already supply the local city market in competition

with PU farms. Similar large transport distances are found in other parts of the world where commercial produce is imported from more distant regions, such as Italian or Spanish fruit and vegetables exported to other parts of Europe. All scenarios, including displaced horticultural production, were required to supply 61 tonnes of lettuce to the city's central market, as currently supplied by PU field farms.

In addition to field lettuce, a lettuce production system using a HTG was included in scenarios developed for the PU area. Data was obtained using a pre-structured questionnaire from a grower in The Netherlands, since no similar production exists in NSW. Corrections to energy consumption were made for the Sydney climate. The HTG was assumed to use solar PV energy for cooling and baseload energy requirements.

2.2.3. Environmental impacts characterised

Five environmental impact categories relevant to urban areas were analysed using the ReCiPe midpoint (hierarchical) method (Goedkoop et al., 2009) in SimaPro: climate change (kg CO2-eq to air)5; freshwater eutrophication (kg P-eq to freshwater)6; photochemical oxidant formation (kg NMVOC to air)7; particulate matter (kg PM10-eq to air); and human toxicity (kg 1,4-DB-eq to urban air)8. A 100-year timeframe for climate impact was adopted. Water scarcity (m3) was determined using the method in SimaPro of Pfister et al. (2009).

2.3. Key assumptions

Fifty-year lifespans were assumed for housing capital infrastructure. Assumptions of farm capital infrastructure lifetimes have been detailed in Rothwell et al. (2015b). The use of data aggregated by geographical region for household operational energy, transport and size may cause some under- or overestimation, since some degree of combination of housing types co-exists in each region; however, it ensures that spatial relevance is maintained when sourcing data from different datasets, such as the travel survey (NSW Bureau of Transport Statistics, 2014), ABS statistics (ABS, 2011) and Energy Regulator (AER, 2015 undated). In scenarios that include afforestation, a biomass carbon sequestration rate of 3 t C/ha/yr was used, this value being in the vicinity of data reported for various Australian conditions (Hobbs et al., 2013; Paul et al., 2008). Soil carbon sequestration rates in eucalypt forest are reportedly relatively low; a value of 0.11 C/ha/yr has been suggested (Paul et al. in Czimczik et al., 2005).

Data for farms and greenfield housing was collected for larger land areas, but reduced for reporting to 1 ha land parcels to ensure a common functional unit in the LCA. The extent of the emphasis on different housing styles or on different agricultural systems may modify the results. An exhaustive analysis of PU housing development and agricultural food displacement would be a considerable venture, and requires further research.

3. Results

The environmental impacts of climate change, freshwater eutrophication, photochemical oxidant formation, particulate matter, human toxicity and water scarcity were analysed for the five land-use scenarios (Fig. 2). The housing system was found to be dominant in each environmental impact category; a clear trend was the better environmental performance of the infill housing

5 CO2: Carbon dioxide.

6 P: Phosphorus.

7 NMVOC: Non-methane volatile organic carbon compound.

8 1,4-DB: 1,4-dichlorobenzene.

0 -.-,

G_F l_F

■ Without renewable target o20% renewable target

Fig. 3. Effect on climate change per hectare of PU land per year for greenfield (G_F) and infill (1_F) housing and food scenarios for present-day electricity mix vs. 20 percent renewable energy mix split 50:50 between solar PV and wind for housing operational energy.

and food scenarios (1_F and 1_HTG_A) compared to greenfield (G_F) across the indicators selected, with the blended 701:30G scenarios falling between them. Afforestation mitigated the climate change impact of the infill scenarios by up to 5 percent per hectare per year (LHTG.A and 70I:30G_HTG_A). A more detailed analysis of the drivers within each impact category is given in the following sections.

3.1. Climate change

The poorer environmental performance of the greenfield housing and food scenario was highlighted, with a 25 percent increase in climate change impact compared to infill housing (Fig. 2(a)). Scenarios using a HTG with afforestation to replace field lettuce production, such as 1_HTG_A and 70I:30G_HTG_A, demonstrated small but measurable gains in emissions reduction compared to their field production counterparts, 1_F and 70I:30G_F. Hotspots significant for climate change were coal-based generated electricity for operational energy in the housing system (45-62 percent of total), motor vehicle transport (21-35 percent), housing construction (7-8 percent) and lettuce production and delivery to the city market (8-9 percent). The dominant influences on motor vehicle transport included petrol during use and steel during manufacture. Transport by bus and train produced only minor contributions, at approximately 1 percent. Afforestation mitigated climate change impacts for 1_HTG_A and 70I:30G_HTG_Aby 5 and 3 percent, respectively.

3.1.1. Sensitivity to renewable energy

Sensitivity to a 20 percent renewable energy target for housing system operational energy was determined; the results are presented for climate change only (Fig. 3). Using a 20 percent renewable energy target split equally between wind and solar PV, reductions in climate change impact of approximately 9 percent for both the greenfield (G_F) and infill (LF) housing and food scenarios occurred. Reductions were observed in other impact indicators, with the exception of slight increases in freshwater eutrophication (2-5 percent) and water scarcity (no apparent change). 1ncreased freshwater eutrophication results from the use of copper in solar PV cell manufacture.

3.2. Freshwater eutrophication

Reductions in freshwater eutrophication of approximately 35-43 percent occurred in the infill scenarios 1_HTG_A and LF, respectively, compared to the greenfield scenario G_F (Fig. 2(b)). Motor vehicle transport (64-81 percent of total eutrophication

impact) and lettuce production (2-31 percent) produced the bulk of freshwater eutrophication. Petrol, steel, copper and wiring boards dominated eutrophication effects from motor vehicle use and manufacture. Phosphate fertilisers and copper-based fungicides strongly influenced field lettuce production. For housing construction (5-9 percent), concrete products, glass wool insulation and copper were the greatest contributors to eutrophication. The HTG (LHTG.A) produced more freshwater eutrophication than field lettuce (I_F) due to its greater electricity-related component, despite lower eutrophication impacts from fertilisers used therein.

3.3. Photochemical oxidant formation

Infill scenarios I_F and LHTG_A showed approximately 25 percent less photochemical oxidant formation than the greenfield scenario G_F (Fig. 2(c)). Coal-generated electricity for house operation (40-55 percent of photochemical oxidant formation), motor vehicles (18-30 percent) and housing construction (9-11 percent) dominated the housing system. Transport of lettuce to market produced a 9-14 percent impact. Less on-farm diesel fuel was used in HTG lettuce production, giving a lower photochemical oxidant count than field farming. Reusable polypropylene crates for harvested lettuce in HTG production resulted in a better environmental performance than the single-use cardboard carton method of several field farmers.

3.4. Particulate matter

The impact of particulate matter was 25 percent lower in the infill housing and food scenarios I_F and LHTG.A than in greenfield G_F (Fig. 2(d)). Coal-based electricity generation for house operation again created the predominant impact (44-60 percent), with motor vehicles (18-31 percent), housing construction (8-11 percent) and lettuce at market (8-9 percent) also being major contributors. Similarly to photochemical oxidant formation, packaging type and on-farm diesel use were the main contributors of partic-ulate matter in the lettuce-growing system.

3.5. Human toxicity

The infill scenarios led to around 28 percent less human toxicity than greenfield, as illustrated in Fig. 2(e). The predominant impact on the system was caused by coal-based electricity for house operation (38-57 percent of total). Other significant influences included motor vehicles (26-42 percent), lettuce at market (9-10) and housing construction (5-7 percent).

3.6. Water scarcity

The infill scenario I_HTG_A exhibited 26 percent lower water scarcity than G_F, whereas scenario I_F increased it by 9 percent. Potable drinking water used by household residents produced high contributions in all scenarios, at 56-84 percent of impact. Motor vehicles (3-7 percent), housing construction (4-5 percent) and electricity for house operation (less than 5 percent) were relatively minor contributors.

Contributions of the horticultural system to water scarcity ranged from 2 percent in scenario G_F and LHTG.A to 34 percent in LF. The extent of water scarcity in field lettuce production depended on both the water source and the water stress index for irrigation water used on the farm. Regarding water source, greater water scarcity impacts resulted if river water was used in the LCI to represent farm irrigation, as was the case in one of the two PU field farms examined in scenarios LF and 70I:30G_F. When surface water was used from dams on the farm, as was done at the remote farm in scenario G_F, lower scarcity impacts resulted. Changing the water

source from river' to surface' water for the given PU farm resulted in significantly lower water scarcity results, as in scenarios LHTG_A and 70I:30G:HTG_A. For example, the contribution to total impact from the lettuce farms in scenario LF dropped from 34 percent to 2 percent when surface water rather than river water was selected in the LCI.

4. Discussion

PU land use conflict is a problem defined differently by stakeholders in an often contradictory public space. Conflicting goals are evident, such as housing an increasing population as opposed to using the land in a productive or ecosystem-regulating capacity. This study proposed a method of integrating opposing goals within a framework of environmental outcomes. Acknowledging that there is no right' solution to PU land management, careful consideration of the environmental impacts of feeding and housing a projected population in a comprehensive life-cycle approach provides an evidence base for addressing competing claims.

Differently from previous studies was the novel use of a cross-sectoral systems-based approach to analyse the environmental consequences of decisions to urbanise PU horticultural land and thereby displace PU food production. The impacts of both aspects were evaluated using this comprehensive approach rather than within isolated sectors, or not at all. LCA has provided an innovative systems-based solution for informing complex decision-making by providing environmental data that may balance social and economic factors as a sustainable policy objective. Results show that when both housing and food displacement are considered, the housing system still dominates environmental impacts.

4.1. Overall environmental impacts

Results for the environmental impact categories of climate change, freshwater eutrophication, photochemical oxidant formation, particulate matter and human toxicity showed that the infill scenarios LF and I_HTG_A are able to provide the same housing and food production as the greenfield scenario G_F, while simultaneously providing improved environmental performance. Depending upon the category analysed, the environmental impact was reduced by 25 to 43 percent. The blended 70I:30G scenarios either permitted retention of some field production (70I:30G_F) or allowed for HTG production to be introduced together with a level of afforestation (70I:30G_HTG_A). Both of the blended scenarios performed markedly better than the greenfield scenario G_F.

4.1.1. Housing system

The trend of poorer environmental performance of outer suburban greenfield housing compared to medium-density infill housing, as occurred within this study, has also been noted elsewhere (Camagni et al., 2002; Duffy, 2009). The better environmental performance of the infill housing scenarios was primarily due to the higher operational energy and transport requirements ofthe greenfield housing type. Car travel and coal-generated electricity for household operation were found to be the dominant drivers of the large impacts of both housing types; bus and train travel typically accounted for less than 2 percent of all the environmental indicators for both systems.

Due to the fossil intensive nature of these drivers, significant opportunity to reduce environmental impact is associated with strategies that either reduce final energy consumption or reduce impacts from primary energy generation. Reducing final energy consumption would require major urban infrastructure changes, such as energy-efficient housing and better public transport networks to service both inner and (especially) outer urban greenfield areas. The high cost of such infrastructure modifications dictates

that these are longer-term strategies. Transforming the generation of primary energy from fossil fuels to renewable energy sources would represent a technology leap with the capacity for more rapid, accessible and sweeping changes in developed urban centres. Existing electricity transmission infrastructure may be utilised. The use of renewable primary energy sources is recommended as a strategy for reducing environmental impacts in PU areas where housing infrastructure is already established, as well as in new greenfield developments.

Continuing to pursue greenfield development rather than infill development in the current resource-intensive culture may have the effect of locking society into emissions-intensive pathways. The sensitivity analysis illustrated in Fig. 3 highlights the fact that even a renewables target of 20 percent applied to housing operational energy in the greenfield scenario G_F cannot deliver the environmental performance of the infill scenarios 1_F and 1_HTG_A before the renewable energy target was applied. This reveals a need to investigate alternatives to greenfield housing that displaces fresh food production, in order to avoid the risk of infrastructure lock-in.

Although this study was limited to environmental performance, evidence in favour of developing more compact cities has also been couched in economic terms: after infrastructure costs, social services and externalities, infill housing costs have been calculated to be less than those for greenfield housing (Biddle et al., 2006; C1E, 2010). The present study, which has taken a unique, new perspective on PU environmental impacts, supports a compact city approach, complementing the findings of others.

Evidence suggests that compact city forms with preserved green space is a more sustainable planning option, yet urban expansion often continues to be the norm, with PU land simplistically seen as land-in-waiting for urban expansion. Housing dominated all the environmental impacts that were analysed. Policies for improving urban environmental sustainability need to focus on structuring cities with appropriate low-impact housing.

4.1.2. Horticultural system

As electricity and fuels shift to renewable sources over time, the dominance of housing operational energy and private car transport may decline and the horticultural system may gain in relative environmental importance; hence the food system should not be overlooked. The need to feed a growing population, together with the potential impacts of rising temperatures and changed water availability in a business-as-usual climate-change scenario will require a diversity of adaptations. Diversity within the supply chain improves resilience to weather extremes. Solutions include integrated local and regional supply networks sourced from both field production and high-technology protected cropping.

Horticultural LCA studies in northern Europe comparing field production to greenhouse technologies have found that when co-generation or waste heat energy sources were used as the greenhouse energy source, environmental results were better than for field-grown food imported from more remote locations (Nordenstrom et al., 2010). Using solar PV technology for cooling and baseload energy in the present study produced similar results for several environmental indicators. Apart from the impact category of freshwater eutrophication, using a modern fresh food production technology of a PU HTG in combination with medium-density infill apartments provided optimal environmental impacts across the range of indicators analysed. However, this optimal typology of housing and food production technology is currently atypical for a Western developed city such as Sydney, where sprawl and low rates of HTG adoption are characteristic.

Surprisingly, freshwater eutrophication in the HTG exhibited a higher impact than field lettuce farms due to manufacture of solar PV cells. The contribution of fertilisers in HTG production to eutrophication was lower per kilogram than for produce from

field farms; however, countering this and increasing the eutrophi-cation impact was the greater electricity usage. Copper used in solar PV cell production and disposal of associated sulfidic tailings were dominant influences on HTG photovoltaic electricity generation. A second influence was the use of lignite (brown coal) to generate the electricity required in the manufacture of PV cells, as well as for the disposal of associated copper mine spoil. These eutrophica-tion effects are likely to impact in the country of manufacture (e.g. China). Obviously, depending on the manufacturing location for the solar cells, lignite may be used in different proportions or not at all, thereby changing the eutrophication impact.

Water scarcity exhibited similarly unexpected results. Water scarcity is a measure of the extent to which water withdrawals by consumers deprive other users of fresh water in a given geographical region (Pfister et al., 2009). 1n the present study, different results were obtained depending on the source of irrigation water in the horticultural system that was selected in the modelling software. Of the two PU field farms in the study, one used water from dams on the farm (surface water') for irrigation; the second farm used river water. The water-scarcity impact per hectare of produce was several orders of magnitude lower when surface water was selected, as opposed to river water, in the LC1, despite river water being a form of surface water. For the PU HTG, no change was seen in the water scarcity result regardless of water source, whether drinking water, surface water or process water. The differences in the impact of river- and surface water were seen despite the water supply for both farms being in the same hydrological catchment, where water scarcity would be expected to be independent of surface water type. This apparent anomaly suggests that the method of modelling water flow within a given inventory in SimaPro may require further analysis.

The sensitivity to different water source types within the bounds of a single hydrological catchment may have ramifications for decision-making. 1t may be necessary to modify water scarcity methods for water sources specific to either urban or PU regions, if its representation in urban regions is to be improved. As it stands, all water within a single hydrological catchment should be handled in the same way by the present water scarcity method. However, the river water used in field farming may potentially be less in demand for urban consumers than potable water, but an HTG may compete with residential or industrial potable and/or recycled water users unless self-contained. 1n urban environments, water taken from different water sources may deprive other users in the same catchment, thereby raising local water-scarcity issues. Different urban water sources may require assessment at the sub-catchment level.

4.1.3. A wider range of environmental impacts

Overall, the examination of a broad range of environmental impacts as set out in this study enables decision-makers to evaluate potential trade-offs with respect to the use of different food production technologies and housing types in a region. Findings from this study did not detect substantial trade-offs between scenarios; rather, the environmental impacts showed a relatively consistent trend. Despite the unexpected water scarcity results discussed above, the infill scenarios consistently showed better environmental performance across a wide range of impact categories relevant to PU areas in a dominant trend.

4.2. Land use integration in peri-urban settlements

Sparing PU land by developing infill housing and, at the same time, using HTG methods to increase food output as in scenarios 1_HTG_A or 70I:30G_HTG_A released some PU land for other purposes — afforestation, in the present case. In those scenarios, multiple functions (food production, housing and afforestation) were co-located in the urban landscape and simultaneously pro-

duced lower environmental impacts than monofunctional land use such as greenfield housing. Together with co-location, multifunctional land use typically requires interaction between the functions (Selman, 2009). Beneficial interactivity is evidenced by the growing of fresh food for the urban population, combined with business and employment opportunities afforded through implementation of competitive advanced food production technologies. At the ecological level, advanced high-output food production methods spare land for afforestation, whose synergistic effects include the support and regulation of eco-services, and better human health and aesthetics.

With regard to climate change, gains due to afforestation are likely to occur over a long period, possibly as long as 100 to 150 years as forests re-establish and reach equilibrium. Carbon sequestration is one of the many benefits brought by afforestation. No complete environmental measure of the co-benefits of establishing parkland by afforestation has been included within this study. Biodiversity and carbon sequestration benefits from afforestation may be accompanied by less tangible benefits such as microclimate control, improved air and water quality, and recreation and visual amenity. Given the historical human interference in native biomes and the resources required to re-establish a pre-anthropogenic state, afforestation within PU landscapes would be more likely to introduce novel ecosystems with characteristics such as changed combinations and relative abundances of species (Hobbs et al., 2006). Novel ecosystems represent an opportunity to improve the functioning of urban systems, to support sustainability, conservation and biodiversity (Francis and Chadwick, 2013) along with other socially desirable amenity values. PU areas are of particular importance to biodiversity: for example, 50 percent of Australia's threatened species reside in PU habitat (Yencken and Wilkinson, 2000).

The present study has demonstrated that multifunctional PU land use is feasible, in which the often conflicting needs of housing, food production and ecosystem services can be met at the same time as delivering improved environmental performance. Further research involving key stakeholders is needed to strengthen relevant scenarios and explore these interactions for environmental outcomes. Such community based scenario development and exploration would provide a new and valuable input into the current debate over competing claims for PU land.

4.3. Future research directions

Since housing systems, and not food production systems, have been found to dominate environmental impacts in PU regions, suggestions for future research include the establishing of a set of housing and food typologies in which a wider diversity of housing types, transport modes and food production scenarios are modelled. Such housing and food typologies may assist modelling of renewable energy sources and climate mitigation targets.

In this paper, three typologies only were identified: greenfield housing with displaced field production; infill housing with retained field production; and infill housing with an advanced food production technology (HTG) combined with afforestation. Suggestions for other typologies may include housing variants such as greenfield or apartment-style housing with off-grid domestic solar photovoltaic electricity supply, electricity supply from other renewable grid sources, and high-density apartments or other housing system variants unique to particular geographical PU areas. Horticultural variants may include high-technology greenhousing using no external water sources, other food production technologies (vertical gardens, aquaponics) or intra-urban, community-supported agriculture. A comprehensive engagement with stakeholders is necessary to develop and agree on scenarios for future modelling within the LCA framework.

Typologies aimed at modelling renewable energy targets maybe used to improve our understanding of how environmental impacts will be manifested across a range of different housing systems or how the environmental performance gap between greenfield and infill systems, highlighted here, may be narrowed. Scenarios that incorporate such typologies are important for developing effective urban and PU emissions mitigation strategies that aim to minimise the lock-in risks associated with urban land use and infrastructure life spans. If there are known target levels for environmental impacts such as climate change, related targets for renewable energy or other mitigation strategies can be approximated by sensitivity analyses using LCA; for example, if the political environment were to support a halving of climate impacts caused by urbanisation and housing systems, suitable renewable energy targets could be modelled that reflect required results. An investigation of typologies may serve to support the current consensus that infill development delivers better environmental outcomes. On the other hand, such an investigation may identify alternative housing and food systems with enhanced environmental performance.

Beyond the scope of this paper but necessary for a more complete analysis of decisions to urbanise PU horticultural land, is the contribution of direct (earthworks, soil carbon changes and shared services infrastructure) and indirect land-use change to environmental impact. PU horticultural-to-housing land-use change may incur additional environmental debt, particularly for the greenfield scenario. However, the environmental impact contribution of such land-use change is not usually assessed. Few empirical studies have been conducted to evaluate its environmental impact, and further research is needed. This particular aspect of direct and indirect land-use change resulting from peri-urbanisation has been examined elsewhere (Rothwell et al., 2015a,b).

5. Conclusion

Integrated land-use options for PU regions were identified in which housing and feeding an expanding population can both be achieved with lower environmental impacts than is the norm for greenfield development. Infill housing combined with HTG food production provide optimal environmental outcomes for the indicators of climate change, photochemical oxidant formation, particulate matter, human toxicity and water scarcity. Co-benefits such as afforestation were then possible on land no longer needed for housing or food production. Unexpectedly, freshwater eutroph-ication results were higher for HTG lettuce production than for field production. Waste produced during the manufacture of solar PV panels, likely to impact in the country of manufacture (e.g. China), used on the HTG contributed to this. Recommendations for enhancing the representativeness of water-scarcity modelling methods are proposed, for example by taking account of water sources specifically related to urban and PU regions.

This study has given examples of alternative food and housing land-use scenarios in the PU context. Supporting the development of region-specific PU housing and food system typologies will inform climate change consequences, and other environmental impacts, of urban development pathways which are currently cited as a knowledge gap (Dhakal, 2010). Furthermore, housing and food production typologies can be used to determine the attributes of systems that will result in a desired reduction in their environmental impact, such as for climate mitigation purposes.

In addition to providing an evidence base for land-use planning decisions, the process of developing and evaluating scenarios could be the basis for community engagement around the contested issue of PU land use. Facilitating stakeholder engagement in such an exercise has the potential to develop a deeper shared understanding of the environmental consequences of specific courses of action, potentially leading to a consensus view of a desired PU future.

Competing interests

The authors declare no competing interests.

Funding

PhD scholarship funding was provided by the University of Western Sydney, Australia and the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia.

Acknowledgements

Part of the data collection for this research was conducted in parallel with the Climate and Health Cluster under a Commonwealth Scientific and Industrial Research Organisation (CSIRO) Flagship program. Specifically, we would like to thank Gillian Hall and Bron-wyn Isaacs (then of the Australian National University) for the collaboration opportunity. Much appreciation is expressed toward the farmers who contributed their time and knowledge toward this project. We wish to thank the two anonymous reviewers and editor for their valuable suggestions.

References

ABS, 2011. Census of Population and Housing Community Profiles, Australia, Basic Community Profile: Table B31 Dwelling Structure. Australian Bureau of Statistics (ABS), Canberra. ABS, 2013.3222.0 - Population Projections, Australia, 2012 (base) to 2101, 'Table 3: Projected Population, Components of Change and Summary Statistics, Greater Sydney', Data Cube: Excel Spreadsheet, Projected Population, Components of Change and Summary Statistics - Australia, State/territory, Greater Capital City/Balance of State, 2012 (Base) to 2101. Australian Bureau of Statistics (ABS), Canberra. ABS, 2014. 7503.0 - Value of Agricultural Commodities Produced, Australia,

2012-13, 'Table 2: Value of Agricultural Commodities Produced, State and nrm region-New South Wales-2012-13', Data Cube: Excel spreadsheet, Cat. no. 75,030D0002 201213. Australian Bureau of Statistics (ABS), Canberra. American Farmland Trust, 2013. Threatened Farmland: What's happening to our farmland, Available from: http://www.farmland.org/resources/fote/default.asp (accessed 15.12.14.). APA, 2007. American Planning Association Policy Guide on Community and Regional Food Planning. American Planning Association (APA), Chicago, IL, Available from: https://www.planning.org/policy/guides/adopted/food.htm (accessed 17.06.14.). AER, Energy Made Easy: Understand and Compare Your Electricity Usage, Australian Energy Regulator (AER) Available from https://www. energymadeeasy.gov.au/bill-benchmark/results/2135/2 (accessed 15.12.14.). Beardsley, K., Thorne, J.H., Roth, N.E., Gao, S., McCoy, M.C., 2009. Assessing the influence of rapid urban growth and regional policies on biological resources. Landscape Urban Plan. 93,172-183. Biddle, T., Bertoia, T., Greaves, S., Stopher, P., 2006. The costs of infill versus

greenfield development - a review of recent literature. Paper presented at the 29th Australasian Transport Research Forum 2006. Bierwagen, B.G., Theobald, D.M., Pyke, C.R., Choate, A., Groth, P., Thomas, J.V., Morefield, P., 2010. National housing and impervious surface scenarios for integrated climate impact assessments. Proc. Natl. Acad. Sci. 107 (20), 20887-20892.

Boume, L., Bunce, M., Taylor, L., Luka, N., Maurer, J., 2003. Contested ground: the dynamics of peri-urban growth in the Toronto region. Can. J. Regional Sci. Toronto: Department of Geography and Program in Planning, University of Toronto.

Brochu, P.J.S., Yanosky, J.D.S., Paciorek, C.J.P., Schwartz, J.P., Chen, J.T.P., Herrick, R.F.S., Suh, H.H.S., 2011. Particulate air pollution and socioeconomic position in rural and urban areas of the northeastern United States. Am. J. Public Health 101, S224-230.

Brown, K., 2002. Urban Agriculture and Community Food Security in the United States: Farming from the City Center to the Urban Fringe. North American Urban Agriculture Committee, Venice, CA, Available from: http://www.ocfoodaccess.org (accessed 7.05.15.). Budge, T., 2013. Is food a missing ingredient in Australia's metropolitan strategies? In: Farmar-Bowers, Q., Higgins, V., Millar, J. (Eds.), Food Security in Australia: Challenges and Prospects for the Future. Springer Science Business Media, New York, pp. 367-379.

Calderón-Garcidueñas, L., Mora-Tiscareno, A., Ontiveros, E., Gómez-Garza, G., Barragán-Mejía, G., Broadway, J., Chapman, S., Valencia-Salazar, G., Jewells, V., Maronpot, R.R., Henríquez-Roldán, C., Pérez-Guillé, B., Torres-Jardón, R., Herrit, L., Brooks, D., Osnaya-Brizuela, N., Monroy, M.E., González-Maciel, A., Reynoso-Robles, R., Villarreal-Calderon, R., Solt, A.C., Engle, R.W., 2008. Air

pollution, cognitive deficits and brain abnormalities: a pilot study with children and dogs. Brain Cognition 68,117-127.

Camagni, R., Gibelli, M.C., Rigamonti, P., 2002. Urban mobility and urban form: the social and environmental costs of different patterns of urban expansion. Ecol. Econ. 40, 199-216.

C1E, 2010. The Benefits and Costs of Alternative Growth Paths for Sydney: Economic, Social and Environmental 1mpacts. Centre For 1nternational Economics, Canberra.

Crawford, R.H., 2011. Life Cycle Assessment in the Built Environment. Spon Press, Oxon.

Czimczik, C., Mund, M., Schulze, E., Wirth, C., 2005. Effects of reforestation, deforestation, and afforestation on carbon storage in soils. In: Griffiths, H., Jarvis, P. (Eds.), The Carbon Balance of Forest Biomes. Taylor & Francis Group, New York, pp. 342-355.

Dhakal, S., 2010. GHG emissions from urbanization and opportunities for urban carbon mitigation. Curr. Opin. Env. Sust. 2, 277-283.

Duffy, A., 2009. Land use planning in Ireland - a life cycle energy analysis of recent residential development in the Greater Dublin Area. Int. J. Life Cycle Ass. 14, 268-277.

EC, 2009. Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and /30/EC (Text with EEA relevance). European Parliament, Council publication date 23/04/2009.

FAO, 1999. Urban and Peri-Urban Agriculture. A Report to the FAO Committee on Agriculture. Food and Agriculture Organisation of the United Nations (FAO), Rome.

Francis, R.A., Chadwick, M.A., 2013. Urban Ecosystems: Understanding the Human Environment, 1st ed. Routledge, New York.

Frontier Economics, 2011. Efficient Water Resource Pricing in Australia: An Assessment of Administered Scarcity Pricing in Urban Areas. Waterlines report, National Water Commission, Canberra.

Gleeson, B., 2012. Critical commentary. The urban age: paradox and prospect. Urban Stud. 49,931-943.

Goedkoop M., Heijings R., Huibregts M., De Schryver A., Struijs J., Van Zelm R., 2009. ReCiPe 2008: A life Cycle 1mpact Assessment Method Which Comprises Harmonised Category 1ndicators at the Midpoint and the Endpoint Level; First Edition Report 1: Characterisation, R1VM, CML, PRe Consultants and Radboud Universiteit Nijmegen, The Hague.

Haapio, A., Viitaniemi, P., 2008. A critical review of building environmental assessment tools. Environ. Impact Asses. 28,469-482.

Hall, G., Rothwell, A., Grant, T., Isaacs, B., Ford, L., Dixon, J., Kirk, M., Friel, S., 2014. Potential environmental and population health impacts of local urban food systems under climate change: a life cycle analysis case study of lettuce and chicken. Agric. Food Secur. 3,6, http://dx.doi.org/10.1186/2048-7010-3-6

Hobbs, R.J., Arico, S., Aronson, J., Baron, J.S., Bridgewater, P., Cramer, V.A., Epstein, P.R., Ewel, J.J., Klink, C.A., Lugo, A.E., Norton, D., Ojima, D., Richardson, D.M., Sanderson, E.W., Valladares, F., Vila, M., Zamora, R., Zobel, M., 2006. Novel ecosystems: theoretical and management aspects of the new ecological world order Global Ecol. Biogeogr. 15,1-7.

Hobbs, T.J., Neumann, C.R., Tucker, M., Ryan, K.T., 2013. Carbon Sequestration from Revegetation: South Australian Agricultural Regions, DEWNR Technical Report 2013/14. Department of Environment, Water and Natural Resources, Adelaide.

Huang, S.-L., Yeh, C.-T., Chang, L.-F., 2010. The transition to an urbanizing world and the demand for natural resources. Curr. Opin. Env. Sust. 2, 136-143.

Hussey, K., Price, R., Pittock, J., Livingstone, J., Dovers, S., Fisher, D., Hatfield-Dodds, S., 2013. Statutory Frameworks, Institutions and Policy Processes for Climate Adaptation: Do Australia's Existing Statutory Frameworks, Associated Institutions and Policy Processes Support or Impede National Adaptation Planning and Practice? National Climate Change Adaptation Research Facility, Gold Coast, Australia.

Imhoff, M.L., Bounoua, L., DeFries, R., Lawrence, W.T., Stutzer, D., Tucker, C.J., Ricketts, T., 2004. The consequences of urban land transformation on net primary productivity in the United States. Remote Sens. Environ. 89, 434-443.

1PART, 2004. Residential Water Use in Sydney, the Blue Mountains and Illawarra. Independent Pricing and Regulatory Tribunal (1PART) of NSW, Sydney.

1PCC, 2014. Summary for policymakers. In: Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler, A., Baum, I., Brunner, S., Eickemeier, P., Krimann, B., Savolainen, J., Schlomer, S., von Stechow, C., Zwickel, T., Minx, J. (Eds.), Climate Change 2014: Mitigation of Climate Change 2014:Contribution of Working Group 111 to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge.

ISO 14040, 2006a. Environmental Management - Life Cycle Assessment -Principles and Framework. International Organization for Standardization (1SO), Geneva.

ISO 14044, 2006b. Environmental Management - Life Cycle Assessment -

Requirements and Guidelines. International Organization for Standardization (1SO), Geneva.

Life Cycle Strategies, 2014. Australasian LCI database. Life Cycle Strategies, Melbourne.

Liu, J.P., Lewis, G., 2014. Environmental toxicity and poor cognitive outcomes in children and adults. J. Environ. Health 76,130-138.

Lovell, S.T., 2010. Multifunctional urban agriculture for sustainable land use planning in the United States. Sustainability 2, 2499-2522.

Lovell, S., Taylor, J., 2013. Supplying urban ecosystem services through

multifunctional green infrastructure in the United States. Landscape Ecol. 28, 1447-1463.

Low Choy, D., Buxton, M., 2013. Farming the city fringe: dilemmas for peri-urban planning. In: Farmar-Bowers, Q., Higgins, V., Millar, J. (Eds.), Food Security in Australia: Challenges and Prospects for the Future. Springer Science Business Media, New York, pp. 397-412.

Martin, S., Merrick, M., Henderson, T., Mylott, E., Haines, K., Price, C., Koski, A., Dann, B., 2008. Planting Prosperity And Harvesting Health: Trade-offs and Sustainability in the Oregon-Washington Regional Food System. The Institute of Portland Metropolitan Studies, Portland State University, OR.

McDonald, J., 2014. Hot in the city: planning for climate change impacts in urban Australia. In: Christoff, P. (Ed.), Four Degrees of Global Warming: Australia in a Hot World. Routledge, New York, pp. 172-189.

Millar, J., Roots, J., 2012. Changes in Australian agriculture and land use: implications for future food security. Int. J. Agric. Sustain. 10, 25-39.

Neuman, M., 2005. The compact city fallacy. J. Plan. Educ. Res. 25,11-26.

Nilsson, K., Nielsen, T.A.S., Aalbers, C., Bell, S., Boitier, B., Chery, J.P., Fertner, C., Groschowski, M., Haase, D., Loibl, W., 2014. Strategies for sustainable urban development and urban-rural linkages, Research briefings, March 2014. Eur. J. Spatial Dev. 4.

Nordenstrom, E., Guest, G., Froling, M., 2010. LCA of local bio-chip fuelled

greenhouses versus Mediterranean open field tomatoes for consumption in northern Scandinavia. Paper presented at EC0-TECHi10, 22-24 November 2010. Kalmar, Sweden.

NSWDOP, 2010. Metropolitan Plan forSydney. NSW Department of Planning (DOP), Sydney.

NSW Bureau of Transport Statistics, 2014. 2011/12 Household Travel Survey (HTS), five years pooled, Summary Transport Statistics By Local Government Area. Electronic Publication No. E2013-38-HTS-LGA, version 1.1. NSW Bureau of Transport Statistics, Available from http://www.bts.nsw.gov.au/Statistics/ Household-Travel-Survey/default.aspx?FolderID=215#top (accessed 15.12.14.).

O'Farrell, P.J., Anderson, P.M.L., 2010. Sustainable multifunctional landscapes: a review to implementation. Curr. Opin. Environ. Sust. 2,59-65.

OECD, 1999. 505.4 Bill of Quantities for an Apartment, Eurostat-OECD PPP

Programme Comparison of Construction Prices for Residential Buildings. OECD, Luxembourg.

Paul, K.I., Jacobsen, K., Koul, V., Leppert, P., Smith, J., 2008. Predicting growth and sequestration of carbon plantations growing in regions of low-rainfall in southern Australia. Forest Ecol. Manag. 254, 205-216.

Pfister, S., Koehler, A., Hellweg, S., 2009. Assessing the environmental impacts of freshwater consumption in LCA. Environ. Sci. Technol. 43,4098-4104.

Piorr, A., Ravetz, J.R., Tosics, I., 2011. Peri-Urbanisation in Europe: Towards European Policies to Sustain Urban-Rural Futures. A Synthesis Report. University of Copenhagen/Academic Books Life Sciences, Denmark.

Pires, V., Burton, P., 2013. Help or hindrance? The relationship between land use planning and urban agriculture on the gold Coast. In: Farmar-Bowers, Q., Higgins, V., Millar, J. (Eds.), Food Security in Australia: Challenges and Prospects for the Future. Springer Science Business Media, New York, pp. 381-396.

PRe Consultants, 2014. SimaPro. PRe Consultants, The Netherlands.

Randolph, B., Tice, A., 2013. Who lives in higher density housing? Study of spatially discontinuous housing sub-markets in Sydney and Melbourne. Urban Stud. 50, 2661-2681.

Richardson, H.W., Bae, C.H.C., 2004. Urban sprawl in Western Europe and the United States. Ashgate, Aldershot.

Rothwell, A., Ridoutt, B., Page, G., Bellotti, W., 2015a. Direct and indirect land use change: prospective environmental indicators for PU development transitions. J. Environ. Plann. Manage., Advance online publication. doi: http://www.tandfonline.com/doi/full/10.1080/09640568.2015.1035775.

Rothwell, A., Ridoutt, B., Page, G., Bellotti, W., 2015b. Environmental performance of local food: trade-offs and food security implications for a developed city. J. Clean. Prod., Advance online publication. doi:http://dx.doi.org/10.1016/j.jclepro.2015.04.096.

Rowley, S., Phibbs, P., 2012. Delivering Diverse and Affordable Housing on Infill Development Sites. AHURI Final Report No. 193. Australian Housing and Urban Research Institute (AHURI), Melbourne.

Russo, P., Tomaselli, G., Pappalardo, G., 2014. Marginal periurban agricultural areas: a support method for landscape planning. Land Use Policy 41, 97-109.

Sala, S., Farioli, F., Zamagni, A., 2013. Progress in sustainability science: lessons learnt from current methodologies for sustainability assessment: part 1. Int. J. Life Cycle Ass. 18,1653-1672.

Schetke, S., Haase, D., Kotter, T., 2012. Towards sustainable settlement growth: a new multi-criteria assessment for implementing environmental targets into strategic urban planning. Environ. Impact Asses. 32,195-210.

Scholz, M., Hedmark, A., Hartley, W., 2012. Recent advances in sustainable

multifunctional land and urban management in Europe: a review. J. Environ. Plann. Manage. 55, 833-854.

Selman, P., 2009. Planning for landscape multifunctionality. SSPP 5, 45-52.

Shearer, A., Mouat, D., Bassett, S., Binford, M., 2009. Discussion and conclusions. In: Shearer, A., Mouat, D., Bassett, S., Binford, M., Johnson, C., Saarinen, J., Gertler, A., Kahyaoglu-Koracin, J. (Eds.), Land Use Scenarios: Environmental Consequences ofDevelopment. CRC Press, Boca Raton, Florida, pp. 373-387.

Sicard, P., De Marco, A., Troussier, F., Renou, C., Vas, N., Paoletti, E., 2013. Decrease in surface ozone concentrations at Mediterranean remote sites and increase in the cities. Atmos. Environ. 79, 705-715.

Siedentop, S., Fina, S., 2012. Who sprawls most? Exploring the patterns of urban growth across 26 European countries. Environ. Plann. A. 44, 2765-2784.

Simon, D., 2008. Urban environments: issues on the peri-urban fringe. Annu. Rev. Env. Resour. 33, 167-185.

Thabrew, L., Wiek, A., Ries, R., 2009. Environmental decision making in

multi-stakeholder contexts: applicability of life cycle thinking in development planning and implementation. J. Clean. Prod. 17,67-76.

Thompson, E., Harper, A.M., Kraus, S., 2008. Think Globally - Eat Locally: San Fransico Foodshed Assessment. American Farmland Trust, Washington, D.C.

Thorne, J.H., Santos, M.J., Bjorkman, J.H., 2013. Regional assessment of urban impacts on landcover and open space finds a smart urban growth policy performs little better than business as usual. PLoS One 8, e65258.

Tong, S.T.Y., Chen, W., 2002. Modeling the relationship between land use and surface water quality. J. Environ. Manage. 66,377-393.

United Nations Human Settlements Programme, 2011. Cities and Climate Change: Global Report on Human Settlements 2011. Earthscan Ltd., London, UK and Washington, DC USA.

van der Schans, J., Wiskerke, J., 2012. Urba agriculture in developed economies. In: Viljoen, A., Wiskerke, J. (Eds.), Sustainable Food Planning. Wageningen Academic Publishers, The Netherlands, pp. 245-258.

Westerink, J., Haase, D., Bauer, A., Ravetz, J., Jarrige, F., Aalbers, C.B.E.M., 2012. Dealing with sustainability trade-offs of the compact city in peri-urban planning across European city regions. Eur. Plan. Stud. 21,473-497.

Weidema, B.P., Bauer, C., Hischier, R., Mutel, C., Nemecek, T., Reinhard, J., Vadenbo, C.O., Wernet, G., 2013. The ecoinvent database: Overview and methodology, Data quality guideline for the ecoinvent database version 3, www.ecoinvent. org, Zurich.

Wiek, A., Binder, C., Scholz, R.W., 2006. Functions of scenarios in transition processes. Futures 38, 740-766.

Yencken, D., Wilkinson, D., 2000. Resetting the Compass: Australia's Journey Towards Sustainability. CSIRO Publishing, Melbourne.

Zasada, I., 2011. Multifunctional peri-urban agriculture - a review of societal demands and the provision of goods and services by farming. Land Use Policy 28, 639-648.