annals of the association of american geographers a...

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Florida State University Libraries]On: 10 June 2011Access details: Access Details: [subscription number 789349894]Publisher RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Annals of the Association of American GeographersPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t788352614

A Geographic Approach to Sectoral Carbon Inventory: Examining theBalance Between Consumption-Based Emissions and Land-Use CarbonSequestration in FloridaTingting Zhaoa; Mark W. Hornera; John Sulikb

a Department of Geography and Institute for Energy Systems, Economics, and Sustainability, FloridaState University, b Department of Geography, Florida State University,

First published on: 28 April 2011

To cite this Article Zhao, Tingting , Horner, Mark W. and Sulik, John(2011) 'A Geographic Approach to Sectoral CarbonInventory: Examining the Balance Between Consumption-Based Emissions and Land-Use Carbon Sequestration inFlorida', Annals of the Association of American Geographers, 101: 4, 752 — 763, First published on: 28 April 2011 (iFirst)To link to this Article: DOI: 10.1080/00045608.2011.567936URL: http://dx.doi.org/10.1080/00045608.2011.567936

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

http://www.informaworld.com/smpp/title~content=t788352614

http://dx.doi.org/10.1080/00045608.2011.567936

http://www.informaworld.com/terms-and-conditions-of-access.pdf

A Geographic Approach to Sectoral CarbonInventory: Examining the Balance Between

Consumption-Based Emissions and Land-UseCarbon Sequestration in Florida

Tingting Zhao,∗ Mark W. Horner,∗ and John Sulik†

∗Department of Geography and Institute for Energy Systems, Economics, and Sustainability, Florida State University†Department of Geography, Florida State University

Carbon accounting is an important analytical task that provides baseline information to assist in establishingemissions targets, developing market-based carbon trading programs, and facilitating sustainable carbon manage-ment at the regional to international scales. Although a substantial amount of research has focused on carbonemissions inventory, limited studies have been conducted to estimate consumption-based emissions and theirspatial distribution in relation to vegetation carbon sinks. In this article, we develop a new approach to modelthe spatially detailed consumption-based carbon emissions from the household energy and transportation sec-tors. Emissions were in turn integrated with vegetation carbon sequestration rates that were modeled throughbiophysical remote sensing techniques. This enables carbon balance analysis at detailed geographic locations.To illustrate our approach for carbon balance analysis, we present a case study in Florida. Results indicate that,in 2001, Florida was able to self-assimilate residential carbon emissions from energy and transportation fuelconsumption. Estimates indicate that net carbon sources (i.e., household emissions exceeding vegetation carbonassimilation) are associated with urban and suburban densities and net sinks with exurban and rural densities.In sum, the research approach can be extended to household energy consumption and carbon assessment forother geographies at alternative scales. Key Words: carbon emissions inventory, GIS, household energy consumption,household transportation fuel consumption, vegetation carbon sinks.

La contabilidad del carbono es una tarea analıtica importante que da informacion de base para ayudar a establecermetas de emisiones, mediante el desarrollo de programas comerciales de carbono basados en el mercado, y facilitarel manejo sostenible del carbono a escalas regionales e internacionales. Si bien una sustancial cantidad deinvestigacion se ha enfocado al inventario de las emisiones de carbono, tambien se han llevado a efecto limitadosestudios para calcular las emisiones basadas en consumo y su distribucion espacial en relacion con los sumiderosde carbono de la vegetacion. En este artıculo desarrollamos un nuevo enfoque para modelar las emisiones decarbono del consumo de los sectores energeticos de hogares y transporte, detalladas espacialmente. A su turno lasemisiones fueron integradas con tasas de secuestro del carbono por la vegetacion, modeladas a traves de tecnicasbiofısicas de percepcion remota. Esto permite el analisis del balance de carbono en localizaciones geograficasdetalladas. Para ilustrar nuestro enfoque de analisis del balance del carbono, presentamos un estudio de casoen la Florida. Los resultados indican que, en 2001, la Florida pudo auto-asimilar emisiones residenciales de carbono

Annals of the Association of American Geographers, 101(4) 2011, pp. 752–763 C© 2011 by Association of American GeographersInitial submission, February 2010; revised submission, August 2010; final acceptance, December 2010

Published by Taylor & Francis, LLC.

Downloaded By: [Florida State University Libraries] At: 20:48 10 June 2011

A Geographic Approach to Sectoral Carbon Inventory 753

provenientes del consumo de combustibles para energıa y transporte. Los estimativos indican que las fuentesnetas de carbono (esto es, las emisiones de hogares que exceden la asimilacion de carbono por la vegetacion)estan asociadas con la densidad urbana y suburbana y con sumideros netos con densidades transurbanas y rurales.En suma, el enfoque de la investigacion puede extenderse al consumo familiar de energıa y a la evaluacion delcarbono para otras geografıas a escalas alternativas. Palabras clave: inventario de emisiones de carbono, SIG, consumode energıa familiar, consumo de combustibles para transporte familiar, sumideros de carbono de la vegetacion.

Anthropogenic climate change challenges soci-eties throughout the world. To slow the in-crease in global temperature, curbing emissions

of carbon dioxide (CO2), which account for nearly 75percent of global annual greenhouse gas emissions, willbe required worldwide (Intergovernmental Panel onClimate Change [IPCC] 2007). The global reductionof CO2 emissions relies heavily on localized mitigationefforts (Knight et al. 2003) for which future targets aredetermined mainly by a country or region’s emissionshistory. For example, the Kyoto Protocol establisheddifferent rates of emissions reduction among industri-alized countries based on their development activitiessince the industrial revolution. The Protocol also re-quired that each participating country provides data toestimate its respective 1990 emission levels to designatetheir future emission targets (United Nations 1998).

A carbon emissions inventory, or the estimation ofa country or region’s historical emissions, quantifiessources and sinks of anthropogenic CO2 over a cer-tain geographic area and time period (U.S. Environ-mental Protection Agency [EPA] 2009a). It requiresapproaches that capture the spatial distribution of CO2sources associated with human activities as well as vege-tation carbon sinks, which can be either anthropogenic(e.g., planted trees) or natural (e.g., undisturbed rain-forests). An assessment of the literature suggests limi-tations with regard to several critical issues involved inthe integrated assessment of historical emissions.

First, the calculation of carbon emissions is oftenbased on production instead of consumption (Blasing,Broniak, and Marland 2005; Petron et al. 2008). In thisstudy we refer to consumption as consumer-based energyand material use. Traditional carbon emissions inven-tories have been performed based on combustion (e.g.,electricity generation) or end use (e.g., iron and steelproduction) of fossil fuels, which are both on the sideof industrial production. The setting of future emissiontargets according to production-based inventories hasits pitfalls, because the consumption and production offossil fuel–related energy and manufacturing productsare often spatially disparate (Rose et al. 2005; We-ber and Matthews 2008). This spatial mismatch be-tween emissions producers and end users might have

impaired international agreements on carbon emissionsallowances (Helm 2008).

Second, emissions estimates are usually performedat coarse geographic scales such as by city, state, orcountry boundaries, with no disclosure of the spatialvariation within those units (Gregg et al. 2009; Sova-cool and Brown 2010). The development of continuousdata sets that describe intraunit geographic and tempo-ral patterns of carbon emissions is still in an exploratorystage. This has been performed for the globe at 1◦× 1◦

spatial resolution by allocating a country’s total CO2emissions based on distribution of that country’s pop-ulation (Andres et al. 1996). More recently, emissionsfrom fossil fuel–based combustion, known as the Vul-can inventory, were generated for the United States ata 10-km spatial resolution by incorporating emissionsdata available at electrical generating units and from ve-hicle miles traveled (Gurney et al. 2009). No spatiallyexplicit accounts of consumption-based emissions haveyet been reported.

Third, the integrated assessment of balance betweenanthropogenic emissions and natural carbon sinks isstill limited to a few urban areas (Wentz et al. 2002;Pataki et al. 2009) as opposed to an inventory that cap-tures carbon balance differentiation along a gradient ofsettlement densities. For the accounting of vegetationcarbon sinks, a recent trend is to produce spatially de-tailed estimates from syntheses of ground surveys andecological models with the addition of remote sensingobservations (Cramer et al. 1999). These inventories,however, exist predominantly for spatially continuousvegetation communities (Imhoff et al. 2004). The ac-counting of carbon balance in human-dominated en-vironments (i.e., cities, suburbs, and exurbia) requiresinterdisciplinary approaches that combine traditionalsocioeconomic and ecological studies (Grimm et al.2008; Churkina, Brown, and Keoleian 2009; Wise et al.2009).

In this article, we present an integrated researchapproach to model the CO2 balance betweenconsumption-based emissions and vegetation carbonsinks. Our study targets current research needs inemissions inventories with respect to the three pointspreviously discussed. Consumption in this analysis


754 Zhao, Horner, and Sulik

includes the household usage of energy and trans-portation fuels, which account for approximately40 percent of the carbon emissions in the UnitedStates (Brown, Southworth, and Sarzynski 2008).This inventory excludes energy and fuel consumptionand associated carbon emissions from the commercialsector. The research approach presented in this articlecaptures the spatial heterogeneity of energy and fuelconsumption, consumption-based CO2 emissions,and vegetation carbon sinks at the U.S. Census tractscale. The spatially explicit accounting of householdconsumption and carbon emissions or sinks werefurther examined among different settlement typescharacterized by their residential densities. Our ap-proach integrates carbon emissions estimates basedon socioeconomic data with ecological modeling ofvegetation carbon sequestration potentials based onbiophysical remote sensing methods. All data used inthis study are publicly available in the United Statesand the research methods can be extended to othergeographies, including the possibility of nationwideanalysis at the Census tract and comparable scale.

Integrated Spatially Explicit CarbonBalance Analysis

This article develops a carbon accounting approachthat (1) investigates sources of carbon emissions at theconsumer end, (2) captures the spatial heterogeneityof energy and fuel consumption and related carbonemissions, and (3) provides a profile of carbon balancebetween emissions from consumption activities andsequestration by natural or managed vegetation. Toillustrate our methods, we demonstrate an applicationinvolving the state of Florida, which is one of thelargest states (by size) in the United States. In 2007,an executive order was issued on carbon emissionsreduction targets for this state (State of Florida 2007).

In the United States, consumption-based emissionsinventories are challenging to conduct at the substatescale. Although the U.S. Energy Information Admin-istration (EIA) has collected point-location data aboutresidential energy consumption since 1978, these dataare distributed in aggregated form at the scale of Censusregions, Census divisions, and states (EIA 2009). It isnecessary to consider consumption patterns associatedwith households and their characteristics to estimateenergy-related residential consumption at finer scales.Our methods make use of average consumption levelsthat vary by household size, as households form the basic

measuring unit of home energy use. When this is com-bined with Census counts of households by householdsize, we are able to produce spatial estimates of residen-tial energy consumption at the Census tract scale.

The different orientations of the social and naturalsciences might explain the limited number of spatiallyoriented carbon balance analyses that account for bothanthropogenic emissions and vegetation sequestration(Pickett et al. 2001). In fact, there are relatively fewsuch endeavors. Among the limited efforts focusing onmodeling carbon balance across heterogeneous land-scapes, some scholars simulate the spatial distributionof the net carbon flow based on actual measures of CO2concentration in the atmosphere without partitioningthe respective contribution of carbon sources and sinks.For example, in one study the spatial and temporal pat-terns of the atmospheric CO2 were derived based onstatistical regression of measured CO2 concentrationand variables such as traffic flow, population density,employment density, and vegetation density (Wentz etal. 2002). Other efforts tend to model carbon emis-sions and sinks separately and then derive the balancebetween those two. A recent study conducted in SaltLake City, Utah, demonstrated an integrated approachthat inventoried CO2 emissions based on consumptionrecords from local utility companies and estimated CO2sequestration rates based on a model of urban forestgrowth (Pataki et al. 2009). In this article, we aim to in-ventory emissions and sequestration separately to iden-tify spatial locations associated with carbon sources andsinks. The vegetation carbon sinks are estimated basedon photosynthesis activities of all green vegetation (in-cluding forests, shrub, grass, agriculture, and wetlands)using a modeling technique of biophysical remote sens-ing. This approach provides a continuous surface of veg-etation carbon sink estimates at the spatial resolutionof one kilometer.

Study Area

Florida is located in the southeast of the UnitedStates (Figure 1). It ranks as the fourth most popu-lous state in this country (U.S. Census Bureau 2009).The climate of Florida ranges from subtropical in thenorth and central parts of the state to tropical inthe southern part of the Keys (Box, Crumpacker, andHardin 1993). Although the per capita total energyconsumption of Florida ranked low (forty-fourth inthe country), Florida’s per capita residential electric-ity demand is relatively high. This could be attributedto the high cooling demand during the hot summer



Figure 1. Florida (highlighted in yel-low in the insert map) is located inthe southeast of the United States.Vegetation and land-cover types wereadapted from the 2001 National LandCover Database (Homer et al. 2007).U.S. topographic map and place namereference came from ArcGIS ResourceCenters.

months and the extensive use of electricity (87 per-cent of share in 2000) for home heating during thewinter months (EIA 2010). Florida’s per capita car-bon emissions, calculated based on fossil fuels used forenergy production and manufacturing products, werebelow the national average during the time periodbetween 1960 and 2000 (Blasing, Broniak, and Mar-land 2005). This might be due partly to Florida’s focuson tourism and agriculture instead of energy-intensivemanufacturing.

Data and Methods

The research approach on carbon accounting forFlorida consists of two parts. First, we investigated CO2emissions released from residential household energyand transportation fuel consumption at the U.S. Cen-sus tract scale. Second, we estimated vegetation carbonsinks measured by plant net photosynthesis at 1-km res-olution using remote sensing and ecological approaches,the results of which were aggregated to Census tracts.The household energy and transportation fuel con-sumption, consumption-based carbon emissions, veg-etation carbon sinks, and carbon balance were thenconnected to four settlement categories as measured byCensus housing unit densities.

Estimating Household Energy and FuelConsumption and Carbon Emissions

In this study, household carbon emissions were mea-sured as CO2 emitted through household energy useand transportation fuel. The total annual CO2 emis-sions from household energy consumption in 2001 (Ee ;kg) were estimated for each Census tract in Florida usingthe following equation:

Ee =(

6∑i =1

H Hi × Btui

)× 74.046 × 0.697 (1)

where∑6

i =1 H Hi × Btui (million Btu) is the annualtotal household energy consumption for a given Censustract. HHi is the Census 2000 total number of house-holds by household size (e.g., one-person householdvs. two-person household) within that Census tract, iranges from one to six (with six representing householdsof six or more residents). Btui (million Btu) is the av-erage per household energy consumption by householdsize in 2001 (EIA 2001a). In Florida, approximately97 percent of the total household energy consumptioncame from electricity; consumption of 1 million Btuwas equivalent to 74.046 kWh generated through elec-tricity (EIA 2001b). The CO2 emission coefficient was



0.697 kg/kWh for electricity generation in Florida (EIA2002). Therefore, the total annual CO2 emissions (kg)from household energy use were approximated basedon energy consumption multiplied by the two scalars inEquation 1.

We used a standard method published by the U.S.EPA to estimate CO2 emissions from household trans-portation (U.S. EPA 2005). According to this method,a typical passenger vehicle was modeled as emitting8.877 kg of CO2 per gallon of gas used. The total an-nual CO2 emissions from household transportation in2001 (Et ; kg) were obtained for each Census tract inFlorida using Equation 2:

Et = V MTMP G

× 8.877 × 365 (2)

where VMT is the daily vehicle miles of travel (VMT)per Census tract calculated based on the NationalHousehold Travel Survey (NHTS) transferability dataset (Oak Ridge National Laboratory 2007). The NHTScollected daily commuting and noncommuting trips ofall members in the sampled households. Those nation-wide point samples were transferred to the aggregatedtravel behavior by Census tract through integrating set-tlement patterns and other socioeconomic data (Huet al. 2007). Households were categorized into clustersbased on settlement patterns, economic characteristics,and demographic characteristics of the Census tractswhere they were located. The sample-based VMT val-ues were assigned to all households within the sameclusters and then summed up by Census tracts (Andrews2008). MPG is the average miles per gallon (MPG) es-timated based on the national average fuel economyfor passenger vehicles (20.3 mpg) used in the EPA Mo-bile6.2 emissions modeling program (U.S. EPA 2005,2009b).

Calculating Vegetation Carbon AssimilationCapacity

The potential vegetation carbon sinks were measuredas gross primary production (GPP) of land multipliedby a scalar 0.5. GPP refers to the total amount of CO2entering an ecosystem through photosynthesis duringa year (Chapin, Matson, and Mooney 2002). Approxi-mately half of the total GPP is used for plant respiration(DeLucia et al. 2007). GPP minus plant respiration (i.e.,net photosynthesis) indicates the maximum potentialvegetation assimilation of CO2 from the atmosphere.

Estimates of GPP were obtained using the light-use efficiency approach (Running et al. 2004; Zhao,Brown, and Bergen 2007). Our analysis includes deriv-ing the actual light-use efficiency (ε) parameters andestimating the absorbed photosynthetically active radi-ation (APAR) for different types of vegetation and landcover (Zhao 2011). The 2001 National Land CoverDatabase (Homer et al. 2007) was used to generateten vegetation and land-cover types throughout Florida(Figure 1). ε was derived for each vegetation and land-cover type based on documented empirical measures(Yang et al. 2007) multiplied by temperature and va-por pressure deficit (VPD) scalars (unpublished data).APAR was modeled based on the 2001 biweekly Ad-vanced Very High Resolution Radiometer (AVHRR)normalized difference vegetation index (NDVI; U.S.Geological Survey EROS Data Center 2006) and solarradiation (Mitchell et al. 2004). The estimates of GPP,calculated at 1-km resolution throughout Florida, wereaggregated to Census tracts for further analyses.

Linking Energy and Fuel Consumption and CarbonBalance to Settlement Patterns

The tract-level household energy and fuel consump-tion and carbon emissions and sinks were comparedacross four settlement categories including urban, sub-urban, exurban, and rural densities. Following the defi-nition of Theobald (2001), urban densities are definedas settlement at 1 or more housing units per acre,suburban densities at 0.1 to 1 units per acre, exur-ban densities at 0.025 to 0.1 units per acre, and ruraldensities at less than 0.025 units per acre. Univari-ate analysis of variance (ANOVA) was constructed todetect effects of settlement category on household en-ergy and fuel consumption, consumption-based carbonemissions, vegetation carbon sequestration, and carbonbalance, respectively. The dependent variable was thetract-level per unit area or per capita measure of con-sumption or CO2, and the factor variable was settlementcategory. Mean values for energy and fuel consumptionand CO2 estimates were derived and reported by settle-ment categories.

Results and Discussion

The State-Level Estimates

Florida, as a whole, appears to contain sufficient car-bon sinks to offset all CO2 emitted through householdenergy and fuel consumption in 2001. The estimatedtotal annual household energy consumption was 0.58



quadrillion Btu for the entire state, which would pro-duce approximately 30 million tons of CO2 emissions.The total annual household vehicle travel was esti-mated to be approximately 74 billion miles for the en-tire state, which was responsible for approximately 32million tons of CO2 emissions. Therefore, the total an-nual CO2 emitted through household energy and trans-portation fuel consumption in 2001 was approximately62 million tons, which is less than the vegetation car-bon assimilation capacity (88 million tons) as measuredby plant net photosynthesis during the same year.

According to the last decennial Census, Florida washome to approximately 16 million residents and 6.3million households in 2000. The per capita annualhousehold energy and transportation fuel emissionswere estimated to be 3.9 tons of CO2 in 2001, whereaseach Florida household accounted for 9.8 tons ofCO2 emissions annually on average. According to ourestimates, the potential vegetation carbon sinks inFlorida were 5.5 tons per person and approximately 14tons per household.

Energy and Fuel Consumption and Carbon Balanceby Census Tract

The spatial variation in household energy andtransportation fuel consumption was high at theCensus tract scale (Figure 2). Tracts with high energyconsumption ranks might not be associated withhigh levels of transportation fuel usage consideringthat Pearson’s correlation is 0.3 (significance =0.01) between the tract-level total household energyconsumption and VMT. The consumption-based CO2emissions normalized by Census tract area (i.e., densityof emissions) followed roughly the distribution ofpopulation centers. The correlation between emissionsfrom energy use and emissions from transportationfuel is relatively high (Pearson’s correlation = 0.774,significance = 0.01) according to the tract-leveldensity-based emissions measures.

The density-based carbon sinks estimate, or plantnet photosynthesis per unit area, also varied signifi-cantly throughout Florida (Figure 3). At the Censustract scale, plant net photosynthesis correlated nega-tively (Pearson’s correlation = –0.506, significance =0.01) with the energy-based emissions measured by den-sity. This indicates that tracts with higher levels of CO2emissions are less likely to be home to vegetation car-bon sinks, which is due possibly to the reduction ingreenspace associated with residential and commercialdevelopment.

Figure 2. The Florida household energy consumption (A) and ve-hicle miles of travel (B) by Census tract in 2001.

Carbon balance, calculated as the total emissionsfrom energy and transportation fuel consumptionsubtracting plant net photosynthesis, was generatedstatewide by Census tract. The spatial pattern of the es-timated carbon balance varies significantly when mea-sured by the total, per unit area, and per capita estimates(Figure 4). When measured by per unit area CO2 bal-ance (Figure 4B), densely populated cities such as Mi-ami, Fort Lauderdale, Tampa, St. Petersburg, Orlando,and Jacksonville were found to be associated with netCO2 sources (i.e., emissions exceeding vegetation car-bon assimilation). Measured by per capita CO2 balance(Figure 4C), the net carbon sources tend to extend awayfrom the densest settlement areas.

Energy and Fuel Consumption and Carbon Balanceby Settlement Densities

Four settlement categories were generated for Floridabased on Census 2000 housing-unit densities (Figure 5).



Figure 3. Vegetation carbon assimilation potential measured asgross primary production (GPP) at 1-km resolution (A) and as plantnet photosynthesis aggregated by Census tract (B).

Approximately half of the state’s total population andhouseholds resided in urban areas, which accounted forless than 4 percent of the total area of Florida. The ur-ban and suburban areas together were home to nearly90 percent of the total state population and households.Exurban densities accounted for approximately 30 per-cent of the state’s total land and 8 percent of the state’stotal population and households. Rural densities occu-pied the largest proportion (45 percent) of the state’stotal land, with less than 4 percent of the total popula-tion and households distributed within those areas.

According to an ANOVA (Table 1), effect of set-tlement category on consumption and CO2 levels isgreater when measured by per unit area estimates thanby per capita measures. The household energy consump-tion over per unit area was significantly higher at urbanand suburban densities than the same measure at ex-urban and rural densities (Table 2), whereas the percapita energy consumption was less distinctive amongthe four settlement categories (Table 3). No variation in

Figure 4. Carbon balance between household emissions and veg-etation sequestration measured per Census tract (A), per squarekilometer (B), and per capita (C) by Census tract.

household transportation measured by per capita VMTcan be attributed to settlement category (Table 1). Ex-urban densities, however, were found to be associatedwith the highest per capita VMT at a rate significantly



Figure 5. Settlement densities based on Census 2000 housing-unitdensity (Urban: � 1 housing unit per acre; Suburban: 0.1–1 housingunits per acre; Exurban: 0.025–0.1 housing units per acre; Rural: <

0.025 housing units per acre).

higher than the same estimate at urban densities (Ta-ble 3). This result corroborates earlier research findingsabout the increasing transportation needs within theperiurban zones (Ewing and Rong 2008; Brownstoneand Golob 2009).

With regard to vegetation carbon sinks, both exur-ban and rural densities were associated with the in-creased amount of plant net photosynthesis, comparedto the urban and suburban densities (Table 2 and Table3). The carbon assimilation potential of exurban andrural densities appears to be similar when measured bycarbon flux density (Table 2), whereas rural densitiesare superior for vegetation carbon assimilation than ex-urban densities when measured by per capita estimates(Table 3).

Settlement category accounted for the balance be-tween household emissions and vegetation carbon as-similation (Table 1). Estimated by carbon flux density,urban and suburban densities appear to be net carbonsources, whereas exurban and rural densities are associ-ated with slight carbon sinks (Table 2). Estimated byper capita measures, urban and suburban densities areshown to be net carbon sources at a much smaller mag-nitude compared to the size of net sinks at exurban andrural densities (Table 3).

Policy Implications and Future Research Directions

Consumption-based emissions analyses offer an ap-proach to capture the carbon sources principally associ-

ated with household activities, a key driver of fossil fueluse. This study investigates emissions from home energyuse and transportation sectors. Going forward, future re-search might consider including other household-levelfossil fuel–related consumption items. These might in-clude goods and products such as food and apparel asreported through consumer spending surveys. Theseanalyses would provide further information to supportemission trading programs, especially those involvinghousehold emission credit transfers based on fossil fuelconsumption (Niemeier et al. 2008).

Another avenue could incorporate information onthe commercial sectors for more comprehensive emis-sions accounting at (supra)national levels. This type ofanalysis would further contribute to helping set emis-sions targets and distribute responsibilities, which aretop priorities for policymakers concerned with carbonemissions reductions and climate change mitigation(Hepburn and Stern 2008). Related to this, time se-ries analysis of energy consumption and consumption-based emissions would provide insights for the efficacyof emissions reduction practices.

Per our results, the highly concentrated energy andfuel consumption levels recorded in urban and subur-ban areas, accompanied by lower amounts of vegeta-tion carbon sinks, suggest the need for emissions reduc-tion in high-density areas through cutbacks in energyuse (possibly through renewable resources or new formsof mass transportation) and the utilization of carbon-efficient plant species. Exurban densities are charac-terized by highly efficient vegetation carbon assimila-tion, yet they are associated with the highest per capitahousehold transportation demand. This suggests thatlowering transportation emissions in low-density ar-eas is a need, although the strategies for doing so arenot clear. These environments tend to be unsupportiveof travel reduction initiatives such as carpooling cam-paigns, because the complex spatio-temporal patternsof individual movement among disparate activity loca-tions limit cooperative transport opportunities (Horner2004). At the same time, mass transit options in theselocations are also very limited.

Besides examining the spatial and temporal patternsof carbon emissions and sequestration, future researchmight wish to explain these patterns as a function ofsocioeconomic variables such as housing-unit charac-teristics (Newton, Tucker, and Ambrose 2000), house-hold income levels (Druckman and Jackson 2008), oreven the lifestyle characteristics of household membersthemselves (Lutzenhiser 1993). This contributes to abetter understanding of social and economic drivers of



Table 1. Effects of settlement category (consisting of urban, suburban, exurban, and rural densities) on consumption and CO2estimates according to the univariate analysis of variance

Type of measures Dependent variable Adjusted R2 df F Significance

Per unit area Energy consumption 0.377 3 420.252 0.000Vehicle miles of travel 0.404 3 470.183 0.000CO2 emissions 0.418 3 498.143 0.000Vegetation CO2 assimilation 0.469 3 613.584 0.000CO2 balance 0.430 3 523.918 0.000

Per capita Energy consumption 0.023 3 17.436 0.000Vehicle miles of travel 0.001 3 1.645 0.177CO2 emissions 0.002 3 2.385 0.067Vegetation CO2 assimilation 0.352 3 376.186 0.000CO2 balance 0.353 3 378.564 0.000

Table 2. The per unit area household energy and fuel consumption and carbon emissions and sinks by settlement densities inFlorida, 2001

Settlement densities Urban Suburban Exurban Rural

Energy consumption (billion Btu/km2) 64.26 (1.37) 9.68 (0.22) 1.12 (0.04) 0.27 (0.02)Vehicle miles of travel (million miles/km2) 9.84 (0.19) 1.62 (0.05) 0.20 (0.01) 0.05 (0.00)CO2 emissions (thousand tons/km2) 7.62 (0.15) 1.21 (0.03) 0.15 (0.00) 0.03 (0.00)Vegetation CO2 assimilation (thousand tons/km2) 0.14 (0.00) 0.31 (0.00) 0.51 (0.01) 0.51 (0.03)CO2 balance (thousand tons/km2) 7.48 (0.15) 0.90 (0.03) –0.36 (0.02) –0.47 (0.03)

Note: Numbers in parentheses are standard error of mean.

Table 3. The per capita household energy and fuel consumption and carbon emissions and sinks by settlement densities inFlorida, 2001

Settlement densities Urban Suburban Exurban Rural

Energy consumption (million Btu/person) 36.91 (0.14) 36.53 (0.17) 35.85 (0.33) 32.36 (1.05)Vehicle miles of travel (miles/person) 5,862.85 (58.91) 5,865.64 (110.13) 6,314.17 (274.53) 5,608.84 (481.39)CO2 emissions (kg/person) 4,468.87 (28.05) 4,450.45 (49.66) 4,611.58 (122.45) 4,123.03 (234.11)Vegetation CO2 assimilation (kg/person) 1,192.00 (128.21) 2,057.51 (90.84) 19,408.97 (1,041.00) 91,689.51 (17,085.57)CO2 balance (kg/person) 4,340.66 (28.04) 2,392.94 (108.51) –14,797.40 (1,037.89) –87,566.48 (17,051.87)

Note: Numbers in parentheses are standard error of mean.

consumption and carbon emissions for effective emis-sions reductions.

Summary and Conclusions

In this study, we presented a new approach for spa-tially detailed carbon balance accounting and appliedthis approach to investigate the carbon balance be-tween household emissions and vegetation assimila-tion by Census tract for the state of Florida. House-hold carbon emissions in the presented Florida studywere composed of CO2 released through residential

consumption of energy and transportation fuels. Weexcluded carbon emissions from the commercial sec-tor for this inventory. The tract-level residential en-ergy and fuel consumption estimates were constructedbased on Census household characteristics and the U.S.Department of Energy household consumption sur-vey. Vegetation carbon sinks were measured as plantnet photosynthesis using biophysical remote sensingtechniques.

The analysis showed that Florida was able to self-assimilate its residential energy and transportation fuel-related carbon emissions (62 million tons) through thepresence of vegetation (88 million tons) according to



estimates in 2001. Household energy and fuel consump-tion, carbon emissions, and vegetation carbon sinks,however, vary significantly by Census tract across thestate. The consumption-based carbon emission sourcestend to be separated from vegetation carbon sinks. Con-sumption measures aggregated by settlement densitiesshowed that the urban and suburban densities (settledat � 0.1 housing units per acre) were associated with thehighest per capita energy consumption, whereas the percapita consumption of household transportation fuelstended to be the highest in exurban densities (settledat 0.025–0.1 housing units per acre). Both exurban andrural densities were associated with a significant amountof vegetation carbon sinks. The CO2 balance indicatesnet carbon sources occurring in the urban and subur-ban densities, with net sinks in the exurban and ruraldensities.

This study contributes a spatial inventory of en-ergy consumption and carbon emissions and sinks(Zimmerer 2011). The spatial inventory approach isused to provide baseline carbon fluxes estimates forestablishing emission targets, distributing emission re-sponsibilities, and establishing carbon trading programsassociated with specific geographic regions or locations.Our methods are broadly applicable and can be inte-grated with other socioeconomic metrics for sustainabil-ity assessments. For example, future efforts might focuson the energy and carbon consequences of various ur-ban development patterns, in particular, the location ofindividualized communities as well as the constituentdensities of residential and commercial development(Horner 2008). This will lead to new insights withrespect to sustainable low-carbon energy consumptionand transportation behavior.

Acknowledgment

Part of this research was sponsored by grants receivedfrom the Institute for Energy Systems, Economics, andSustainability (IESES) at Florida State University.

References

Andres, R. J., G. Marland, I. Fung, and E. Matthews. 1996.A 1 degrees × 1 degrees distribution of carbon diox-ide emissions from fossil fuel consumption and cementmanufacture, 1950–1990. Global Biogeochemical Cycles10 (3): 419–29.

Andrews, C. J. 2008. Greenhouse gas emissions along therural–urban gradient. Journal of Environmental Planningand Management 51 (6): 847–70.

Blasing, T. J., C. Broniak, and G. Marland. 2005. State-by-state carbon dioxide emissions from fossil fuel use inthe United States 1960–2000. Mitigation and AdaptationStrategies for Global Change 10:659–74.

Box, E. O., D. W. Crumpacker, and E. D. Hardin. 1993. Aclimatic model for location of plant species in Florida,U.S.A. Journal of Biogeography 20 (6): 629–44.

Brown, M., F. Southworth, and A. Sarzynski. 2008. Shrinkingthe carbon footprint of metropolitan America. Washington,DC: The Brookings Institution.

Brownstone, D., and T. F. Golob. 2009. The impact of res-idential density on vehicle usage and energy consump-tion. Journal of Urban Economics 65:91–98.

Chapin, F. S., P. Matson, and H. A. Mooney. 2002. Principlesof terrestrial ecosystem ecology. New York: Springer.

Churkina, G., D. G. Brown, and G. Keoleian. 2009. Carbonstored in human settlements: The conterminous UnitedStates. Global Change Biology 16 (1): 135–43.

Cramer, W., D. W. Kicklighter, A. Bondeau, B. Moore,G. Churkina, B. Nemry, A. Ruimy, and A. L. Schloss.1999. Comparing global models of terrestrial net primaryproductivity (NPP): Overview and key results. GlobalChange Biology 5 (Suppl. 1): 1–15.

DeLucia, E. H., J. E. Drake, R. B. Thomas, and M. Gonzalez-Meler. 2007. Forest carbon use efficiency: Is respirationa constant fraction of gross primary production? GlobalChange Biology 13 (6): 1157–67.

Druckman, A., and T. Jackson. 2008. Household energyconsumption in the UK: A highly geographically andsocio-economically disaggregated model. Energy Policy36:3177–92.

Energy Information Administration (EIA). 2001a. Total en-ergy consumption and expenditures by household mem-ber and demographics, 2001. http://www.eia.doe.gov/emeu/recs/recs2001/detailcetbls.html (last accessed 29July 2010).

———. 2001b. Total energy consumption in U.S. house-holds by four most populated states, 2001. http://www.eia.doe.gov/emeu/recs/recs2001/detailcetbls.html(last accessed 29 July 2010).

———. 2002. Updated state-level greenhouse gas emissioncoefficients for electricity generation 1998–2000. http://www.eia.doe.gov/pub/oiaf/1605/cdrom/pdf/e-supdoc.pdf(last accessed 29 July 2010).

———. 2009. Residential energy consumption survey(RECS) 2009 update. http://www.eia.doe.gov/emeu/recs/contents.html (last accessed 29 July 2010).

———. 2010. State energy profiles: Florida. http://tonto.eia.doe.gov/state/state energy profiles.cfm?sid=FL (lastaccessed 29 July 2010).

Ewing, R., and F. Rong. 2008. The impact of urban form onU.S. residential energy use. Housing Policy Debate 19 (1):1–30.

Gregg, J. S., L. M. Losey, R. J. Andres, T. J. Blasing, and G.Marland. 2009. The temporal and spatial distribution ofcarbon dioxide emissions from fossil fuel use in NorthAmerica. Journal of Applied Meteorology and Climatology48:2528–42.

Grimm, N. B., D. Foster, P. Groffman, J. M. Grove, C. S.Hopkinson, K. J. Nadelhoffer, D. E. Pataki, and D. P.C. Peters. 2008. The changing landscape: Ecosystem re-sponse to urbanization and pollution across climatic and



societal gradients. Frontiers in Ecology and the Environ-ment 6 (5): 264–72.

Gurney, K. R., D. L. Mendoza, Y. Zhou, M. L. Fischer, C. C.Miller, S. Greethakumar, and S. D. Du Can. 2009. Highresolution fossil fuel combustion CO2 emission fluxes forthe United States. Environmental Science & Technology43 (14): 5535–41.

Helm, D. 2008. Climate-change policy: Why has so littlebeen achieved? Oxford Review of Economic Policy 24 (2):211–38.

Hepburn, C., and N. Stern. 2008. A new global deal onclimate change. Oxford Review of Economic Policy 24 (2):259–79.

Homer, C., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Lar-son, N. Herold, A. McKerrow, J. N. VanDriel, and J.Wickham. 2007. Completion of the 2001 National LandCover Database for the conterminous United States.Photogrammetric Engineering and Remote Sensing 73 (4):337–41.

Horner, M. W. 2004. Spatial dimensions of urban commut-ing: A review of major issues and their implications forfuture geographic research. The Professional Geographer56 (2): 160–74.

———. 2008. “Optimal” accessibility landscapes? Develop-ment of a new methodology for simulating and assessingjobs–housing relationships in urban regions. Urban Stud-ies 45 (8): 1583–602.

Hu, P. S., T. Reuscher, R. L. Schmoyer, and S.-M.Chin. 2007. Transferring 2001 National HouseholdTravel Survey. Oak Ridge, TN: Oak Ridge NationalLaboratory.

Imhoff, M. L., L. Bounoua, R. DeFries, W. T. Lawrence, D.Stutzer, C. J. Tucker, and T. Ricketts. 2004. The con-sequences of urban land transformation on net primaryproductivity in the United States. Remote Sensing of En-vironment 89 (4): 434–43.

Intergovernmental Panel on Climate Change (IPCC). 2007.Climate 2007: Synthesis report. Geneva, Switzerland:IPCC.

Knight, C. G., S. Cutter, J. DeHart, A. Denny, D.Howard, S.-L. Kaktins, D. E. Kromm, S. E. White,and B. Yarnal. 2003. Reducing greenhouse gas emis-sions: Learning from local analogs. In Global changein local places: Estimating, understanding, and reduc-ing greenhouse gases, ed. Association of American Ge-ographers Global Change in Local Places ResearchGroup, 192–213. Cambridge, UK: Cambridge UniversityPress.

Lutzenhiser, L. 1993. Social and behavioral aspects of energyuse. Annual Review of Energy and the Environment 18:247–89.

Mitchell, K. E., D. Lohmann, P. R. Houser, E. F.Wood, J. C. Schaake, A. Robock, B. A. Cosgroveet al. 2004. The multi-institution North Ameri-can Land Data Assimilation System (NLDAS): Uti-lizing multiple GCIP products and partners in acontinental distributed hydrological modeling system.Journal of Geophysical Research—Atmospheres 109 (D7),D07S90. doi: 10.1029/2003JD003823.

Newton, P., S. Tucker, and M. Ambrose. 2000. Housing form,energy use and greenhouse gas emissions. In Achievingsustainable urban form, ed. K. Williams, E. Burton, andM. Jenks, 74–83. London: E & FN Spon.

Niemeier, D., G. Gould, A. Karner, M. Hixon, B. Bachmann,C. Okma, Z. Lang, and D. H. Del Valle. 2008. Rethink-ing downstream regulation: California’s opportunity toengage households in reducing greenhouse gases. EnergyPolicy 36:3436–47.

Oak Ridge National Laboratory. 2007. 2001 NHTStransferability national files. http://nhts-gis.ornl.gov/transferability/ (last accessed 29 July 2010).

Pataki, D. E., P. C. Emmi, C. B. Forster, J. I. Mills, E. R.Pardyjak, T. R. Peterson, J. D. Thompson, and E. Dudley-Murphy. 2009. An integrated approach to improving fos-sil fuel emissions scenarios with urban ecosystem studies.Ecological Complexity 6:1–14.

Petron, G., P. Tans, G. Frost, D. Chao, and M.Trainer. 2008. High-resolution emissions of CO2 frompower generation in the USA. Journal of Geophys-ical Research-Biogeosciences 113 (G4): G04008. doi:10.1029/2007JG000602.

Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. H. Nilon,R. V. Pouyat, W. C. Zipperer, and R. Costanza. 2001.Urban ecological systems: Linking terrestrial ecological,physical, and socioeconomic components of metropoli-tan areas. Annual Review of Ecology and Systematics32:127–57.

Rose, A., R. Neff, B. Yarnal, and H. Greenberg. 2005. Agreenhouse gas emissions inventory for Pennsylvania.Journal of the Air and Waste Management Association55:1122–33.

Running, S. W., R. R. Nemani, F. A. Heinsch, M. Zhao, M.Reeves, and H. Hashimoto. 2004. A continuous satellite-derived measure of global terrestrial primary production,BioScience 54:547–60.

Sovacool, B. K., and M. A. Brown. 2010. Twelve metropoli-tan carbon footprints: A preliminary comparative globalassessment. Energy Policy 38 (9): 4856–69.

State of Florida. 2007. Executive Order 07-127. Office of theGovernor.

Theobald, D. M. 2001. Land-use dynamics beyond the Amer-ican urban fringe. Geographical Review 91:544–64.

United Nations. 1998. Kyoto Protocol to the United NationsFramework Convention on Climate Change. New York:United Nations.

U.S. Census Bureau. 2009. Annual estimates of theresident population for the United States, regions,states, and Puerto Rico: April 1, 2000 to July 1, 2009(NST-EST2009–01). http://www.census.gov/popest/states/NST-ann-est.html (last accessed 29 July 2010).

U.S. Environmental Protection Agency. 2005. Emissionfacts: Greenhouse gas emissions from a typical pas-senger vehicle. EPA420-F-05-004. http://www.epa.gov/OMS/climate/420f05004.htm (last accessed 29 July2010).

———. 2009a. Inventory of U.S. greenhouse gas emis-sions and sinks: 1990–2007. http://www.epa.gov/climatechange/emissions/usinventoryreport.html (lastaccessed 29 July 2010).

———. 2009b. MOBILE6 vehicle emission modeling soft-ware. http://www.epa.gov/OMS/m6.htm (last accessed29 July 2010).

U.S. Geological Survey EROS Data Center. 2006. Theconterminous U.S. and Alaska weekly and biweeklyAVHRR composites. Sioux Falls, SD: EROS DataCenter.



Weber, C. L., and H. S. Matthews. 2008. Quantifying theglobal and distributional aspects of American householdcarbon footprint. Ecological Economics 66:379–91.

Wentz, E. A., P. Gober, R. C. Balling, and T. A. Day. 2002.Spatial patterns and determinants of winter atmosphericcarbon dioxide concentrations in an urban environment.Annals of the Association of American Geographers 92 (1):15–28.

Wise, M., K. Calvin, A. Thomson, L. Clarke, B. Bond-Lamberty, R. Sands, S. J. Smith, A. Janetos, and J. Ed-monds. 2009. Implications of limiting CO2 concentra-tions for land use and energy. Science 324:1183–86.

Yang, F., K. Ichii, M. A. White, H. Hashimoto, A. R.Michaelis, P. Votava, A. Zhu, A. Huete, S. W. Running,and R. R. Nemani. 2007. Developing a continental-scale

measure of gross primary production by combing MODISand AmeriFlux data through Support Vector Ma-chine approach. Remote Sensing of Environment 110:109–22.

Zhao, T. 2011. Impacts of urban growth on vegetation carbonsequestration. In Urban remote sensing: Monitoring, syn-thesis and modeling in the urban environment, ed. X. Yang,275–86. Hoboken, NJ: Wiley.

Zhao, T., D. G. Brown, and K. M. Bergen. 2007. Increasinggross primary production (GPP) in the urbanizing land-scapes of southeastern Michigan. Photogrammetric Engi-neering & Remote Sensing 73 (10): 1159–67.

Zimmerer, K. S. 2011. New geographies of energy: Intro-duction to the special issue. Annals of the Association ofAmerican Geographers 101 (4): 705–11.

Correspondence: Department of Geography, Florida State University, 113 Collegiate Loop, Tallahassee, FL 32306, e-mail: [email protected](Zhao); [email protected] (Horner); [email protected] (Sulik).


annals of the association of american geographers a...

Documents