Comparing city-scale greenhouse gas (GHG) emission accounting methods

Material Information

Comparing city-scale greenhouse gas (GHG) emission accounting methods implementation, approximations, and policy relevance
Chavez, Abel Antonio
Place of Publication:
Denver, CO
University of Colorado Denver
Publication Date:
Physical Description:
1 electronic file. : ;


Subjects / Keywords:
Greenhouse gas mitigation ( lcsh )
Greenhouse gas mitigation ( fast )
non-fiction ( marcgt )


More than 1,200 cities worldwide are embarking on low carbon goals. However, currently there are no protocols in place to consistently account for GHG emissions associated with cities. Thus, this thesis explores mathematical relationships, approximations, implementation challenges, and policy relevance for three city-scale GHG emission accounting methods: Purely Territorial, Trans-Boundary Infrastructure Supply-Chain Footprint (TBIF), and Consumption-Based Footprint (CBF). Mathematical relationships using Single-Region Input-Output, and Multi-Region Input-Output models showed that neither TBIF nor CBF provided a more holistic accounting of trans-boundary GHG. A typology of cities defined as: net-producers, net-consumers, and trade-balanced in terms of their GHGs embodied in trade is important for understanding the trans-boundary supply-chains serving cities. Data inputs for TBIF are found to be more robust and readily available, compared to CBF. A meta-analysis of 21 US cities showed that trans-boundary electricity generation, air travel, fuel refining, along with the production of food, cement, and iron & steel, may be well-suited for allocation to cities based on their use in city-wide residential-commercial-industrial activities in the TBIF method. Territorial GHGs captured as little as 37% of the total (in-boundary plus trans-boundary) footprint for net-consumer cities, and as large as 68% for net-producers. On average, TBIF captured 75% (n=2) of the total footprint for net-producers, 63% (n=11) for trade-balanced, and 62% (n=8) for net-consumer cities. In contrast, CBF captured an average of 35% (n=2), 57% (n=11), and 71% (n=8) of the total footprint for net-producers, trade-balanced, and net-consumer cities, respectively. Various metrics of GHG emissions computed for the three methods were assessed for their ability to appropriately compare cities'. For territorial GHG, neither GHGTerritorial/capita nor GHGTerritorial/GDP reflected urban efficiency of cities. For TBIF, GHGTBIF/GDP with only electricity allocated (R2=0.62), and GHGTBIF/GDP with the additional suitable infrastructures allocated (R2=0.77), correlated well with an urban efficiency index (UEI) composed of commercial-industrial production efficiency, household energy efficiency, and transportation system efficiency. However, GHGTBIF/capita showed poor correlation (R2=0.1) with the UEI as expected from a production-based account. In contrast, for CBF, GHGCBF/capita and GHGCBF/GDP showed an improved correlation (R2=0.4) with the UEI. However, GHGCBF/capita correlated more strongly (R2=0.76) with per capita expenditures. These data suggest that GHGTBIF/GDP is the appropriate metric for comparing cities based on their urban efficiency, and that GHGCBF/capita is appropriate for viewing cities from a consumption perspective. For the 21 cities modeled, GHGTBIF/GDP ranged from 154 mt-CO2e/GDP to 747 mt-CO2e/GDP, and GHGCBF/capita ranged from 15 mt-CO2e/cap to 32 mt-CO2e/capita. The TBIF was implemented in Delhi, India to explore issues of data availability and transferability of methods from the US to rapidly industrializing nations. Fieldwork showed sufficient availability and adaptability of TBIF methodology from the US to India yielding GHGTBIF equal to 948 mt-CO2e/GDP in Delhi vs. 413 mt-CO2e/GDP in Denver. Broad energy use metrics between Delhi and Denver help explain differences between the two cities. All GDP in this thesis represent 2008 real USD. Given that TBIF captured the majority of the total GHG footprint (62%-75%) in 21 cities in the meta-analysis, was well correlated with the urban efficiency performance of cities, and could be readily implemented in the US and internationally, this thesis finds TBIF to be well suited for international GHG protocols that aim to compare city-efficiencies.
Thesis (Ph.D.)--University of Colorado Denver. Civil engineering
Includes bibliographical references.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Abel Antonio Chavez.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
857463187 ( OCLC )


This item is only available as the following downloads:

Full Text


! COMPARING CITY SCALE GREENHOUSE GAS (GHG) EMISSION ACCOUNTING METHODS: IMPLEMENTATION, APPROXIMATIONS, AND POLICY RELEVANCE by Ab Ž l Antonio Ch ‡ vez Bachelor of Science, Mechanical Engineering, University of Colorado Denver 2002 Master of Business Administration University of Houston 2005 A thesis submitted to the University of Colorado Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Ph.D.) Civil Engineering 2012


"" This thesis for the Doctor of Philosophy degree by Ab Ž l Antonio Ch ‡ vez has been approved for the College of Engineering and Applied Sciences by Dr. Anu Ramaswami Chair and Advisor Dr. Bruce Janson Dr. Deborah Main Dr. Christopher Weible Dr. Jana Milford Dr. Jennifer Thorvaldson April 24, 2012 Date _____________________


""" Chavez, Abel A. (Ph.D., C ivil Engineering) Comparing City Scale Greenhouse Gas (GHG) Accounting Methods: Implementation, Appro ximations, and Policy Relevance Thesis directed by Professor Anu Ramaswami ABSTRACT More than 1 200 cities worldwide are embarking on low carbon goals. However, currently there are no protocols in place to consistently account for GHG emissions associated with cities. Thus, this thesis explores mathematical relationships, approximations, implementati on challenges, and policy relevance for three city scale GHG emission accounting methods: Purely Territorial, Trans Boundary Infrastructure Supply Chain Footprint (TBIF), and Consumption Based Footprint (CBF). Mathematical relationships using Single Region Input Output and Multi Region Input Output models showed that neither TBIF nor CBF provided a more holistic accounting of trans boundary GHG. A typology of cities defined as: net producers net consumers and trade balanced in terms of their GHGs embodie d in trade is important for understanding the trans boundary supply chains serving cities. Data inputs for TBIF are found to be more robust and readily available, compared to CBF. A meta analysis of 21 US cities showed that trans boundary e lectricity gener at ion, air travel, fuel refining, along with the production of food, c ement and iron & s teel may be well suited for allocation to cities based on their use in city wide residential comm ercial ind ustrial activities in the TBIF method. Territorial GHGs ca p tured as little as 37 % of the total (in boundary plus trans boundary) footprint for net co nsumer cities, and as large as 68 % for net producers. On average, TBIF captured 75 % (n=2) of the total footprint for net producers, 63 % (n=11) for trade balanced and 62 % (n =8) for net consumer cities. In contrast, CBF cap t ured an average of 35 % (n=2), 57% (n=11), and 71 % (n=8) of the total footprint for net producers, trade balanced, and net consumer cities, respectively. Various m et r i cs of GHG emissions computed for the three methods were assessed for their ability to appropriately compare cities For territorial GHG neither


"# GHG T erritorial /capita nor GHG T erritorial /GDP reflected urban efficiency of cities. For TBIF, GHG TBIF /GDP with on ly ele ctricity allocated (R 2 =0.62), and GHG TBIF /GDP with the additional suitable infras tructures allocated (R 2 =0.77), correlated well with an u r ban efficie ncy index (UEI) composed of comm ercial ind ustrial production efficiency, household energy effi ciency and transportation system efficiency. However, GHG TBIF /cap ita showed poor correlation (R 2 =0.1) with the UEI as ex pected f r o m a production based account In contrast, for CBF, GHG CBF /cap ita and GHG CBF /GDP showed an improved correlation (R 2 =0.4) with the UEI However, GHG CBF /capita correlated more strongly (R 2 =0.76 ) with per capita expenditures. These data suggest that GHG TBIF /GDP is the appropriate metric for comparing cit i e s based on their urban efficiency, and that GHG CBF /capita i s a ppropriate for viewi ng cities f r o m a consu mption perspective For the 21 cities modeled, GHG TBIF /GDP ranged from 154 mt CO 2 e/GDP to 747 mt CO 2 e/GDP, and GHG CBF /capita ranged from 15 mt CO 2 e/cap to 32 mt CO 2 e/capita. T he TBIF was implemented in Delhi, India to explore issues of data availability and transferability of methods from the US to rapidly industrializing nations Fieldwork showed sufficient availability and adaptability of TBIF methodology from th e US to India yielding GHG TBIF equal to 948 mt CO 2 e/ GDP in Delhi vs. 413 mt CO 2 e/GDP in Denver. B road energy use metrics between Delhi and Denver help explain differences between the two cities All GDP in this thesis represent 2008 real USD. Given that TBIF capture d the majority of the total GHG f ootprint (62% 75%) in 21 cities in the meta analysis, was well correlated with the urban efficiency performance of cities, and could be readily implemented in the US and internationally, this thesis finds TBIF to be well suited for international GHG protocols that aim to compare city efficiencies. The form a nd content of this abstract are approved. I recommend its publication. Approved: Dr. Anu Ramaswami


# DEDICATION To everyone w ith a desire to achieve The future is in your hands. "One of the greatest things you have in life is that no one has the authority to tell you what you want to be. You're the one who'll decide what you want to be. Respect yourself and respect the integrity of others as well." Jaime Alfonzo Escalante Gutier rez


#" ACKNOWLEDGEMENT Many individuals and organizations have made this research possible First I would like to acknowledge the US National Science Foundation's Interdisciplinary Graduate Education and Res earch Traineeship Grant (IGERT; Grant # DGE 0654378) for providing funding to carry out this research. The US NSF Alliances for Graduate Education and the Professoriate (AGEP) program for their support. My advisor Dr. Anu Ramaswami, thank you for pushing me beyond my limits it has made me bette r in every way. I thank you from the bottom of my heart for this transformational experience. Dr. Bruce Janson who spent a number of hours with me helping conceptualize the IO framework early on. Dr. Debbi Main whose expertise in community engagement has b een essential in working with cities. Dr. Chris Weible who has provided me with invaluable feedback and recommendations D r. Jana Milford whose feedback in assembling this research has been extremely helpful. Dr. Jenn ifer Thorvaldson for the answers to my countless questions regarding economic analysis and IO modeling. I would also like to thank the following organizations. MIG Inc. (IMPLAN) for their generous contribution of data. ICLEI USA for establishing a conduit to several US communities, and provid ing an active policy platform that has made this research relevant. ICLEI South Asia for their very warm hospitality and help in access ing data M r. Dwarakanth Nath and the Delhi Government for help ing me in data collecti on Dr. Muthukrishnan from the Delh i Airport who was instrumental in accessing all airport data. Each of the US cities, thank you for taking my calls and providing invaluable support. Lastly, and certainly not least, I thank my family. My wife Naomi who has never complained about her count less hours of work that allowed me to fulfill this research Thank you. C Ž sar and Na ’ ma who have provided me with a getaway from the grind. Pap‡ tus consejos me siguen guiando y te agradezco por todo lo que me has dado. Tal como me dec’as, el camino al Žxito siempre est ‡ bajo construcci—n. Sigues con nosotros. Madre, el apoyo incondicional, y tu Žtica de trabajo incomparable solo son dos de los cienes elementos claves que me han tra’do aqu’. Te quiero mucho.


! #"" TABLE OF CONTENTS Chapter $% &'()*+,-(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%% $ .% /&(0)1(,)0!)02&03!*4 -&(5 6 7-1/0!898!0:&77&*'7! 1--*,'(&'8!:0(9*+7 %%%%%%%%%%% ; .%$ & '()*+,-(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%% ; .%. ) 02&03!*4! &(5 6 7 -1/0! *::,'&(5 6 3 &+0! 898 0 :&77&*'7! : 017,)0:0'( %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%% < .%= 898 0 :&77&*'7! 1 --*,'(&'8!1'+! 4 **(>)&'(&'8! : 0(9*+7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % $$ .%=%$ 8 0*8)1>9&6 ? 170+! 1 --*,'(&'8 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% $= .%=%. ( )1'7 6 ? *,'+1)5! & '4)17(),-(,)0! 4 **(>)&'(! @(?&4A %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% $= .%=%= *'7,:>(&*' 6 ? 170+! 4 **(>)&'(! @-?4A %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % $B .%; >+1(0!*'! > )*(*-*/! + 020/*>:0'( %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% .. .%C *'-/,7&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%% .; =% :1(90:1(&-1/!)0/1(&* '79&>7!1'+!:0(9*+*/* 85!4*)!-*:>1)&'8!898 0:&77&*'!4**(>)&'(7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% .< =%$ & '()*+,-(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%% .< =%. 1 '! & '4)17(),-(,)0 6 ? 170+! 7 ,>>/5 6 91&'! 4 **(>)&'(!4*)! *::,'&(&07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % .B =%= : 1(90:1(&-1/! ) 0/1(&*'79 &>7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%% =. =%; &:>/1' &* ) 08&*'1/&D&'8! : 0(9*+!1'+! + 1(1! 7 *,)-07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % =E =%C : 0( 9*+*/*85! F &* ( 1?/0! 1/&?)1(&*'!1'+! 898 4 **(>)&'(! *:>,(1(&*'7 %%%%%%%%%%%%%%%%%%%%%%%%% ;$ =%C%$ : 0(9*+! F &* + 1(1! 1/&?)1(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% ;$ =%C%. : 0(9*+! F 898 4 **(>)&'(! *:>,(1(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%% ;< =%< + 1(1! 91//0'807 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%% ;G =%G ) 07,/(7!1'+! & '7&89(7!4)*:! ( 9)00! &(5! 1 '1/57&7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%% ;E =%E *'-/,7&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%% CC ;% 1'1/57&7!*4!.$!,7!-& (&07!1'+!&:>/&-1(&*' 7!4*)!:0()&-7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%% CG ;%$ & '()*+,-(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%% CG ;%. 170! &(&07!1'+! + 1(1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%% CB ;%.%$ & '&(&1/! + 1(170( %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%% CB ;%.%. &(5! 91)1-(0)&7(&-7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% <.


! #""" ;%.%= &* ( 1?/0! + 1(1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%% << ;%= : 0(9*+7!1'+! ) 07,/(7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%% G. ;%=%$ ) 1(&*'1/0!4*)! & '4)17(),-(,)0! 1 //*-1(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% G. ;%=%. &(5! ( 5>*/*85 H ) 0/1(&*'79&>7!(*! > *>,/1(&*' H 1'+! > )*I&07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%% GE ;%=%= ( )1'7 6 ? *,'+1)5! 7 ,>>/5 6 91&'7!*4! &(&07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%% E. ;%=%; *:>1)&'8! *20)180!*4! )?1'! : 0(1?*/&7:!?5! (?&4 J -?4 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%% EG ;%=%C : 0()&-7!4*)! *:>1)&'8! 898 7! 1 77*-&1(0+!3&(9! &(&07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% B. ;%; *'-/,7&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%% $K$ C% &'(0)'1(&*'1/!1>>/&1(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% $$G C %$ & '()*+,-(&*' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%% $$E C%. ( )1'7 6 ? *,'+1)5! & '4)17(),-(,)0! 4 **(>)&'(! @(?&4A : 0(9*+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%% $.. C%= 7 *-&* 6 0 -*'*:&-! > )*4&/0!1'+! 20)2&03!*4! 0 '0)85! 70! J 898 7!4*)! + 0/9& H & '+&1 %%%%%%%%%%%%%%%%% $.; C%; + 1(1! 7 *,)-07!1'+! ) 07,/(7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%% $.< C%;%$ + 0/9&! 7 (1(&7(&-1/! 9 1'+?**L %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%% $.< C%;%. + 0:*8)1>9&-7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%% $.G C%;%= ? ,&/+&'87! 0 '0)85! 70!1'+! & '+,7()&1/! > )*-077! 0 :&77&*'7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%% $.G C%;%=%$ ) 07&+0'(&1/! 0 '0)85! 70! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%% $.E C%;%=%. *::0)-&1/! 0 '0)85! 70! ? 0' -9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% $.B C%;%=%= & '+,7()&1/! 0 '0)85! 70! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% $.B C%;%=%; & '+,7()&1/! > )*-077! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% $.B C%;%=%C 0 :&77&*'7! 4 1-(*)7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% $=K C%;%; ( )1'7>*)(1(&*'! 0 '0)85! 70 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%% $=. C%;%;%$ 7 ,)41-0! ( )120/! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%% $=. C%;%;%. 1 &)! ( )120/! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%% $=; C%;%;%= ) 1&/! ( )120/! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%% $=C C%;%;%; 0 :&77&*'7! 4 1-(*)7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% $=< C%;%C 0 :?*+&0+! 0 '0)85!*4! : 1(0)&1/7! 70 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% $=G C%;%C%$ 0 :?*+&0+! 0 '0)85!*4! : 1(0)&1/7! ? 0'-9:1)L7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%% $=G C%;%C%. 0 :&77&*'7! 4 1-(*)7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% $=E C%C *'-/,7&*'7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% $=B <% -*'-/,7&*'H!-*'()&?, (&*'7H!4,(,)0!3*)L!J >)*(*-*/7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%% $;;


! "M <%$ *'-/,7&*'7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% $;; <%. *'()&?,(&*'7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%% $;C <%= 4 ,(,)0! 3 *)L! J > )*(*-*/7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%% $;< 1>>0'+&I!1N!0'0)85!, 70!+1(1!)0()&020+!4) *:!898!&'20'(*)&07H! 1'+!&:>/1' %% $;E 1>>0'+&I!?N!70-(*)!( )1+ 0!4*)!(9)00!-*::,'&( &07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% $C. 1>>0'+&I!-N!&:>/1'!7 0-(*)!7-90:0 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% $C= )040) 0'-07 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%% $<;


! M LIST OF FIGURES 4&8,)0!. 6 $N!()1'7 6 ?*,'+1)5!&'4)17(),-( ,)0!4**(>)&'(!@(?&4A 4*)!+0'20)H!-*/*)1+* % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% $C 4&8,)0!. 6 .N!(?&4!4**(>)&'(!4* )!+0'20)H!-*%!.KKC!1 '+!.KKG% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% $< 4&8,)0!. 6 =N!&//,7()1(&*'!*4!( ?&4!4*)!1'5!-*::,'&( 5% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%% $B 4&8,)0!. 6 ;N!&//,7()1(&*'!*4!?4!4*)!1'5!-*: :,'&(5% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% $B 4&8,)0!. 6 CN!+0'20)!-?4!898!0: &77&*'7%!>)0/&:&'1)5 )07,/(7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% .. 4&8,)0!. 6 >)*1-90 7!1'+!&-/0&!>)*(*-*/ % %%%%%%%%%%%%%%%%%%%%%%%% .= 4&8,)0!= 6 $N!7-90:1(&-!*4!(0)) &(*)&1/H!(?&4H!1'+!?4!1>>)*1-907%!0I>*) (!)0/1(0+! 1-(&2&(&07!1)0!79*3' &'!@8)15O791+0+A% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% =$ 4&8,)0!= 6 .N!?17&-!)0>)070'(1( &*'!*4!1'!0-*'*:5!(9 )*,89!0-*'*:&-!&'>,( 6 *,(>,(!@&*A% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% =E 4&8,)0!= 6 =N!7-90:1(&-!*4!?17& -!&*!(1?/0!&//,7()1( &'8!(90!0'0)85!70-(* )7!-1/&?)1(0+% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ;$ 4&8,)0!= 6 ;N!&//,7 ()1(&*'!*4!)0/1(&*'7 9&>7!4*)!898!0:&77&* '!1--*,'(&'8!+0)&20+ &'! (9&7!(907&7N!1A!)*,( (H!-*P!?A!+0'20)H!-* P!-A!71)17*(1H!4/% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% C$ 4&8,)0!= 6 CN! )*,((! F &//,7()1(&'8!898 ?4 1/*'8!3&(9 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%% C= 4&8,)0!= 6 !*4!.$!,7!-&(& 07!&'!(9&7!:0(1 6 1'1/57&7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%% 0!&'!: 0(1 6 1'1/57&7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%% <$ 4&8,)0!; 6 =N!-*::0)-&1/ 6 &'+,7()&1/!0/0-()&-& (5!,70!4*)!.$!-&(&07 +&7()&?,(0+!1-)*77! (90!,7!120)180% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%% <= 4&8,)0!; 6 ;N!-*::0)-&1/ 6 &'+,7()&1/!'1(,)1/!8 17!,70!4*)!.$!-&(&07 +&7()&?,(0+!1-)*77! (90!,7!120)180% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%% <= 4&8,)0!; 6 CN!-*::0)-&1/ 6 &'+,7()&1/!0'0)85!,7 0!4*)!.$!-&(&07!+&7( )&?,(0+!1-)*77!(90!, 7! 120)180% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% <;


! M" 4&8,)0!; 6 *)(1(&*'!75 7(0:!044&-&0'-5!4*)! .$!-&(&07!+&7()&?,(0 +!1-)*77!(90!,7! 120)180% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% << 4 &8,)0!; 6 $KN!0/0-()&-&(5! 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%%%%%%%%%%%%%%%%%% GC 4&8,)0!; 6 $$N!4**+!18)&-,/(,)0 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%% GC 4&8,)0!; 6 $.N!4,0/!)04&'&'8! 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%%%%%%%%%%%%% G< 4&8,)0!; 6 $=N!70)2&-07! 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % G< 4&8,)0!; 6 $;N!901/(9! 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% GG 4&8,)0!; 6 $CN!&)*'O7(00/! 6 8+>!27%!8987!4)*:!,7 0!@/04(A!1'+!0I>*)(! @)&89(A% %%%%%%%%%%%%%%%%%%%%%%%%%% GG 4&8,)0!; 6 $*/*85!+08)00!27 %!>*>,/1(&*'!4*)!.$! -*::,'&(&07%! /1)80)!-*::,'&(&07!1 )0!79*3'!(*!1>>)*1-9 ()1+0!?1/1'-0+% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%% EK 4&8,)0!; 6 $GN!-*))0/1(&*'!&'!*::,'&(5!(5>*/*85!:1 8'&(,+0!20)7,7!(*(1/ -*::0)-&1/ 6 &'+,7()&1/!0/0-()&-& (5!,70!>0)!-1>&(1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% E$ 4&8,)0!; 6 $EN!-*))0/1(&*'!&'!*::,'&(5!(5>*/*85!:1 8'&(,+0!20)7,7!-*::0 )-&1/ 6 &'+,7()&1/!0'0)85!,7 0!@&'!L?(,A!>0)!)07& +0'(&1/!0'0)85!,70!@ &'!L?(,A% %%%%%%%%%%%%%%%%%%%%%% E$ 4&8,)0!; 6 $BN!-*))0/1(&*'!&'!*::,'&(5!(5>*/*85!:1 8'&(,+0!20)7,7!0:>/* 5:0'(! &'(0'7&(5% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%% E$ 4&8,)0!; 6 .KN!'0( 6 >)*+,-0)7%!8987!)0/1 (&'8!(*!(0))&(*)&1/H &:>*)(H!1'+!0I>*)(!7 ,>>/5 6 -91&'7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%% E; 4&8,)0!; 6 .$N!()1+0 6 ?1/1'-0+%!8987!)0/1( &'8!(*!(0))&(*)&1/H! &:>*)(H!1'+!0I>*)( 7,>>/5 6 -91&'7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% EC 4&8,)0!; 6 ..N!'0( 6 -*'7,:0)7%!8987!)0/1 (&'8!(*!(0))&(*)&1/H &:>*)(H!1'+!0I>*)(! 7,>>/5 6 -91&'7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% E<


! M"" 4&8,)0!; 6 .=N!8987!-*20)0+!?5! (?&4!1'+!-?4H!)07>0(&20/5%!79*3'!17!(90 120)180!*4! -&(&07!?5!(5>*/*85% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%% EB 4&8,)0!; 6 .;N! '0( 6 >)*+,-0)! 6 Q!*4!(*(1/!4**(>)&'( 8987!-*20)0+!?5!&'4) 17(),-(,)0! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% EB 4&8,)0!; 6 .CN!()1+0 6 ?1/1'-0+! 6 Q!*4!(*(1/!4**(>)&'( 8987!-*20)0+!?5!&'4) 17(),-(,)0! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% BK 4&8,)0!; 6 .)&'( 8987!-*20)0+!?5!&'4) 17(),-(,)0! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% BK 4&8,)0!; 6 .GN!'0( 6 >)*+,-0)! 6 Q!*4!(*(1/!4**(>)&'( 8987!-*20)0+!?5! -*'7,:>(&*'! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% B$ 4&8,)0!; 6 .EN!()1+0 6 ?1/1'-0+! 6 Q!*4!(*(1/!4**(>)&'( 8987!-*20)0+!?5!-*'7 ,:>(&*'! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% B$ 4&8,)0!; 6 .BN!'0( 6 -*'7,:0)! 6 Q!*4!(*(1/!4**(>)&'( 8987!-*20)0+!?5!-*'7 ,:>(&*'! 7,>>/5 6 -91&'% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% B. 4&8,)0!; 6 =K N!:0()&-!R$! F -*))0/1(&*'!*4!(0))& (*)&1/!8987!27%!,0&% 1A!>0)!8+>H!1'+!?A!> 0)! -1>&(1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% BC 4&8,)0!; 6 =$N!:0()&-!R.! F -*))0/1(&*'!*4!7-*>0 $S.!8987!27%!,0&%!1A >0)!8+ >H!1'+!?A!>0)! -1>&(1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% B< 4&8,)0!; 6 =.N!:0()&-!R=! F -*))0/1(&*'!*4!7-*>0 $S.S=!8987!27%!,0&%! 1A!>0)!8+>H!1'+!?A!> 0)! -1>&(1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% BG 4&8,)0!; 6 ==N!:0()&-!R;! F -*))0/1(&*'!*4!(?&4! 8987!27%!,0&%!1A!>0) 8+>H!1'+!?A!>0)!-1>& (1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% BE 4&8,) 0!; 6 =;N!:0()&-!RC! F -*))0/1(&*'!*4!-*'7, :>(&*' 6 ?170+!8987!27%!,0&%! 1A!>0)!8+>H! 1'+!?A!>0)!-1>&(1% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%% BB 4&8,)0!; 6 =CN!-*))0/1(&*'!*4!8 98 -?4 O-1>!27%!4&'1/!-*'7, :>(&*'! 0I>0'+&(,)07!@TO-1>A % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% $KK 4&8,)0!C 6 $N!(?&4!4*)!+0/9&H!& '+&1H!.KKB!0I>1'+0+! 898!4**(>)&'(% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%% $;.


! M""" LIST OF TABLES (1?/0!$ 6 $N!,%7%!-0'7,7!+04&' &(&*'7!*4!,)?1'H!:0( )*>*/&(1'H!1'+!),)1/ @-0'7,7H!.KKGA% %%% $ (1?/0!. 6 $N!+&440)0'-07!&'!-& (5 6 7-1/0!898!0:&77&*'!4 **(>)&'(7!&'!>00) 6 )02&030+! /&(0)1(,)0% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% $. (1?/0 6 .N!()1'7 6 ?*,'+1)5!&'4)17(),-( ,)0!1-(&2&(&07!1--*, '(0+!4*)!?5!(90!(?&4 % %%%% $; (1?/0!. 6 =N!+0:*8)1>9&-!1'+!> 0)!>0)7*'!,70!()0'+7 &'!+0'20)H!-*%!1'',1 /!Q!-91'807! 1)0! -1/-,/1(0+!4)*:!.KKK 6 .KKG% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%% $G (1?/0!. 6 ;N!0I1:>/07!*4!-&(5 6 7-1/0!0'0)85!1'+!:1( 0)&1/!044&-&0'-5!:0( )&-7% %%%%%%%%%%%%%%%%%%%%% .C (1?/0!= 6 $N!-*::,'&(5 6 3&+0!898!0:&77&*'!7( ,+&07!&'!-&(&07!&'-* )>*)1(&'8! &'4)17(),-(,)0!7,>>/ 5 6 -91&'7!70)2&'8!(90!3 9*/0!-*::,'&(5% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%% .B (1?/0!= 6 .N!>*/&-5!)0/021'(!1 (()&?,(07!1'+!+08)00 *4!)0/021'-0!4*)!019!*4!(90! (9)00!898!1--*,'(&'8 :0(9*+7%! UVVV!)0>)070'(7!8)01 (07(!)0/021'-0P! W0I>/1'1(&*'7X!1)0!> )*2&+0+!4*)!)0+,-0+! )0/021'-0Y% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%% == (1?/0!= 6 =N!7,::1)5!(1?/0!*4! (90!+1(1!7*,)-07!,70 +!?5!&:>/1 '!4*)!-*'7(),-(&*'! +*3'7-1/0+!&*!(1?/07 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%% ;K (1?/0!= 6 ;N!0/0-()&-&(5!80'0) 1(&*'!4*)!(90!(9)00! -170 6 7(,+5!-&(&07%!1A!/*1/! 0/0-()&-&(5!80'0)1(& *'N!,' 6 -1/&?)1(0+!&:>/1'!27 %!08)&+P!?A!-1/&?)1( 0+!&:>/1'! 1'+!-*::,'&(5 6 3&+0!,70% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % ;= (1?/0!= 6 CN!,'1+Z,7(0+!1'+!1+ Z,7(0+!&:>/1'!4*)!/* -1//5!80'0)1(0+!0/0()&-&(5!,70!&'! +0'20)%!,70!?5!9*,70 9*/+7H!1'+!(*>!4&20! -*::0)-&1/ 6 &'+,7()&1/!,70)7% %%%%%%%%%%%%%%%%%% ;< (1?/0!= 6 /1'!1'+!-*:: ,'&(5!898!&'20'(*)5! )0>*)(7!4*)!(9)00!,7 -*::,'&(&07% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% C< (1?/0!; 6 $N!-&(&07!4)*:!*)&8& '1/!CC!0/&:&'1(0+!4) *:!:0(1 6 1'1/57&7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% 9&-!J!0-* '*:&-!7(1(&7(&-7!4*) (90!.$!,7!-&(&07!&'! (9&7!:0(1 6 1'1/57&7% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% <$


! M"# (1?/0!; 6 =N!7,::1)5!*4!:*7(!0 44&-&0'(H!1'+!:*7(!& '044&-&0'(!-&(&07!?5 0'0)85!,70! 044&-&0'-5!>1)1:0(0) % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%% <. (1?/0!; 6 ;N!(*>!(9)00!*,(>,(! ()1+0!70-(*)7!4*)!.$ ,7!-&(&07% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%% 0)!-1>&(1!8+>!1' +!&'-*:07!4)*:!&:>/1 '!1'+!?01H!4*)!.$!,7 -&(&07% %%%%%%%%%%%%%%%%%% GK (1?/0!; 6 0)!-1>&(1!8+>!1' +!&'-*:0!4*)!(90!$K! -*,'(&07!*4!(90!+0'2 0) 6 1,)*)1 6 ?)**:4&0/+!:71% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%% G$ (1?/0!; 6 GN!0I1:>/07!*4!&'+,7 ()&1/!>)*-07707!/*-1 (0+!&'!70/0-(0+!,7!&(&07% %%%%%%%%%%%%%%%%%%%%% G. (1?/0!; 6 EN!)08)077&*'!-*))0/ 1(&*'7!@) A!4)*:!,70!1'+!0I>*) (!4)*:!1//! 70-(*)7!4*)!.$! -&(&07H!1/*'8!3&(9!( 90!Q!-*'()&?,(&'8!(* '1(&*'1/!@,%7%A!8987 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%% G; (1?/0!; 6 BN!+08)00!?5!39&-9!*::,'&(&07!1)0!:017, )0+!17!'0( 6 >)*+,-0)H!1/*'8!3&(9 1/(0)'1(0!:0()&-7!4* )!)0>)070'(&'8!(5>*/ *85% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%% GB (1?/0!; 6 $KN!+08)00!?5!39&-9! -*::,'&(&07!1)0!:017 ,)0+!17!()1+0!?1/1'0+H!1/*'8! 3&(9!1/(0)'1(0!:0()& -7!4*)!)0>)070'(& '8!(5>*/*85% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% GB (1?/0!; 6 $$N!+08)00!?5!39&-9! -*::,'&(&07!1)0!:017 ,)0+!17!'0( 6 -*'7,:0)H!1/*'8!3&(9 1/(0)'1(0!:0()&-7!4* )!)0>)070'(&'8!(5>*/ *85% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%% GB (1?/0!; 6 $.N!8987!0:?*+&0+!&' 0I>*)(7%!-1/-,/1(0+! 27%!&:>/1'!(*(1/!0I> *)(!20-(*)% %% E= (1? /0!; 6 $=N!1(()&?,(07!1'+!, '&(7!*4!(90!,)?1'!04 4&-&0'-5!&'+0I!@,0&A % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%% B= (1?/0!; 6 $;N!:0()&-7!021/,1(0 +!4*)!-*:>1)&'8!8987 177*-&1(0+!3&(9!-&(& 07!181&'7(! (90!,0&% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%% B= (1?/0!; 6 $CN!,0&!4*)!.$!,7!-& (&07% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % B; (1?/0!; 6 $1-(7!*4!&'4)1 7(),(,)07H!?05*'+!0/0-() &-&(5H!912&'8!(90!:* 7(!&:>1-(! *'!,)?1'!044&-&0'-5% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%% $KK (1?/0!C 6 $N!?17&-! 1'+!0I>1'+0+!)0>*)(& '8!4)1:03*)L7!4*)!&/0& 6 ,71!-*::,'&(5 6 7-1/0! 898! 0:&77&*'7!1--*,'(&'8 1'+!)0>*)(&'8!>)*(**/% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% $.$ (1?/0!C 6 .N!(?&4!0'0)85!J!:1( 0)&1/7!,70!?0'-9:1)L 7H!1'+!047% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% $.. (1?/0!C 6 =N!+1(1!'00+7!4*)!?0 '-9:1)L&'8!(90!(?&4! (*!)0>)070'(!0'0)85! J!:1(0)&1/7! ,70% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%% $.;


! M# (1?/0!C 6 ;N!-*:>1)&7*'7!*4!L0 5 +0:*8)1>9&-!1'+!0-*' *:&-!21)&1?/07!&'!,7 1H!&'+&1H!1'+! +0/9&% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%% $.C (1?/0!C 6 CN!7,)41-0!()1'7>*)( 4,0/!,70!&'!+0/9&%!? 5!4,0/!(5>0!1'+!209& -/0!(5>0% %%%%%%%%%%% $=; (1?/0!C 6 *)(!7,)205!-*'+,-(0+ 1(!(90!+0/9&!&'(0)'1 (&*'1/! 1&)>*)(%!$$$!()120/0 )7N!C.!+*:07(&-!()12 0/0)7H!CB!&'(0)'1(&* '1/!()120/0)7% %% $=< (1?/0!C 6 GN!-*:>1)&7*'!*4!?)* 1+!0'0)85!J!:1(0)&1/ ,70H!1'+!+0:*8)1>9&! :0()&-7!4*)! +0/9&H!&'+&1!1'+!+0' 20)H!-*H!,71% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%% $;=


! 1 1. I ntroduction T he United Nations (UN) reported the world's population in 2007 at 6.4 billion people, a nd projects 8 billion people by 2025, and 9.2 billion by 2050. Also in 2007, for the first time in human history, more than half of the world's population was living in urban areas ( UN, 2007 ) According to the UN, between 2007 and 2050, the world's urban population will increase from 3.3 billion people, to 6.4 billion people. By 2025, 57% of the world's population will be located in urban areas, and by 2050, 70% of the world will be living in urban areas ; suggesting most of the forecasted population growth will occur in urban areas. Although the definitions of urban/metropolitan take various meanings around the world ( UN, 2008 ) the definitions as acknowledged by the US Census are presented in ([\]^! $ 6 $ Table 1 1 : U.S. Census definitions of urban, metropolitan, and rural ( Census, 2007 ) Type Population Criterion Density (peopl e/square mile) Urban Area 50,000 1,000 Metropolitan Area 100,000 1,000 Rural All those not classified as Urban Urban areas in the United States (US) house 80% of the total national population of 300 million people ( Census, 2007 ) This growing urban population is forecasted to reach 86% by 2025, and 90% in 2050 ( UN, 2007 ) The US South and West regions have had the highest rates of population growth, roughly 1.5 times the national average ( Census, 2010 ) The high urban population growth is expected to place much stres s and create great challenges in terms of infrastructure provisioning in the context of limited natural resources. Pri mary energy, water, food, and land, are some of the resources that urban dwellers will be competing for. C ities are also significant contributors to the global Greenhouse Gas (GHG) emissions. For example, it has been estimated that 70% of GHG emissions globally are attributed to energy use in cities ( IEA, 2008b ) Because cities are major con tributors to GHG emissions, cities can also be important in GHG emission mitigation ( Alberti & Susskind,


! 2 1996 ; N. Grimm et al., 2008 ) Treaties, such as the Kyoto protocol ( ICLEI, 2009 ) and the US mayor's climate protection agreement ( Mayors, 2010 ) must establish rigorous baseline GHG emission inventories that provide holistic GHG accounting at the city scale, and support regionally measurable GHG mitigation However t here are three core issues confound ing city scale GHG emission accounting : 1) The smaller spatial scale of cities causes artificial truncation of human activities such as commuter travel at the geograp hic boundary 2) Essential infrastructures serving urban needs such as electricity power plants airports, petroleum refineries, etc., cross city boundaries, hence termed trans boundary infrastructures 3) Beyond infrastructures, t here are significant exchanges (i.e. trade) of goods and services across city boundaries. Thus, GHG emissio ns associated with cities can broadly be classified as in boundary or trans boundary GHG emissions. The idea of scopes is one way of separating in boundary & trans boundary, as was introduced by the World Resources Institute ( WRI ) GHG protocol for corporat e GHG emissions ( WRI, 2004 ) Sco pes help define organizational boundaries for GHG emissions. The three scopes, and the ir definitions are: Scope 1 : in boundary, direct GHG emissions resulting from natural gas combustion, and tailpipe emissions from vehicle fuel combustion. Scope 2 : indire ct GHG emissions from purchased electricity used in the community Scope 3 : all other indirect GHG emissions linked to activities within the city While S copes 1 and 2 are required reporting, EPA & WRI recommend a small number of Scope 3 items to create win win supply chain GHG mitigation strategies ( EPA, 2007 ) But as will be shown in Chapter 2, there is considerable v ariability on how to allocate in boundary & trans boundary GHG emissions to cities in the form of GHG inventories & footprints.


! 3 At the national scale a significant amount of in boundary GHG emissions from c ommercial industrial activities are consumed within the country's boundary (e.g., the U.S.) O n the other hand, cities have unique specializations which lead to unique trans boundary flows For instance, ski resort communities have disproportionately higher commercial energy use result ing from the export of services. Suburban communities have higher residential energy use minimal commercial industrial energy use and higher imports of consumer goods and services. Larger metro cities provide both jobs and residence for many, and may be balanced ( combination of producers & consumers). The challenge then, is in developing a consistent GHG accounting framework for diverse city types. The overall goal of this dissertation is to compare the three leading city scale GHG emission accounting met hods side by side, evaluating implementation of methods, testing mathematical relationships that may allow for approximations, and identifying suitable metrics for increased policy relevance. This thesis is organized into four (4) additional chapte rs, and conclusion. Chapter 2 : Literature review and overview of city scale GHG emission accounting methods. Chapter 3 : Determine mathe matical relationships between the methods to facilitate approximations (simplifications), when appropriate based on city typolog y Chapter 4 : Meta analysis of 21 US cities to explore the nature and size of the ir trans boundary supply chains, and explore the relevance to metrics. Chapter 5 : Address translation of geographic based methods to rapidly industrializing countries, with specific attention to data availability in India Chapter 6 : Conclusions Contributions and Future Work


! 4 2. Literature Review of City Scale GHG Emissions Accounting Method s C ities are home to a large propor tion of the world's people, as a result they are being recognized as major contributors to global greenhouse gas (GHG) emissions. There is a need to establish baseline GHG emission accounting protocols that provide consistent, reproducible, comparable, and holistic GHG accounts that incorporate in boundary and trans boundary GHG impacts of urban activities and support policy intervention. This chapter provides a synthesis of previously published GHG accounts for cities by organizing them according to their in boundary and trans boundary inclusions, and reviewing three broad approaches that are emerging for city scale GHG emissions accounting: Geographic accounting, Trans Boundary Infrastructure Footprint (TBIF) and Consumption Based Footprint (CB F ) The TBI F and CBF footprints are two different approaches that result in different estimates of a community's GHG emissions, and inform policies differently, as illustrated with a case study of Denver, CO. The conceptual discussions around TBIF and CBF indicate that one single metric (e.g., GHG/person) will likely not be suitable to represent GHG emissions associated with cities, and it will take a combination of variables for defining a low carbon city 2.1 Introduction In 2007, for the first time in human history, more than half of the world's population was living in urban settings ( UN, 2007 ) According to the UN, the world's urban population is projected to increase f rom 3.3 billion people in 2007, to 6.4 billion people in 2050. By 2025, 57% of the world's population will be located in urban settings, and by 2050, 70% of the world will be living in urban settings. The US is also witnessing large rates of growth in urba n population, particularly in western states such as Colorado and Arizona. US metropolitan areas are home to 80% of the total national population of 300 million


! 5 people ( Census, 2007 ) The percentage of the US urban population is forecasted to reach 86% by 2025, and 90% in 2050 ( UN, 2007 ) Cities are home to a large proportion of the world's people, and as a result are being recognized as major contributors to the global Greenhouse Gas (GHG) emissions ( N. Grimm, et al., 2008 ; IEA, 2008b ) as well as a critical part of the solution, addressing both GHG mitigation and climate risk adaptation ( Alberti & Susskind, 1996 ) Cities worldwide have signed onto the Kyoto protocol, pledging to reduce greenhouse gases (GHG) by 7% by 2012 from 1990 baseline levels ( ICLEI, 2009 ) More recently, Mexico City mayor Marcelo Ebrard and ICLEI, convened 138 cities and signed the Global Cities Covenant on Climate otherwise known as the Mexico City Pact. The pact promises to have cities report on their respective GHG emissions and climate mitigation act ivities ( WMSC, 2010 ) In the US, one thousand forty four mayors' have also c ommitted their communities into some type of GHG mitigation ( Mayors, 20 10 ) However, such treaties must establish baseline GHG emission inventories that provide consistent, reproducible, comparable, and holistic (in boundary and trans boundary) GHG emission accounts with support for policy intervention. There are three primary challenges in holistic GHG accounting at the city scale that considers the full impact of urban activities on global GHG emissions. The three challenges are ( Hillman & Ramaswami, 2010 ; G. P. Peters, 2010 ; Ramaswami, Hillman, Janson, Reiner, & Thomas, 2008 ) : 1) Due to the relatively small spatial scale of cities, important human activities such as commuter travel and air travel, etc., are artificially truncated at the city's geographic boundary, 2) Cities are also served by trans boundary infrastructures such as electric power plants, oil refineries and pipelines, etc., that extend beyond city boundaries; 3) Last, beyond infrastructures, there are significant exchanges (i.e. trade) of goods and ser vices across boundaries.


! 6 Because of the above three trans boundary phenomena, it is being increasingly recognized that human activities in cities can stimulate emissions within their geographic boundary, as well as those outside, i.e., trans boundary GHG e missions. Thus, measuring only energy use and GHG emission strictly within a city's boundary can provide an incorrect and even misleading picture. In some cases establishing a purely geographic measurement approach may create unintended incentives to simpl y move GHG emissions outside the boundary. As society considers new technologies, design strategies and policies for low carbon cities, it is imperative that we have clearly define d methods for holistic measurement of GHG emissions associated with cities, addressing both in boundary and trans boundary emissions. In the past, the complexity of dealing with in boundary and trans boundary emissions has led to various inconsistent ways of GHG accounting at the city scale. The objective of this chapter is twofol d: To provide an overview of past literature on GHG accounting, listing the in boundary and trans bound ary inclusions. To highlight two leading theoretical and emerging approaches for GHG emissions footprinting at the city scale, incorporating trans boundary inclusions. Where other works discuss GHG footprinting methods in a general sense, the value added of this chapter is in the exemplification of GHG footprinting methods and results through actual data from a Denver, Colorado case study. We conclude by briefly discussing how ICLEI USA is incorporating these leading approaches into the community scale GHG emissions accounting and reporting protocol ( ICLEI, 2011 ) which is a framework being developed to help standardize GHG emi ssions accounting in US cities 2.2 Review of City Scale Community Wide GHG Emissions Measur e ment The lack of standardized method s for city scale GHG emissions accounting to date has produced inconsistent accounting approaches for cities throughout the world. This


! 7 inconsistency is seen both in the wide variation of inclusions in city scale GHG emissions accounting in the peer reviewed literature, and lack of explicit statements on what the unit of analysis is i.e., who is the accounting being done for? Is the unit of analysis, household consumption, community wide energy use, etc.? A first lack of clarity (or con fusion) arises between GHG measurements produced for city municipal governments and those that attempt to measure cities as a whole, i.e., whole communities that comprise a city. This issue is easily dealt with by referring to the city government emissions as Local Government Operations ( ICLEI, 2010 ) while the citywide analysis can be referred to as Community Wide GHG accounting Thi s chapter (and thesis) is solely concerned with community wide GHG accounting for cities; however, we mention this distinction to develop a consistent vocabulary in the literature going forward. Moreover, even the term community wide GHG accounting does not clearly address who the GHG accounting is being done for, i.e., what is the unit of analysis? Sometimes it is done for the "entire community" encompassed within a city's geopolitical boundary, i.e., residences, businesses and industries located within the geopolitical boundary, termed the geographical based approach in our review. At other times, GHG accounting appears to address primarily the consumption by households within a community a subset of a full consumption based footprint approach although many researchers are not explicit in such delineation in their papers. Lastly, when GHG accounting addresses economic final consumption (i.e. households government and capital expenditures within a community), this is termed full consumption based acco unting. So, what is a GHG footprint? Broadly speaking a "footprint" describes GHG emission of an activity beyond the boundary of the organization or entity for which the footprint is being computed ( Hillman & Ramaswami, 2010 ; G. P. Peters, 2010 ; Wright, Kemp, & Williams, 2011 ) Thus GHG emission footprints associated with cities seeks to measure and allocate the in boundary and trans boundary GHG emissions associated with cities in a manner that provides rigorous data and informs policy making. One way of describing in boundary & trans boundary GHG emissions, is through the idea of sco pes, developed by the World Resources Institute (WRI) for corporate GHG emissions reporting ( WRI,


! 8 2004 ) The concept of scopes helps define organizational boundaries for GHG emissions, and can be mapped to activities associated with cities as shown below: Scope 1 GHG emissions include all direct GHG emissions resulting from in boundary fossil fuel combustion (natural gas, fuel oil, gasoline, diesel, etc.), non energy industrial processes, and waste. Scope 2 indirect GHG emissions from imported electricity used within the community boundary Scope 3 all other indirect GHG emissions linked to supply chain lifecycle of materials and energy carriers used within the boundary that are produced outside. Note, in a consumption based approach, one must also subtract t he lifecycle GHG emissions from products that are produced within the boundary that are exported for consumption elsewhere, and can be shown as a Scope 3 subtractio n. Apart from helping define organizational boundaries, the concept of scopes also provides the means for preventing double counting of GHG emissions. For example, a city generating some of its total electricity use through a power plant located wit hin its boundary would count tho se GHGs as Scope 1 Meanwhile, power plant GHG from generating any additional imported electricity (not locally generated) used in the c ity would be c ounted as Scope 2, as these GHG are physically occurring outside of the city boundary Moreover, a city's GHGs resulting from the generation of surplus electricity (i.e., exports) should be allocated out based on demand to avoid double count ing both at the source and point of use In other words, the surplus electricity generated locally for use elsewhere should be subtracted from the Scope 1 of the gener ating city, and reported as Scope 2 for the using city If allocation of Scopes 1 and 2 are done correctly for all cities and regions in a country, the ir sum should total that country's territorial ( Scope 1 ) GHGs Similar issues arise when accounting for indirect GHG from other infrastructures (Scope 3). Scope 3 GHG may be added directly to a city's Scope 1+2 GHG, avoiding double counting with the c ommunity's GHG that may double count another community's Scope 1+2 GHG. This is why indir ect supply chain GHG s from infrastructures are shown as


! 9 Scope 3. In summary, i nclusion of indirect GHG emissions (Scopes 2 and 3) warrants careful allocation of GHG to avoid double counting ( Ramaswami, et al., 2008 ) Infrastructures such as large electric power plants, or oil refineries are easily recognized within city boundaries, and their GHG can be readily allocated based on local demand, thus reducing the potential for double counting. S copes 1 and 2 are required reporting for corporate accounting, though EPA & WRI recommend a small number of Scope 3 items to create win win supply chain GHG mitigation strategies ( EPA, 2007 ) However, cities are not like corporations, and there is considerable variation on how to allocate in boundary & trans boundary GHG emissions to cities. ([\]^! 6 $ shows peer reviewed studies that have accounted for various subsets of in boundary (Scopes 1+2) and trans boundary (Scope 3) GHG emissions relating to activities within cities. Note, although we follow typical nomenclature by showing Scope 2 as in boundar y, most electricity used in cities is generated outside, thus potentially allowing to be classified as trans boundary. Brown et al. (2009) inventoried GHG emissions for 100 US cities, and in their method, accounted for emissions resulting from in boundary residential electricity use and fossil fuel (cooking and heating) use, and fuel combustion in road transport & freight within each city. Neither commercial nor industrial activities within the boundary were included due to "complex processing issues", as s tated by the authors. Parshall et al. (2010) also considered multiple US cities, and sought to evaluate the GHG Vulcan data product and its ability to measure fossil fuel energy use in combustion in US urban areas. Due to Vulcan's focus on point of combust ion, emissions from direct energy use within a community are accounted for, but imported electricity are not, which is significant in most US cities. Thus both ( Bro wn, Southworth, & Sarzynski, 2009 ) and ( Parshall et al., 2010 ) provide a partial accounting of in boundary energy use and associated GHG emissions. In ( Sovacool & Brown, 2010 ) the authors inventoried geographic based GHG emissions of 12 international metropolitan areas. The study covered energy use in buildings (residential, commercial, industrial), road transport, agriculture within the boundary, and


! 10 waste accounting for almost all in boundary GHG emissions, non en ergy processes were not accounted however. No trans boundary activities were accounted. The city of Denver, Colorado is the first known city to have included trans boundary GHG emissions in their community wide GHG emissions estimates ( Greenprint, 2007 ) as was published and articulated by Ramaswami et al. (2008) The method accounted for all in boundary emissions, and included trans boundary emissions from airline travel, fuel refining, water/wastewater treatment, and production of cement & food for in boundary use. There are two known studies t o have accounted for GHG emission for Los Angeles ( Ngo & Pataki, 2008 ) and Chicago ( McGraw, Haas, Young, & Evans, 2010 ) respectively. Both accounted for a comprehensive set of in boundary emissions; the former included trans boundary emissions from food production and wastewater treatment, whereas the latter did not cover these categories, but did account for freight. In a study of ten global cities, Kennedy et al. (2009) inventoried GHG emissions from electricity, heating & industrial fuels, industrial processes, road transport, aviation, marine, and waste, in a method that fully accounted for in boundary GHG emissions. Lifecycle, upst ream emissions from refining the fuels used within each city were the trans boundary emissi ons considered. The authors cited the need to evaluate upstream GHG emissions from use of other critical materials in cities (e.g. food, buildings materials, etc.), which is now being addressed. Hillman and Ramaswami (2010 ) developed an approach that accou nted for in boundary GHG emissions, plus lifecycle emissions associated with key trans boundary infrastructures serving cities: water/wastewater pumping & treatment, fuel refining, and embodied emissions from cement & food production, and commuter, air, an d freight travel. Applying their method across eight US cities elucidated that the in boundary plus trans boundary accounting methodology provides a more holistic account of GHG emissions approaching national per person GHG emissions of 25 mt CO 2 e/cap for large US metro cities, with a presumed balance of carbon in remaining imports and exports. Very small cities with disproportionally low industrial activity were found to be outliers.


! 11 More recently, certain US states (e.g., Oregon ; ( Stanton et al., 2011 ) ) and cities (e.g., King County, WA ; ( Stanton et al., 2012 ) ) are embarking on full consumption based approaches for GHG emissions footprinting that tracks trade of goods and services in and out of cities, i.e., all imports and exports. Such approaches base d on economic IO have been used at national scales ( Peters & Hertwich, 2008 ) but city scale applications have been sparse due to challenges in downscaling IO data to the city level. ([\]^! 6 $ presents a summary of the above studies. 2.3 GHG Emissions Accounting and Footprinting Methods As seen in the literature review in ([\]^! 6 $ there are three emerging methods for city scale GHG emissions accounting. The three methods are Geographic boundary limited accounting, Trans Boundary Infrastructure Supply Chain Footprint (TBIF) and Consumption Based Footprint (CBF) This section discusses each of the three methods within the context of their theoretical origins, followed by their advantages & disadvantages. The discussion builds upon recent article s ( Wright, Coello, Kemp, & Williams, 2011 ; Wright, Kemp, et al., 2011 ) who describe advantages & disadvant ages of production & consumption based footprints, in general. In ( Wright, Coello, et al., 2011 ) its a cknowledge d that city scale footprint s are in their infancy, making this research a timely addition by covering the newer TBIF method (not previously covered in ( Wright, Coello, e t al., 2011 ) ) and providing city specific data as illustrative examples. We begin by discussing t he Geographic Based Accounting


! 12 Table 2 1 : Differences in city scale GHG emission footprints in peer reviewed literature.


! 13 2.3.1 Geographic Based Accounting Boundary limited geographic approaches to GHG emissions accounting are those used in national inventories, which are largely considered "production based", even though they include fuel combustion GHG s by final consumption (i.e., in homes, and personal vehicles, etc.). In other words, this me thod accounts for GHG emissions from all production activities within the nations geopolitical boundary, although direct GHG emissions from end use of energy in households are also included. These national GHG accounts are typically related to metrics of productivity, particularly Gross Domestic Product (GDP), and illustrated as GHG/$GDP ( EPA, 2010 ) Purely geographic based is not suited per se for reporting GHG/person; in order to truly represent an individual's impact on global GHG emissions, carbon embodied in trade to and from the country (imports/exports) must be included. For larger nations such as the US, about 90% of GHG emissions resulting from in boundary production are consumed within the boundary, and net import GHG emissions (imports less exports) are about 7% of the country's GHG emissions ( Peters & Hertwich, 2008 ) Therefore, strictly geographic and consumption based methods may be numerica lly similar for large countries or populations. However, strictly geographic approaches are not really suited for small cities because many of their infrastructures (e.g. transport networks, power plants, etc.), extend well beyond the city. For example, mo re than 60% of workers in Denver commute from other cities in the region ( R amaswami, et al., 2008 ) electricity transmissions can exceed 200 miles in the US ( Hirst, 2000 ) while freight travel averages 600 miles ( BTS, 2009 ) and US food travel averages 1,500 miles ( Weber & Matthews, 2010 ) 2.3.2 Trans Boundary Infrastructure Footprint (TBI F ) The Trans Boundary Infrastructure Supply Chain Footprint (TBIF ) is an innovative method developed by Ramaswami et al. (2008) which recognizes that cities are not like large natio ns, in that energy use to provide essential infrastructures like electricity, fuel, etc., often occurs outside the geographic boundary of the city The TBIF method therefore borrows from the concept of Scopes used in c orporate GHG accounting


! 14 (described previously in the introductory section), to account for essential trans boundary infrastructures serving cities. The method can be thought of as an infrastructure based supply chain footprint for cities, accounting for GHG emissions from buildings infrastr uctures (residential, commercial, and industrial) within the city (Scope 1) and trans boundary electric power supply, trans boundary transportation (road, air, and freight), fuel supply, water supply, waste management, and construction materials infrastruc tures serving cities (Scopes 2 & 3). See ([\]^! 6 Table 2 2 : Trans boundary infrastructure activities accounted for b y the TBI F Energy Use and Direct GHG Emissions within the Boundary (Scope 1) Energy Use in Various Trans boundary Infrastructure Sectors Serving Cities (Scopes 2 & 3) Buildings Infrastructure Direct Fuel Combustion and GHG Emissions From: Residential Buildings Commercial Buildings Industrial Facilities Energy Sector : Electric Power Production Fuel Refining Water and Waste Sector : Water Supply pumping Water & Wastewater treatment Landfill emissions from waste disposal Construction Sector : Cement Production Supply chain of other major materials to cities Transportation Sector : Air Travel Long Distance Freight Food Production Sample Results & Policy Impact: Results from applying the TBIF method to the City and County of Denver are shown in 4"_`a^! 6 $ TBIF results shown in 4"_`a^! 6 $ were obtained using bottom up end use of electricity and natural gas for buildings within the city, obtained from the local utility's billing data. Energy use in surface transportation was co mputed using regional vehicle miles traveled (VMT) across the commuter shed,


! 15 and allocated to Denver based on origin destination of trips. Emission Factors (EF) of energy carriers were consistent with IPCC. Often, material and energy flows (e.g., energy us e in air travel, cement use in city) were obtained from local data such as airports or economic census. EF s relating to the embodied energy of materials were obtained from regional scale LCA (for cement) and national EIO LCA (for food production), as discu ssed in Ramaswami et al. (2008) and Hillman and Ramaswami (2010) Results show GHG from the buildings sector corresponded to about 51% of which Scope 2 electricity related GHG emissions are 36%, while GHG emissions from surface transport tailpipe emissions were about 19%. The additional Scope 3 emissions (hatched) were attributed to trans boundary activities such as air travel, fuel processing, cement production, and food production. With these inclusions, Denver's GHG emissions footprint approached a broad er GHG footprint that is in line with the national average per person GHG emissions of 25 mt CO 2 e/person, suggesting the method is effective in capturing dominant trans boundary emissions associated with Denver. Similar convergence with national scale was seen in 6 other large US cities ( Hillman & Ramaswami, 2010 ) Figure 2 1 : Trans Boundary Infrastructure Footprint (TBIF) for Denver, Colorado.


! 16 The TBIF method has been shown to be highly policy relevant, resulting in innovative actions taken by the city ( Hillman & Ramaswami, 2010 ; Ramaswami, et al., 2008 ) In addition to focusing in on energy efficiency and conservation within the boundary, cities are now also able to focus on cross scale infrastructure efficiencies, e.g., related to water supply, regional transport, and the materials supply chain, etc. For example, Denver using the TBIF method, has implemented a green concrete policy aimed at reducing GHG embodied in concrete with the use of fly ash substitution for cement. Denver is also conducting a pilot project to evaluate conversion of food waste to energy. As a result of the TBIF method, cities such as Denver (in 2007) and more recently San Franc isco (in 2009) have developed voluntary travel offset programs at their airports ( Cabanatuan, 2008 ) Cross sector strategies such as tele presence that can displace airline travel are also particularly amenable for accounting in the TBIF method, wherein the trade off s between buildings energy use for tele confer encing programs can be shown to offset airline travel emissions. Lastly, as seen in 4"_`a^! 6 (a b) and in ([\]^! 6 = the TBIF method can be used in tracking GHG emissions over time, which further emphasizes the illustration of trade offs. Figure 2 2 : TBIF Footprint for Denver, CO. 2005 and 2007. !" !! #"$


! 17 Table 2 3 : Demographic and Per Person Use trends in Denver, CO. Annual % changes are calculated from 2000 2007. Demographic Trends Per Person Use Measure Annual % Change Data Source Measure Annual % Change Data Source Population + 0.95% U.S. Census Electricity + 0.9% Xcel Energy New Home Stock + 1.26% CCD Assessor Natural Gas 1.36% Xcel Energy New Comm. Area + 0.19% CCD Assessor Motor Gasoline 0.7% DOR Diesel + 1.2% DOR CCD: City and County of Denver; DOR: Colorado Department of Revenue. Advantages: The primary advantage of the TBIF method is that all activities within the city, residential, commercial, and industrial, are considered together, along with the trans boundary infrastructures critical for t hese activities. See 4"_`a^! 6 = Thus the method is relevant for city and region al planners who consider transport, power, water and materials supply in the region, as a whole. The manner in which the TBIF method addresses trans boundary infrastruct ures serving the entire community is illustrated schematically in 4"_`a^! 6 = Thus the advantages of the TBIF are concisely show n as ( Ramaswami, Chavez, Ewing Thiel, & Reeve, 2011 ) : Relevant to city and regional planning for whole communities considering residences, businesses and industries together. Well suited for showing impacts of infrastructure changes, linking local and regional actions. Cross sector strategies, such as teleconferencing are visible. Easy for public communication in that the major activities in home carbon calculators (e.g. airline travel etc.) are now also included in city accounting. The method yields sector specific benchmarks developed for each city, useful for comparing s ectoral efficiencies across cities.


! 18 The method is effective for tracking climate change impacts such as urban heat island effect that relate to direct in boundary Scope 1 fuel combustion Metrics pertaining to risk, vulnerability, and adaptation, can be qua ntified for both in boundary infrastructures (e.g., urban heat island) and trans boundary supply chain risks (e.g., risks to a city's electricity system due to climate water impacts). The method is particularly useful in linking local Scope 1 GHG emissions with local health impacts, e.g., increases in local ozone concentration ( Jacobson, 2010 ) and in potential future inclusions of short lived climate forcers (SLCF). As shown in figure 2, the TBIF method used locally specific da ta and is suitable for tracking a city's GHG emissions over time. Indeed with its capacity to address local health impacts of GHGs and SLCF, and provide input on supply chain vulnerabilities, the TBIF method is well suited to address both GHG emissions and climate adaptation in cities. Disadvantages: The primary shortcoming of this method is that it requires improved metrics for inter city comparisons on a consistent basis. Because the TBIF method is based upon geographic production based inventories, the often used per capita (same as per resident) metric is not appropriate for inter city comparisons using this method, particularly when a city with high industrial commercial activity is com pared with a solely residential community. GHG per unit gross regional product (or gross metropolitan product) is likely a better option, but for many smaller cities and towns such data are not reported. In such cases, normalizing community wide emissions by residents plus jobs could be an alternate for comparing cities. Per capita GHG emissions for this method may also be used if a typology of cities is created, representing producer consumer and energy balanced cities, such that cities are only compare d within their peer group. These are explored and discussed in a forthcoming chapter


! 19 Figure 2 3 : Illustration of TBIF for any community. Figure 2 4 : Illustration of CBF for any community. 2.3.3 Consumption Based Footprint ( CBF ) The consumption based approach accounts for global GHG emissions resulting from economic final consumption (households, government, and capital investments), within a city, including GHG emissions in imports, but excludes GHG from the production of


! 20 exports within the boundary. This method traces GHG emissions fully upstream, outside of the community boundary, accounting for all trans boundary activities that serve economic f inal consumption in the community. Of the economic final consumption sectors, households have been estimated to be responsible for the vast majority of the consumption ( Weber & Matthews 2008 ) (i.e. 80% final demand in the US), thus the method becomes well suited for evaluating household impacts on GHG emissions. A schematic showing activities in a typical consumption based applic ation can be seen in 4"_`a^! 6 ; Its been recognized there are two general approaches for conducting CBF for cities. The two CBF approaches are Consumer Expenditure Survey (CES) and Input Output (IO) The CES approach as sesses the impacts of household consumption, only linking purchases by households from a number of goods and services to economic sectors (i.e., North American Industry Classification System (NAICS)) While direct energy end use and associated emission factors are applied from local (territorial) data, GHG emissions from all other purchases apply national average production emission intensities regardless of production location ( Jones & Kammen, 2011 ; Weber & Matthews, 2008 ) Because CES' report on aggregate household consumption th e approach is not equipped to inform on the true location of production. The aggregate household GHG emissions are normalized per households or per capita yielding CBF from CES. The other approach to CBF is the economic input output (IO) approach. In this application national IO tables are downscaled (also call ed regionalized ) to counties, creating consumption profiles that address all c omponents of final consumption ( household s government and business investments). Such an approach, which is often referred to as a non survey approach for it s use of national statis t i cs, has been made commercially available by IMPLAN Inc for every US county. However, the accuracy of downscaled IO tables in representing material and energy flows in cities remains to be explored. Note that the IO approach for CBF has been adopted in t his thesis. Sample Results & Policy Impact: The preliminary GHG emission results from final consumption in Denver were computed using commercially available downscaled IO d ata


! 21 from IMPLAN. IMPLAN estimates monetary transactions and expenditures throughout the local economy across 440 economic sectors. The monetary expenditures for Denver were then converted to GHG emissions using a single region model and GHG emissions by economic sector (in mt CO 2 e/million$) from the EIO LCA tool ( CMU, 2008 ) T he total life cycle emissions associated with final consumption can be separated to show the contribution by scope, i.e., in boundary fuel combustion (Scope 1), electricity imports (Scope 2), other infrastructure imports denoted in TBIF (Scope 3) The other represents GHGs occurring while fulfilling all oth er goods and services to meet final consumption. Advantages: The primary policy relevance of this method is that it makes the full trans boundary impact of household consumption visible. Because most of the final consumption comes from households, this is theoretically the most rigorous method for comparing per person GHG emissions from household expenditures. Further, the method can help inform greening of government operation supply chain s With detailed and accurate IO data, imports and exports to/from a community can be traced. Overall an IO approach to life cycle assessment has been recognized as one producing fast holistic and mostly acceptable results A lthough the approach must be completed with caution as large variances may exist across scales ( Hendrickson, Horvath, Joshi, & Lave, 1998 ) Disadvantages: However, the full consumption based IO method is valuable only if accurate IO analysis and data are available at the city scale. Misallocations in local IO tables can occur when physical flows of energy and materials do not match the flow of economic activ ity; often occurring when large corporate headquarters in a city report economic activity well outside city boundaries. For many US cities, IO data are not published at a scale smaller than the county scale (now also available at the zip code level) U nlike the TBIF the CBF divides the community in two, with commercial industrial activities for exports not included in the unit of analysis. See 4"_`a^! 6 ; For some communities (e.g. resort towns & industrial towns), this excludes a sizable portion of their local economy that could be shaped by local policies. The application of IO tables for GHG emission accounting is new at the city level, and researchers are learn ing about its application to smaller spatial scales where downscaling national data poses challenges.


! 22 The difficulty of tracking GHG emissions via this method is triggered by the low publishing frequency of national IO tables; at every 5 7 years for the U. S. Figure 2 5 : Denver CBF GHG E missions Preliminary Results 2.4 Update on Protocol Development In summary and as seen in 4"_`a^! 6 = & 4"_`a^! 6 ; and in the discussions in this chapter the TBIF and CBF are two different approaches which give distinctly different estimates of a communities GHG emissions, and inform policies differently The TBIF method accounts for all in boundary emissions within the geographic boundary of a city, along with key trans boundary infrastructures serving the community as a whole The method is suited to future infrastructure planning that address es the who le community, and to address regional cross scale and cross infrastructure strategies across city boundaries. CBF GHG emissions accounts for all (in boundary and trans boundary) GHG emissions resulting from economic final consumption in the community, whil e the in boundary commercial industrial activities exported elsewhere along with their supply chains, are excluded, even though these local activities generate jobs and may also be shaped by local regulations. The method is especially suited to educate ho useholds about the global nature of their consumption.


! 23 Recognizing that both methods provide useful and different information, ICLEI USA has published a draft framework for community scale GHG emissions accounting and reporting ( ICLEI, 2011 ) The framework aims to help local US governments in planning and demonstrating GHG emissions reductions, by es tablishing standardized approaches for which communities can use to create holistic baseline GHG emissions measures. Because the protocol is in development, it is subject to future revisions. Recognizing that local governments have distinct reasons for mea suring GHG emissions, the protocol has varying tiers of reporting. The reporting approaches for community wide GHG emissions are Basic Reporting Standard (Basic), Expanded Community Impact Reporting (Expanded), and Consumption Based Reporting. The below sc hematic illustrates the ICLEI reporting framework, and how it links with the methodological approaches described in this chapter The basic reporting standard is expected to describe a minimum level of inclusions for co mmunity GHG emission accounting to es tablish consistency across cities. The expanded community impact reporting provides guidance on measuring energy use and GHG emissions more holistically, by incorporating all key trans boundary infrastr uctures as described in the TBIF method. Lastly, the p rotocol allows for an optional and separate accounting of GHG emissions f rom community final consumption using the CBF Figure 2 6 : Link between GHG accounting approaches and ICLEI protocol.


! 24 2.5 Conclusion Establishing a goal to develop low carbon cities requires good measurement tools for GHG accounting in cities. While it is obvious that a low carbon city must improve the energy efficiency of its buildings and transport system within the boundary, this c hapter asks what other trans boundary sectors are important in considering and defining a low carbon city? Many of the sectors that are trans boundary may in fact offer further and more innovative opportunities for GHG mitigation e.g., waste and indust rial symbiosis and innovative technologies such as tele presence. Changing the nature of consumption in communities, e.g., changing food diets, also becomes a part of the low carbon strategy toolkit. Recalling the adage What Gets Measured Gets Done mea surements tools play a major role in shaping the available strategy set, and vice versa. Recent advances in trans boundary GHG accounting ([\]^! 6 $ and the inclusions of such emerging knowledge into community wide GHG protocols being developed by ICLEI USA and others, is a major step in developing improved measurement tools. The discussion presented in this chapter show s that one single metric (e.g., GHG/person) will likely not be suitable to represent GHG emissions associated with cities. A combination of variables such as GHG per unit city residents plus city employees, or the totality of economic output may all serve a s potential metrics for defining a low carbon city. In addition to aggregate citywide metrics, such as GHG/person or GHG/GRP, sector specific efficiency and consumption measures are also useful. Hillman and Ramaswami (2010) have quantified efficiency and c onsumption measures in buildings, transport, and materials sectors in cities, at no additional cost or effort beyond TBIF ([\]^! 6 ; illustrates some of these efficie ncy metrics.


! 25 Table 2 4 : Examples of city scale energy and material efficiency metrics. Sector End Use Efficiency Metric Household Energy Use kWh/HH/mo, therms,/HH/mo, kBTU/HH/mo kWh/ cap /mo, therms,/ cap /mo, kBTU/ cap /mo Commercial Building Energy Use kWh/sq ft/mo, kBTU/sq ft/mo kWh/ GDP /yr kBTU/ GDP /yr Community Transport VMT/person/day VMT/ (resident s +job s ) /day Material Use tons MSW/capita/yr, mt cement/capita/yr Particular metrics will need to be ranked and weighted across cities. Efficiency benchmarks already existing in the literature ( Ramaswami, et al., 2008 ) could be expanded on. It is likely to take a combination of various metrics, together, to help define a low carbon city both for rigor and for policy relevance ( Zhou, Pric e, & Ohshita, 2010 ) Other sustainability metrics such as health and well being, Amartya's Sen's concepts of human capabilities approach reflected in the Human Development Index ( Anand & Sen, 2000 ) and emerging metrics of risk and vulnerability must also be considered in defining a low carbon goal.


! 26 3. Mathematical Relationships and Method ology for Comparing GHG Emission Footprints This chapter compares the policy relevance and derives math ematical relationships between three approaches for GHG emissions accounting associated with cities. The three approaches are: a) Purely Geographic Inventory, b) Trans Boundary Infrastructure Footprint (TBIF), and c) Consumption Based Footprint (CBF). In a case study of three U.S. communities (Denver Colorado, Routt Colorado, and Sarasota Florida), mathematical derivations coupled with data analysis show s that no one method provides a larger or more holistic estimate of GHG emissions associated with commun ities. A net producing community (Routt) demonstrates higher TBIF GHG emissions relative to the CBF, while a net consuming community (Sarasota) yields the opposite. Trade balanced communities (Denver) demonstrate similar numerical estimates of TBIF and CBF as predicted by the mathematical equations. Knowledge of community typology is important in understanding trans boundary GHG emission contributions 3.1 Intro duction Different types of greenhouse gas (GHG) emission footprints have been referenced in the li terature, often referred to in shorthand as "carbon footprints". Technically, carbon footprints address only carbon dioxide (CO 2 ) and methane (CH 4 ) emissions, while GHG footprints address the global warming potential of all six Kyoto GHGs (CO 2 CH 4 N 2 O, HFCs, PFCs, and SF 6 ), represented as CO 2 eq ( Wright, Kemp, et al., 2011 ) While these definitions are important, this chapter address es the larger issue of allocating GHG emi ssions to various segments of societies producers, consumers, nations and cities. The assignment of GHGs associated with the full life cycle of a product to a unit of production has been well understood in the industrial ecology literature, e.g., ( Eide, 2002 ) Recent efforts at the World Resources Institute (WRI) have incorporated life cycle approaches to inform GHGs reporting by corporations (producers) using the concept of scopes ( WRI, 2004 2011 )


! 27 Consumption based footprints (CBF) have also been articulated, wherein GHGs in commercial and industrial sectors are not assig ned to producers, but to economic final consumption represented by household expenditures, government expenditures, and business capital investments. At the national scale, GHG embodied in trade between nations has been assigned to final consumption activi ties in each nation, yielding CBF of residents in nations ( Peters & Hertwich, 2008 ) More recently, downscaled input output models at the scale of cities and states are being tested to develop CBF ( Stanton, et al., 2011 ) Final consumption is dominated (>80%) by household expenditures; hence a large number of CBF studies have developed GHG footprints of households using readily available consumer expenditure data and tracing the full life cycle GHG associated with th ese expenditures using EIO LCA, e.g., ( Jones & Kammen, 2011 ; Weber & Matthews, 2008 ) When nations report GHG emissions, however, territorial accounting is employed, i.e., direct GHGs within national boundaries are report ed in national GHG inventories, e.g., ( EPA, 2010 ) These territorial accounts are often re ferred to as production based accounts, but also include final household consumption of fuel (i.e., fuel combustion). Territorial accounts yield GHG intensity per unit productivity of nations, but are also reported on a per capita basis, although territori al GHG/capita does not reflect worldwide emissions associated with the residents of any nation. There is wide recognition that strict territorial accounting of GHGs employed in national scale GHG accounting is not meaningful for the s maller spatial scale o f cities, e.g., ( Kennedy et al., 2009 ; Ramaswami, et al., 2011 ; Ramaswami, et al., 2008 ) Cities are relatively small compared to nations, and also small compared to the larger scaled infrastructure systems in which they are embedded, e.g., transportation commuter sheds, and power supply networks. Consequently, important infrastructures ser ving cities that provision electricity, commuter travel, water supply, etc., are artificially truncated at the city's geographic boundary. Thus, GHGs from energy use in these key trans boundary infrastructures often occur outside the boundary of the city u sing these services (e.g., electricity used in a city is often generated outside of that city).


! 28 Several cities and associated research papers (see ([\]^! = 6 $ ) have sta rted incorporating the embodied energy in infrastructure supply chains serving the city as a whole, an approach that is formally being articulated in this chapter as the Trans Boundary Infrastructure Footprint (TBIF). TBIF studies have shown that energy us e in key trans boundary infrastructures serving cities can be as large, or larger than, the direct energy use & GHGs within city boundaries ( Hillman & Ramaswami, 2010 ; Kenne dy, et al., 2009 ; Ramaswami, et al., 2008 ) The TBIF supports citywide cross scale infrastructure pla nning for low carbon cities addressing infrastructure supply chains that serve both producers (e.g., industries) and consumers (e.g., households) that are co located in a community, provisioning infrastructures (e.g., energy supply, commuter travel, etc.) to the community as a whole. Through these trans boundary infrastructures, local and higher level governments are uniquely positioned to influence not only infrastructure related household activities, but also infrastructure related industrial commercial production activities (e.g., energy efficient offices) in a city that may subsequently export goods/services elsewhere. In contrast, CBF focuses more narrowly on city resident household and government consumption, examining their full supply chain impact s worldwide Increasingly, researchers are suggesting that both a TBIF and a parallel CBF be employed to inform a full spectrum of GHG mitigation strategies in cities ( Baynes, Lenzen, Steinberger, & Bai, 2011 ; Ramaswami, et al., 2011 ) However, the two footprint approaches are often considered to be entirely separate, when in fact, they are mathematically related in important ways. The refore the objectives of this chapter are to: Articulate the TBIF in the context of purely territorial and purely consumption based accounting, addressing the policy relevance of all three approaches. Elucidate mathematical relationships between the three methods, enabling approximations and simplifications between them, a s appropriate.


! 29 Table 3 1 : C ommunity wide GHG emission studies in cities incorporating infrastructure supply chains serving the whole community. Trans Boundary Infrastructures Serving Whole Community Researchers Electricity Water Fuel Cement Food Air Travel Freight ( Sovacool & Brown, 2010 ) Ramaswami et al. (2008) ( Ngo & Pataki, 2008 ) ( McGraw, et al., 2010 ) Kennedy et al. (2009) Hillman & Ramaswami (2010) Baynes et al. (2011) Chavez et al. (2012) ( Paris, 2009 ) ( Sharma, Dasgupta, & Mitra, 2002a ) 3.2 An Infrastructure Based Supply Chain Footprint for Communities Trans Boundary Infrastructure Footprints (GHG TBIF ) overcome the shortcomings of strictly boundary limited approaches by using the WRI concepts of scopes described earlier. A TBIF for cities reports direct community wide energy use and GHGs within city boundaries as Scope 1 emissions, GHGs from electricity generation for local use in all sectors (residential, commercial, ind ustrial) as Scope 2 emissions, while trans boundary life cycle emissions associated with other essential infrastructures serving the community are incorporated as Scope 3 emissions. Introduced by Ramaswami et al. (2008), TBIF quantifies Scope 3 GHGs from trans boundary commuter and airline travel, and from supply chains providing drinking water, wastewater treatment, transportation energy, food supply, and building construction materials in cities. Hillman & Ramaswami (2010) added impacts from long distan ce


! 30 freight infrastructure. Baynes et al. (2011), a nd Chavez et al. (2012) have quantified supply chains of electricity, water, fuel, cement, food, and air travel infrastructures serving the cities of Melbourne and Delhi, respectively. Several others incorp orated upstream GHG emissions from a smaller subset of infrastructures ( ([\]^! = 6 $ ). While studies have included different infrastructure supply chain inputs, articulat ing the method explicitly as a trans boundary infrastructure supply chain GHG emissions footprint for cities, while elucidating its policy relevance, helps clarify the method. The infrastructures covered by TBIF are widely accepted as essential for any city to functioning through provision of water, energy, food, transportation, waste treatment and b uilt environment materials (shelter). Developing trans boundary GHG emissions footprints associated with these key infrastructures enables multi level governance ( Betsill & Bulkeley, 2006 ) rangin g from the city scale (e.g., building codes) to the city region (e.g., mass transit) to the state and national scales that set standards for electric power generation, transportation fuel standards, etc. Although the status of food production as an infras tructure sector is fuzzy, cities are considering structural changes that formalize "green infrastructure" for urban food production ( J. Grimm, 2009 ) Moreover, food may also be viewed as another form of energy required by cities to be productive. Care must be taken to avoid double counting when incorporating GHG embodied in infrastructure supply chains. For example, supply chain GHG em bodied in gasoline would double count with any oil refineries operating within the city. Most infrastructures are large and visibly distinct (e.g., oil refineries), that their GHGs can be carefully allocated based on use/de mand. In the case of f ood production, TBIF as modeled in this thesis is unique compared to other approaches Most prior research has adopted the Bureau of Labor Statistics (BLS) Consumer Expenditure Survey (CES) ( BLS, 2012a ) as the source for the material flow analysis (MFA) of food consumed in a city. The CES reports economic expenditures for food consumption by homes, only, but does not reflect community wid e use. I n this research th e MFA of Food' represents true community wide use by homes as well as a ll local businesses obtained by trac k ing local & import interindustry flows and final consumption expenditures in IMPLAN ( sectors 1 14) Another important detail is that the


! 31 GHGs from use of food incorporated agriculture/livest ock portions, only (IMPLAN sectors 1 14) and not food manufacturing (IMPLAN sectors 43 69) such as the making of bread, beer, tortillas, etc. T he reason for this is to keep from double counting the agriculture/livestock portions within food manufacturing For example, supply chain GHGs from Beer production within a community would also include the agriculture GHG from wheat, hops, rice, etc A ny small (limited) agriculture within the boundary can be caref ully addressed avoiding double count ( Chavez, Ramaswami, Nath, Ranjan, & Kumar, 2012 ) Consumption Based GHG Footprints (GHG CBF ) go beyond allocating infrastructure, to allocate the trade of all goods and services across cities, however, focusing only on supply chains serving final consumpt ion (see 4"_`a^! = 6 $ ). Thus, local commercial industrial activities that produce goods and services for export elsewhere are allocated out, and excluded from the city's CBF. Figure 3 1 : Schematic of Territorial, TBIF and CBF approaches. Export related activities are shown in (gray/shaded).


! 32 Both TBIF and CBF provide different types of policy relevant information. TBIF is particularly relevant to future infrastructure planning across spatial scales. The potential for greening of infrastructures and supply chains, made visible by the TBIF, can be facilitated by multi level governance from the city to region to state an d national scales. For example, providing facilities for recharge/refueling of alternate fueled vehicles in cities r equires government facilitation at all levels and LCA based footprint computations to calculate net GHG benefits. TBIF is also very effectiv e in addressing multi scale risks that arise from fossil energy use by all sectors in a city homes, businesses and industry. These risks range from indoor air pollution from poorly ventilated stoves in homes, to local scale air pollution from traffic and industrial emissions, to regional haze and climate change induced risks to a city's coupled water energy system. In contrast, CBF conceptually provides the most holistic assessment of per capita GHG emissions that fully reflects an individual's impact on global GHG. CBF informs households and governments of the full impact from their consumption activities, which can promote shifts in consumption patterns, as well as encourage purchases from cleaner producing regions, i.e., greening the supply chain beyond the infrastructure sectors already addressed in TBIF. Because CBF excludes exported industrial commercial output and their supply chains (grey shaded areas in 4"_`a^! = 6 $ ) the stimulus to greening the supply chain is limited to households and governments. ([\]^! = 6 summa rizes the policy relevance of the purely geographic inventories, the TBIF, and the CBF. 3.3 Mathematical Relationships TBIF and CBF are often treated as completely separate methods, when in fact they are mathematically related. This section highlights mathem atical relationships between the two using a single region IO (SRIO) model for simplicity of illustrating the derivation, followed by a uni directional multi region IO (MRIO).


! 33 Table 3 2 : Policy relevant attributes and degree of relevance for each of the three GHG accounting methods {*** represents greatest relevance; [Explanations] are provided for reduced relevance} Desired Policy Relevant Attributes Utility of Greenhouse Gas Accoun ting Methods to Policy Attribute "###!$%&$%'%()'!*$%+)%')!$%,%-+(.%/!012&,+(+)34('5!+$%!&$4-36%6!74$! $%68.%6!$%,%-+(.%9 Purely Geographic Trans Boundary Infrastructure Footprint (TBIF) Consumption Based Footprint (CBF) Informs future city & regional infrastructure (multi level ) planning and policy [Most infrastructures transcend city boundaries] *** [Excludes infrastructures serving local businesses and industries that export goods.] Linkage of energy use to local urban heat islands, local air quality, and public health *** ** [Energy use in key infrastructures is allocated based on use, not location] [Energy use in all industries and businesses are allocated based on consumption, not location] Informs supply chain vulnerability for future p lanning [Most infrastructures transcend city boundaries] *** [Allocates GHG after consumption occurs, but does not address future planning for local supply vulnerability] Enables Inter city comparisons using per capita metrics to inform residents N/A [Per capita metric is incorrectly applied] N/A [Per capita metric is incorrectly applied] *** Enables Inter city comparisons using economic productivity metrics [Most infrastructures transcend city boundaries] *** N/A Data availability, quality and ability to benchmark or verify energy use and GHG emissions data ** [Remote sensing (e.g., Shepson et al., 2011) may enable independent verification] ** [IO models are calibrated to personal consumption and other data, not separately verifiable] SRIO Derivation The equation for computing consumption based GHG emissions, GHG CBF is ( Peters & Hertwich, 2008 ) :


! 34 where: F is the portion of local final consumption met by local production, and M F is the portion of local final consumption met by imports. The sum of F and M F yields total final consumption by households, government, and capital investments in the community. L is the Leontief Total Requirements Matrix ($ output/$ final demand) and in the SRIO model L is assumed to be equal to the national (U.S.) L matrix. B is the GHG intensity vector (mt CO 2 e/$ output) EF use is the use phase combustion emissions factor of fuels consumed by final consumption (e.g. natural gas, transport fuels). The pr oduction balance of an economy is written as: where: E are community exports M Z are imports to local industries used in meeting final demand, and Z are local interindustry transactions. Therefore, u pon substituting the term [L][F] from 0b`[c"de! = 6 into 0b`[c"de! = 6 $ and recognizing that total net imports, M net equals imports to local industry plu s to final consumption less exports ( M Z + M F E ), GHG CBF in (1) can be re written as: G H G C B F = B [ ] L [ ] + E F u s e # $ { } % F [ ] + M F [ ] { } L [ ] F [ ] + E [ ] { } = L [ ] M Z [ ] + Z [ ] + F [ ] + E [ ] = T L O [ ] + L [ ] M Z [ ] / "f^ 6 -gh]^O7`ii]g 6 -j["e!898! 0k"ll"del!&ec^el"cg! : ,l^! >j[l^!0k"ll"del!4[hcda ()!df!&kidacl! cd!/dh[]! &em`lca"^l (dc[]!4"e[]! -del`kic"de!"e! -dkk`e"cg (dc[]!)^b`"a^k^ecl! @()A!df!4"e[]!+^k[em (dc[]!/dh[]!*`ci`c! @(/*A Equation 3 1 Equation 3 2


! 35 G H G C B F = B [ ] T L O [ ] + E F u s e # $ % F [ ] + M F [ ] { } + B [ ] L [ ] M n e t i n f r a # $ + B [ ] L [ ] M n e t n o n & i n f r a # $ where: Term 1 [B][TLO] represent in boundary GHG emissions from direct energy use in commercial industrial production within the boundary serving final demand (includes exports). Term 2 [EF use ][F+M F ] captures use phase GHG emissions from final consumption, e.g., GHG from natural gas combustion by households. The sum of terms 1 and 2 yields GHG emissions from direct energy use, typically represented as S cope 1. Term 3 [B][L][ M n e t i n f r a ], quantifies the lifecycle emissions from key infrastructures serving cities, including Scope 2 (electricity) and Scope 3 (other infrastructures) GHG emissions. Term 4 [B][L][ M n e t n o n i n f r a ], quantifie s the lifecycle GHG emissions from all other, non infrastructure sectors. Note that M n e t i n f r a ( = M Z i n f r a + M F i n f r a E i n f r a ) represents net infrastructure imports (for electricity, natural gas and petroleum production, water/WW facilities, cement and food production agricu lture, air and freight transportation sectors), including imports to industry (M z ) and to Final Consumption sectors (M F ), less exports (E). Likewise net non infrastructure imports are represented as: M n e t n o n i n f r a ( = M Z n o n i n f r a + M F n o n i n f r a E n o n i n f r a ) Note, since infrastructures provide basic services to all communities, their GHG contributions are being allocated based on use, prior to evaluating the net productivity of cities. )^ia^l^ecl! (?&4!898!0k"ll"del!4ddcia"ec )^ia^l^ecl!8^d_a[ij"h!@(^aa"cda"[]A!898!0k"ll"del! &e#^ecdag 898!^k\dm"^m!"e!e^c!"kidacl!df!ede 6 "efa[lca`hc`a^l!cd!h"cg Equation 3 3


! 36 Note that : G H G T B I F = B [ ] T L O [ ] + E F u s e # $ % F [ ] + M F [ ] { } + B [ ] L [ ] M Z i n f r a + M F i n f r a & E i n f r a # $ Substituting 0b`[c"de! = 6 ; into 0b`[c"de! = 6 = yields the following relationship between GHG CBF and GHG TBIF G H G C B = G H G S c o p e s 1 + 2 + 3 T B I F + G H G M n e t n o n i n f r a ! 0b`[c"de! = 6 C implies that: In a trade balanced community, where G H G M n e t n o n i n f r a << GHG TBIF GHG CBF In a producer community where G H G M n e t n o n i n f r a is a large negative, GHG TBIF > GHG CBF In a consumer community where G H G M n e t n o n i n f r a is a large positive, GHG TBIF < GHG CBF Complementary sets of equations for the MRIO analysis are presented next MRIO Derivation We now derive mathematical relationships between GHG TBIF and GHG CBF using a uni directional MRIO model. A uni directional MRIO assumes that direct trade to local industries dominates. For details on uni directional MRIO, the reader is referred to ( Lenzen, Pade, & Munksgaard, 2004 ; Peters & Hertwi ch, 2008 ; Weber & Matthews, 2008 ) ). MRIO attempts to attribu te impacts to a particular region by considering a number of trade partners with different production characteristics (i.e., L matrix). For simplicity, we begin by writing MRIO GHG CBF using a two region model where Region 1 is the local community, and Reg ion 2 is the rest of world (ROW). G H G S c o p e s 1 + 2 + 3 T B I F Equation 3 4 Equation 3 5


! 37 G H G C B F = B [ ] L 1 [ ] F [ ] + B [ ] L 2 [ ] M F [ ] + E F u s e # $ F + M F [ ] where: L 1 = (I A 1 ) 1 = (I [A 11 +A 21 ]) 1 and is the full production matrix of the local/base economy, L 2 is the ROW production matrix in which following uni direction MRIO, is assumed equal to the national (US) production matrix ( L ). The production balance of an economy in the MRIO framework is written: where: A 11 are the direct requirements on local production, A 21 are the direct requirements on production of industrial imports from region 2 to 1, and x 1 is region's 1 output. Further, A 11 x 1 = Z, and A 21 x 1 equals th e total industrial imports into the local economy, region 1. Next we assume that all industrial imports into region 1 are exclusive of region 1 exports, and that A 21 x 1 [L 2 ][M Z ] Then, upon substituting [L 1 ][F] from 0b`[c"de! = 6 G into 0b`[c"de! = 6 < MRIO GHG CBF are shown as: G H G C B F = B [ ] T L O [ ] + E F u s e # $ F + M F [ ] + B [ ] L 2 [ ] M Z i n f r a + M F i n f r a # $ % B [ ] L 1 [ ] E i n f r a # $ { } + B [ ] L 2 [ ] M Z n o n % i n f r a + M F n o n % i n f r a # $ % B [ ] L 1 [ ] E n o n % i n f r a # $ { } where, [B][TLO]+[EF use ][F+M F ] + { [B][L 2 ][M infra ] [B][L 1 ][E infra ] } should approximate GHG TBIF and { [B][L 2 ][M non infra ] [B][L 1 ][E non infra ] } are the GHG embodied in net imports of non infrastructures to the city. These relationships can be directly related to those obtained from 0b`[c"de! = 6 C L 1 [ ] F [ ] + E [ ] { } = A 1 1 x 1 + A 2 1 x 1 + F + E = T L O + A 2 1 x 1 Equation 3 6 Equation 3 7 Equation 3 8


! 38 Figure 3 2 : Basic repre sentation of an economy through economic input output (IO) 3.4 IMPLAN IO Regionalizing Method and Data Sources T here are two general approaches for constructing economic IO tables ( 4"_`a^! = 6 ) 1) Survey (also known as Primary), and 2) Non Survey (also known as Secondary). Survey approaches collect data on economy wide transactions directly from businesses and other users within an economy. Even though survey approaches are thought to yield the most accurate representation of an economy, they are rarely completed due to the high level of resource requirements. An example of a survey approach is seen through the US benchmark IO table ( BEA, 2008 ) compiled by the Bureau of Economic Analysis (BEA) and p ublished every 7 years with a lag of the same (i.e., the benchmark IO table representing the US economy in 2007 will be released in 2014). On the other hand, non survey approaches rely on publicly available data collected by others. Using non survey approaches an economy can be modeled in relatively short time, and with substantially less res ources. IMPLAN IO tables are non survey based, and make use of several data


! 39 sources and techniques. The following summarizes the steps along with data sources used by IMPLAN in developing downscaled IO tables for US counties. Employment statistics used in IMPLAN are derived from three sources, US Department of Labors' Covered Employment and Wages (CEW) ( BLS, 2012b ) BEA Regional Economic Information System (REIS) ( BEA, 2012 ) and the US Census County Business Patterns (CBP) ( Census, 2012b ) CEW counts for t hose employees covered by unemployment insurance only, thus missing the self employed and exempt industries (e.g., railroad). As a result, REIS data is used for estimating this additional employment in sectors such as agriculture, construction, and railroa d, all of which are not subject to unemployment insurance. However, REIS data is only available at the semi aggregated (3 digit NAICS). CBP is used to estimate government employment from national statistics, where CBPs employment statistics are based on fi rst quarter employment. Note, employment data reported in IMPLAN are full and part time employees. Employment statistics are used by IMPLAN for estimating county level compensation, local outputs, and government & business capital expenditures. Value Added consists of employee compensation, property type income, and indirect business taxes, which IMPLAN estimates as follows. Employee compensation (wage and salary income) is estimated using state level income per employee ratios by sector (from REIS) multiplied by number of county employees in that same sector Other prope rty type income (OPTI) ( payments from interests), and Indirect business taxes (IBT) (sales taxe s) each are obtained from BEA's Gross State Product (GSP) ( BEA, 2012 ) for each sector S tate level OPTI/income and IBT/income ratios are multiplied with county income estimates for that same sector to compute county OPTI and IBT, respectively Total Industry Output (TIO) is computed using national data from the BEA ( BEA, 2011 ) National outputs are distributed to counties via national output per employee by sector multiplied with the local employment for a particular sector. Final Consumption (Households, Government Expenditures, and Capital Inves tments) are also gathered nationally, and distributed to counties as follows HH expenditures are estimated from the diary and survey of the Consumer Expenditure Survey (CES) ( BLS,


! 40 2012a ) and distributed to counties based on number of households and income. Government expenditures are obtained from the Federal Procurement Data Center (FPDC), and the Annual Survey of Governments (state governments only). In its default form t he FPDC is provided at the county scale while the ASG is compiled nationally and is distributed to counties based on employment. Business capital investments use BEA Wealth and the BEA Benchmark W orkfile. Because capital investments are generally closely linked to construction activity, these national data are distributed to counties using local construction employment Trade (Imports and Exports) is estimated in the following manner. Foreign trade nationally are obtained from the US Department of Commerce Foreign Trade Statistics ( USDOC, 2012 ) and distributed to counties through the ratio of local TIO to national TIO, by sector. Domestic trade between counties are estimated from IMPLANs National Trade Flow Model, which is a doubly constrained gravity mo del ( Lin dall, Olson, & Alward, 2005 ) ([\]^! = 6 = presents a summary of data sources and techniques used by IMPLAN for estimating downscaled IO tables Table 3 3 : Summary table of the data sources used by IMPLA N for construction downscaled IO tables. Input Output Table Element Data Source Technique for Local Distribution Used for Value Added Salary Income: REIS Other Income & Taxes: BEA GSP Income: from state level Other Income & Taxes: from state GSP Local GDP Industry Output BEA's output series Local sector employment Local TLO Household Consumption CES Local households and income Local household final consumption Government Expenditures FPDC, and ASG FPDC by county; ASG via employment Local government final consumption Business Capital BEA Wealth, and BEA Benchmark workfiles Construction employment Local business capital final consumption Trade Domestic: IMPLAN N ational Trade Model Foreign: USDOC IMPLANs trade model is a gravity model. National USDOC thru local industry output. Local Imports and Exports Employment CEW, REIS, CBP Local data retrieved from data sources. Derive TLO, Govt & Business Capital expenditures


! 41 3.5 Methodology IO Table Calibration and GHG Footprint Computations IO tables are n ot presently downscaled with the intent of being used for energy and GHG analysis, as being used here, and thus may be suscept ible to misallocations that mis represent energy use in communities sub nationally To identify such instances the following steps were implemented and where required, the respective IO tables were updated/ calibrated accordingly 4"_`a^! = 6 = is a high level schematic of an economy, showing the energy sec tors calibrated in our method C I is Commercial Industrial; and HH is Residential (or household) Figure 3 3 : Schematic of basic IO table illustrating the energy sector s calibrated. 3.5.1 Method IO Data Calibration Step 1 Building Energy Use : Th is step describes the calibration for community wide electricity and natural gas us e. Retrieving from IO : Building energy use reported in the default IO was obtained from the Z F M Z and M F data files. Electricity is repr esented by IMPLAN sector 3031 ; Natural


! 42 Gas is IMPLAN sector 3032. Building energy used by commercial industrial users is obtained from Z + M Z while residential from F + M F Monetary IO expenditures were converted to physical units using state average prices for electricity, and natural gas, by end user (res idential, comm ind) for the respective year ( EIA, 2011 ) For example, the Colorado state average price for electricity in 2007 was 9.3 cents/kWh, and 6.8 cents/kWh for residential, and commercial industrial, respectively ( EIA, 2011 ) Electricity, and natural gas use from the unadjusted IO tables were then compared to geographic building ener gy use obtained from local data, separated by community wide commercial industrial, and residential use Community wide uses rep orted in the IO data for each (electricity & natural gas) are forced to match the amount retrieved from geographic data by manually adjusting Z & M Z (commercial industrial), and F & M F (residential) in IMPLAN For natural gas we maintain the local/imported proportion defined by IMPLAN. For electricity however, we simultaneously adjust community wide use and t he locally generated amount of community wide use as data to do so is available through eGRID This is described in the following step. This thesis does not repeat the specifics for adjusting & regenerating IMPLAN IO tables, as they are described elsewhere ( MIG, 2004 ) Step 2 Loca lly Generated Electricity : The US EPA's Emissions & Generation Resource Integrated Database (eGRID) ( EPA, 2011 ) was used to calibrate for the amount of community wide electricity use that is locally generated Locally generated electricity is estimated from the ratio of Electricity Generation reported in eGRID, by total local electricity use retrieved from local energy use data. In equation form: % l o c a l l y g e n e l e c t r i c t y c o m m w i d e = e G R I D c o u n t y g e n t o t a l e l e c t r i c i t y u s e l o c a l d a t a Retrieving from IO : Default I O values for locally generated electricity use were obtained as shown in 0b`[c"de! = 6 $K for commercial industrial, and 0b`[c"de! = 6 $$ for residential L o c a l G e n C o m m I n d d e f a u l t = Z Z + M Z ( ) Equation 3 10 Equation 3 9


! 43 L o c a l G e n r e s d e f a u l t = F F + M F ( ) I n order to set the IMPLAN IO data to match the ratio computed from local data and eGRID ( 0b`[c"de! = 6 B ) both Z and F were manually adjusted (IMPLAN sector 3031 ) before regeneration. At the point the adjustments are made to community wide electricity use (step 1) and locally generated electri city (this step) the following balance holds. e G R I D c o u n t y g e n = Z + F + E ( ) 3 0 3 1 Thus, if the % locally generated electricity > 1 ( 0b`[c"de! = 6 B ) then exports (E) ( 0b`[c"de! = 6 $. ) are greater than 0, as the community produces surplus electricity. Similarly, if the % locally generated electricity <1, then exports (E) are equal to 0, as the community requires electricity imports to fulfill local use. ([\]^! = 6 ; illustrates electricity generation for the three cities in this case study (Routt, Denver, and Sarasota); where ([\]^! = 6 ; a compares the un calibrated IMPLAN with eGRID, and ([\]^! = 6 ; b shows the calibrated values in IMPLAN along with community wide electricity use for the three cities. Table 3 4 : E lectricity generation for th e th ree case study cities. a) Local electricity generation: Un calibrated IM PLAN vs. eGRID ; b) Calibrated IMPLAN and community wide use. +; ! Electricity Generation (in GWh) Routt, CO Denver, CO Sarasota, FL Z (local interindustry) 282 10,265 752 F (local final consumption ) 121 3,616 912 E (exports ) 877 5,415 7 Total Local Generation from Unadjusted IMPLAN 1,280 19,296 1,671 Total Local Generation from eGRID 3,654 1,269 0 % Error in local electricity generation between Unadjusted IMPLAN and eGRID 65% 1,421% Infinitesimal Equation 3 11 Equation 3 12


! 44 <; ! Note, similar databases for other infrastructures were not identified; therefore such analysis was carried out for electricity, only. Step 3 Road Transport Energy U se : This step describes the calibration for gasoline & diesel used in motorized road transportation. Retrieving from IO Data : Fuel used in motorized road transp ortation reported in the default IO data was obtained from the Z, F, M Z and M F data fi les. Petroleum Refining is represented by IMPLAN sector 3115 In this analysis it was assumed that all expenditures made from the petroleum refining sector were towards gasoline and diesel. Fuel used in road transportation by commercial industrial users is obtained from Z + M Z while residential from F + M F Fuel used in motorized transportation obtained from local data is allocated to users (residential, and commercial industrial) using national statistics ( DOE, 2009 ) which report that 96.8% and 3.2% of Gasoline is used by HH and non HH users, respectively. The same data set also shows that HH and non HH users use 7.5% and 92.5% of Diesel respectively. S t ate average gasoline and diesel prices ( EIA, 2011 ) are used for the volum e monetary unit conversion for comparing IMPLAN and local data Community wide uses of gasoline and diesel reported in the IO data are forced to match the amount retrieved from geographic data by manu ally adjusting Z & M Z and F & M F in Electricity Generation (in GWh) Routt, CO Denver, CO Sarasota, FL Z (local interindustry) 251 957 0 F (local final consumption ) 137 312 0 E (exports ) 3,265 0 0 eGRID 3,653 1,269 0 M Z ( imports to interindustry) 0 4,081 1,860 M F ( imports to final consumption ) 0 1,331 2,792 Total Community Wide Use a 388 6,681 4,652 a. Total Community Wide Electricity Use = [Z+F]+[M Z +M F ]


! 45 IMPLAN for the total expenditures (gasoline plus diesel) for each commercial industrial, and residential We maintain the local/imported proportion defined by IMPLAN. Detailed steps for adjusting & regenerating IMPLAN IO tables a re described elsewhere ( MIG, 2004 ) Step 4 Sectoral Output : Sector outputs as reported by IMPLAN can be in error due to 2 factors: 1) when large corporate headquarters are situated in a city, or 2) when self employed persons in a city operate energy assets (e.g., oil drills) located elsewhere The effects of such scenario s are seen in a community's "exports", thus giving the illusion of highly producing sector(s). To identify some of these erroneous sectors exports from each county were ranked by monetary value. Concentrations of top community sectors were validated throug h public information obtained from community web sites. For example, public sites confirmed economic specializations for Routt, CO ( mining /electricity / entertainment/recreation ). The physical condition f or others, such as no Oil & Gas Mining in Denver, crea ted a discrepancy with data reflected in IO tables. In such a case our approach zeroed the sector in question. In the future, better local data on local industrial outputs can yield more effective approach es to address these types of challenges. A ppendix B s hows top ten producing sectors, by exports, for t h e three c ities Data on community imports is not readily available, making imports difficult to verify Our approach thus relies on local knowledge stated above for identifying potential misallocations. Step 5 Adjusting IMPLAN Data File : After adjusting IMPLAN for the misallocations discussed in Steps 1 4, the respective IMPLAN file required regeneration in order to reconstruct all matrices R ecall that adjusted local uses of energy are visible in Z and F and adjusted import values would be made in M Z and M F Below is a sample of how sectors may be impacted after such adjustments. Here we show a portion of these impacts through an example of Denver's electricity use (sector 3031), as illustrated via loc al use by households as well as the top five comm ercial ind ustrial sectors. eGRID shows that 19% of Denver's electricity use is locally generated. After calibrating local electricity use ( Z & F ), it s noted that the amount used as a percent


! 46 remains constant (i.e., Real estate uses 21% of Denver's comm ercial ind ustrial local electricity). Table 3 5 : Unadjusted and Adjusted IMPLAN for locally generated e lectricity use in Denver. Use by h ouseholds, and top five commercial i ndustrial users. User Unadjusted Flow (million $) [91% locally generated] Adjusted Flow (million $) [19% locally generated] Household $335 $29 Co mmercial Industrial (top 5 ) $665 $62 Real estate buying and selling, leasing, managing, and related services $145 $13 Oil and natural gas $82 $8 Restaurant, bar, and drinking place services $41 $4 Wholesale trade distribution services $29 $3 Education from junior colleges, colleges, universities, and professional schools $20 $2 3.5.2 Method GHG Footprint Computation After addressing the required calibrations, and after the adjusted IO table had been regenerated, t he following methods were applied for estimating GHGs using the above SRIO equations. Step 1 Download IO Data : Final Consumption (F) is retrieved from ILCD (local), and M F from INSM. Imports to local businesses & industries (M Z ) are retrieved from INDM. Exports (E) from the city are obtained from ILCD. Each is in the form of a column vector, in commodity basis. Note, Institution Local Commodity Demand (ILCD), Institution Imports (INSM), and Industry Imports (INDM). Step 2 Convert to Industry Basis : As the GHG intensity vector (B) is derived from the US benchmark IO table (industry basis), IO data (step 1) is converte d from commodity basis (C) to industry basis (I) by multiplying each of the column vectors with the respective market shares matrix ( MSM) MSM represents the proportion of a c ommodity th at is produced by each industry, and is derived from the make matrix (IxC)


! 47 by dividing each row by the total commodity output. IMPLAN automatically calculates MSM for each model Step 3 Total Requirements : The column vectors now in industry basis, are diagonalized creating a square ( nxn ) matrix and multiplied by the total requirements matrix (L) Note, computing TLO is achieved by multiplying F and E each, by the respective city's l ocal L All others are multiplied by the national L. Both local and national L are retrieved from IMPLAN. This step yields total ou t puts in current year prices (i.e., 2008 model yields outputs in 2008 prices) Step 4 Price Adjustment : As B is built from the 2002 economy, total requirements (step 3) are price adjusted to 2002$ using sector specific prices ( BEA, 2009 ) M ultiplying the ratio of $ 2002 /$ current by the current year total re quirements computed from step 3 yields the price adjusted outputs. Step 5 Computing GHGs : Lastly, multiplying the price adjusted outputs with B yields GHG emissions attributed to each of the city's activities (e.g., F, M F M Z E). 3.6 Data Challenges Computing the CBF and TBIF for the three communities presented in this chapter using downscaled IO data revealed significant data chall enges in using IO tables D ownscaled IO tables are primarily used for economic development planning and are not specifically designed to match actual energy flows associated with electricity and fossil fuel use in local communities. Thus, several mismatche s between monetary and energy flows were observed, summarized in ([\]^! = 6 G Nationally downscaled home energ y use did not match locally observed data and had to be corrected with the locally obtained data ( Denver, 2010 ; Routt, 2010 ; Sarasota, 2008 ) Monetary energy purchases retrieved from IO tables were converted to physical units using state ave rage prices by end use sector ( EIA, 2011 ) As seen in ([\]^! = 6 G the percentage of local electricity use that is locally generated as projected by IMPLAN, significantly exceeded the local electricity generated based on eGRID ( EPA, 2011 ) in


! 48 two of the three communities. For electricity, this mismatch became visibl e because comparison with eGRID was possible. Thus, city scale IO model applications for GHG accounting that espouse the ability to highlight supply chains within versus outside the community may find that mismatches seen for electricity may also exist in other sectors, but remain unverifiable. Other mismatches between monetary and physical flows were also observed when large corporate headquarters are situated in a city, or when self employed persons in a city operate energy assets located elsewhere, as was found to be the case for Denver. For D enver, Oil & Natural Gas Sector (exports) and construction, generated exceedingly high economic activity caused by the entrepreneur residents and large headquarters located in the community, respectively. More collaboration with developers of IO models suc h as IMPLAN can help flag these mis matches and develop tools specific for city scale energy use and GHG analysis, as the IO models are not currently designed to represent energy/material flows. 3.7 Results and Insights from Three City Analysis The mathematical derivations ( Equations 3 1 thru 3 5 ) are tested for three US communities, Denver Colorado, Routt Colorado, and Sarasota Florida. Downscaled IO tables for these three communities were obtained from IMPLAN and calibrated with actual household en ergy use, transportation energy use and commercial industrial energy use reported in their respective GHG inventories ( Denver, 2010 ; Routt, 2010 ; Sarasota, 2008 ) The calibrated IO tabl es had to be further corrected, after which Equations 3 1 thru 3 5 were evaluated; results are shown in ([\]^! = 6 <


! 49 Table 3 6 : R esults for TBIF and CBF for three U.S. communities. County (Typology) GHG CBF (mt CO 2 e/cap) : [eq. 1]; { eq. 3 } GHG TBIF (mt CO 2 e/cap) [eq. 4] Numeric ratio: GHG TBIF of GHG CBF GHG non infra Mnet (mt CO 2 e/cap) Comm Ind Electricity use per capita (kWh/cap) Routt, CO (Net Producer) [32.2]; { 31.9 } 52 163% 20 Large negative (N et Producer ) 13,271 Denver, CO (Balanced) [31.6]; { 29.9 } 28 94% 2 Approaches zero (~Balanced) 8,704 Sarasota, FL (Net Consumer) [28.8]; { 29.7 } 22 74% 8 Larger positive ( Net Consumer) 5,123 U.S. Average 28 26 93% 2 (~Balanced) 7,704 GHG non infra Mnet are the GHG embodied in net imports of non infrastructures As expected, 0b`[c"de! = 6 $ and 0b`[c"de! = 6 = yield estimates of GHG CBF compute d in two different ways that are in line with each other for each of the three communities (column 2, ([\]^! = 6 < ). Moreover, GHG CBF is also similar across the three com munities, ranging from 29 mt CO 2 e/cap in Sarasota, to 32 mt CO 2 e/cap in Routt reflecting similar per household expenditures in the three communities. However, GHG TBIF (computed from 0b`[c"de! = 6 ; ) is vastly different across the communities, ranging from 22 mt CO 2 e/cap in Sarasota, to 52 mt CO 2 e/cap in Routt (column 3, ([\]^! = 6 < ) the latter containing a high proportion of commercial industrial activities engaged in exports. The ratio of GHG TBIF to GHG CBF (column 4, ([\]^! = 6 < ) shows that Routt (163%) has a much larger GHG TBIF relative to GHG CBF consistent with 0b`[c"de! = 6 C because Routt is a net producing community after essential infrastructures are evened out (column 5, ([\]^! = 6 < ) For Sarasota GHG TBIF

! 50 this metric may be a suitable proxy to represent net producing, net consuming, or GHG trade balanced communities. ([\]^! = 6 < shows that establishing a typology of communities as net producers, net consumers, and trade balanced in terms of GHG embodied in trade after allocating basic infrastructures is important in understanding the relative magnitude of the trans boundary GHG contributions in different types of cities. 4"_`a^! = 6 ; shows the relationship across city types, whe re: Routt, a net producing community reports GHG TBIF > GHG CBF Denver, a larger metro community, estimated to be roughly trade balanced reports GHG TBIF GHG CBF ; and Sarasota, a community dominated by residences (net consumer) reports GHG TBIF < GHG CBF Net Producing communities have higher territorial GHG emission, and are served by relatively smaller trans boundary supply chains. In contrast, highly consuming cities have smaller territorial GHG and much larger trans boundary GHG.


! 51 Figure 3 4 : I llustration of relationships for GHG emission accounting derived in this thesis : a) Routt, CO ; b) Denver, CO ; c) Sarasota FL.


! 5 2 Recall, Scope 1 are GHGs from end use of fossil fuels with the boundary ; Scope 2 are indirect GHG emissions from purchased electricity used in the community; and Scope 3 are all other indirect GHG emissions linked to the supply chain s for products used within the community Also recall that CBF includes GHGs occurring while provisioning of final consumption, only, and excludes activities relating to exports. Thus, it is expected th at in net consuming communities a large portion of local activities (energy use and GHGs) be fo r the fulfillment of final consumption. Meanwhile, in net producing and GHG trade balanced communities, where a larger portion of local activities fulfill exports, the opposite is expected lesser portions of local activities in support of final consumpti on and greater towards exports. The above is true among the three cities in this analysis. For Routt it is estimated that 22% of electricity, 33% of natural gas, and 40% of motor fuel used by local commercial industrial users is for serving final consumpti on. F or Denver it is estimated that 53% electricity, 63% natural gas, and 50% motor fuel used by local commercial industrial users is for serving final consumption. For Sarasota it is estimated that 63% of electricity, 76% of natural gas, and 85% of motor fuel used by local commercial industrial users is for serving final consumption. The following t hree figures highlight the differences described above. Each illustration shows GHG CBF by Scope, attributed to serving final consumption only, for each of the three communities.


! 53 4"_`a^! = 6 C illustrates GHG CBF for Routt (net producer) by consumption category. Note that GHGs shown are a result from serving Routt's final consumption, only. Here we compute Scopes 1+2+3 serving final consumption equal 17.9 mt CO 2 e/cap (or 55% of CBF). Figure 3 5 : Routt Illustrating GHG CBF along with GHGs by Scopes in serving final con sumption. 4"_`a^! = 6 < illustrates GHG CBF for Denver (trade balanced) by consumption category. GHGs shown are a result from serving Denver's final consumption, only. For Denver we compute Scopes 1+2+3 serving final consumption equal 16 mt CO 2 e/cap (or 53% of CBF). Figure 3 6 : Denver Illustrating GHG CBF along with GHGs by Scopes in serving final consumption.


! 54 /[lc]gH! 4"_`a^! = 6 G illustrates GHG CBF for Sarasota (net consumer) by consumption category. GHGs shown are a result from serving Sarasota's final consumption, only. For Sarasota we compute that Scopes 1+2+3 serv ing final consumption equal 20 mt CO 2 e/cap (or 70% of CBF). Figure 3 7 : Sarasota Illustrating GHG CBF along with GHGs by Scopes in serving final consumption. The perspective shown in 4"_`a^! = 6 C thru 4"_`a^! = 6 G reinforces some of the notable differences among community types. For example, lower activity in support of exports, and larger amounts of local energy used towards final consumption is highlighted in net consumer Sarasota.


! 55 3.8 Conclusion This preliminary case study of 3 cities suggests using caution in a pplying downscaled IO data to the city scale because current IO downscaling methods do not incorporate energy materials mapping/verification capabilities that are essential to show the percent local consumption that is being met by local production. Analysis of the IO tables, however, provided a useful side by side theoretical comparison of territorial, TBIF and CBF GHG emission footprints for three types of communities: for a net producing community (Routt), large metro community (Denver), and net co nsuming community (Sarasota). Along with the mathematical relationships Equations 3 5 the data analysis reinforces and offers the following insights: No one method provides a larger or "more holistic" account of GHG emissions associated with communities. For high net producing communities, which are net exporters of embodied GHG in trade after evening their supply chains of basic infrastructures, TBIF will yield a larger GHG footprint compared to the CBF. For high net consuming communities, TBIF will yield a lower GHG footprint compared to CBF. Most large metro areas are likely trade balanced communities (after essential infrastructures are evened out), wherein TBIF and CBF would estimate similar GHG footprints. For such communities, given that there are er rors and uncertainty in downscaling IO tables, for practical purposes, TBIF may provide a simplified approximation of CBF. Understanding the nature of communities as highly producing, highly net consuming and net Carbon trade balanced, after allocating out basic infrastructures, is essential for a more scientific understanding of their trans boundary impacts.


! 56 Table 3 7 : Differences between Unadjusted IMPLAN and Community GHG Inventory reports for three US communities. Electricity Use % community wide electricity use that is generated locally Unadjusted IMPLAN 1 Community GHG Inventory [state benchmark] 2 Unadjusted IMPLAN EPA eGRID 3 ( generation) Routt Residential Intensity 980 kWh/HH/mo 833 [554] kWh/HH/mo 98% 100% Total Commercial Industrial Use 287 GWh 251 GWh Denver Residential Intensity 1,284 kWh/HH/mo 546 [768] kWh/HH/mo 91% 19% Total Commercial Industrial Use 11,313 GWh 5,038 GWh Sarasota Residential Intensity 952 kWh/HH/mo 1,403 [1,367] kWh/HH/mo 43% 0% Total Commercial Industrial Use 1,730 GWh 1,861 GWh 1. Unadjusted IMPLAN data was retrieved from each of communities input output data file, provided by MIG, Inc. 2. Each of the three communities GHG Inventory Report as used to extract geographic energy use. 3. Local electricity generation retrieved from EPA eGRID (EPA, 2011).


! 57 4. Analysis of 21 US Cities and Implications for Metrics 4.1 Introduction In the previous chapter mathematical relationships were derived and used to estimate city GHG emissions using three methods: Territorial, Trans Boundary Infrastructure Footprint (TBIF), and Consumption Based Footprint (CBF), each with a unique treatment of trans boundary GHG emissions. While "Territorial" strictly measure s GHGs from sources within the city boundary (in the territory), TBIF also accounts for the GHGs embodied in net imports of infrastructures. CBF goes further to account for GHGs embodied in net imports of non infrastructures. Ramaswami et al. (2008) have a rticulated that TBIF is akin to production based GHG accounting with key infrastructures that are hypothesized to be essential for economic production, allocated across cities based on their "use". Thus the metric for comparing cities using TBIF is expecte d to be GHG/GDP. These key infrastructures (water, energy, food, transportation) have previously been hypothesized as being basic services supporting every city ( Hillman & Ramaswami, 2010 ; Ramaswami, et al., 2008 ) Their use in residential commercial industrial sectors in a city is added to territo rial as Scope 2 (electricity), or Scope 3 (other essential infrastructures), while surpluses being exported are subtracted. The allocations of these infrastructures were based on use, and on practical knowledge; this chapter explores the rationale for this approach, using meta data for 21 cities. The allocation of infrastructures (presented in chapter 3) yielded a 3 way typology of cities based on the GHGs embodied in net imports of non infrastructures. Where the net imports of non infrastructures are large then the city is said to be net consuming, but if the city is a large net exporter (or large negative in net imports) of non infrastructures, then it is said to be a net producer. The relative size of non infrastructure net imports are compared to TBIF, and if within +/ 15% the city is GHG trade balanced, >15% net consumer, and < 15% net producer. But why is a typology for cities important?


! 58 Beyond classifying cities, a typology for cities is important to understand the size of the trans boundary supply chains serving cities, enabling insight into the environmental impacts beyond the boundaries stimulated by in city activities. The nature (i.e., what is the make up of the supply chain) along with size of the supply chains are important to better un derstand GHG emissions associated with cities. For instance, chapter 3 showed that a net producer has lower embodied GHGs in their supply chains (imports) relative to in boundary GHG. This is in contrast to a net consumer city that has small in boundary GH G relative to that embodied in imports. For the trade balanced city, the GHG embodied in imports and exports (after allocating infrastructures) is about equal. But do these patterns hold for most cities within typologies? This chapter explores the nature a nd size of trans boundary supply chains for 21 US cities classifying them by typology. Additionally, other factors that shape typology, such as city population, and economy, will be explored to enhance our understanding of city types. We also explore suita ble metrics for comparing cities, and GHG emissions associated with them. The common metric presenting a city's GHGs has largely been per capita. Hillman & Ramaswami (2010) have indicated that GHGs per capita (i.e., GHG/resident) is not appropriate as seen in many studies. Cities with minimal commercial industrial activities will artificially appear more efficient, while highly producing cities with larger amounts of industries will appear as less efficient (e.g., see Fig 3 in Hillman & Ramaswami). Ramaswam i et al. (2011) propose GHG/GDP or GHG/job is appropriate for TBIF, while GHG/cap is suited for CBF. The availability for the first time of robust data for 21 cities can help identify best metrics to compare cities by. Each of the above core topics are ex plored through a meta analysis of 21 US cities. The cities are technically defined as US counties ( Census, 2012a ) but are called cities in short hand. The specific objectives of this chapter are: 1. Examine and articulate a rationale for allocating speci fic infrastructures based on "use" in a city's residential commercial industrial sectors. 2. Evaluate relationships between city typology (net producers, net consumers, and balanced) and city population size along with other proxies, to explore if larger citi es tend to be balanced.


! 59 3. Compute the size of trans boundary supply chain footprints serving cities, by typology, relative to in boundary GHGs. 4. Explore suitable metrics for comparing GHG emissions associated with cities that appropriately reflect urban efficiency characteristics. We first discuss the meta study dataset after which methods and results addressing each of the objectives are presented as: a) Infrastructure rationale, b) Typology, c) Trans Boundary Supply Chains, and d) Metrics. 4.2 Case C ities and Data 4.2.1 Initial Dataset A dataset on 55 citi es was provided by ICLEI representing all ICLEI members that had completed GHG inventories by 2010. The cities in the dataset needed to be designated as counties for which Input Output (IO) tables could be obtained. The dataset was first audited for cities that were not counties, of which 13 were identified, reducing the list to 42. Immediate local energy data required for the meta analysis were residential commercial industrial building energy use, and motorized transportation energy use (gasoline/diesel), for which was unattainable for 5 cities, thus reducing the list to 37 cities. Lastly, 16 counties did not assign a representative to assist in acquiring local knowledge, required for contextual u nderstanding of IO data. The dataset was reduced to 21 cities, all of which energy use data was benchmarked with comparable state data (see Supplemental Inform ation (SI) in back of chapter).


! 60 Table 4 1 : C itie s from original 55 eliminated from meta analysis. City/County Reason for not including New Haven, CT Missoula, MT Urban classification : City rather than County Blaine County, ID Providence, RI Rock Island, IL Franklin, TN New Orleans, LA Dallas, TX Nantucket, MA Richmond, VA Worcester, MA Spokane, WA Baltimore, MD Sonoma County, CA Alachua County, FL Missing Data : Building energy use, or Transportation Marin County, CA Leon County, FL Alameda County, CA Fairbanks North Star, AK Chatham County, NC No county representative for local knowledge San Luis Obispo, CA Orange County, NC Los Angeles County, CA Clackamas County, OR San Mateo County, CA Montgomery County, PA La Plata County, CO Williamson County, TX Montgomery County, MD Albemarle County, VA Queen Anne's County, MD Skagit County, WA Hennepin County, MN Whatcom County, WA Figure 4 1 : Map of 21 US cities in this meta analysis


! 61 Table 4 2 : Demographic & economic statistics for the 21 US cities in this meta analysis. County /Region Population a Population Density (people/sq mile) a GDP (mill $) b Jobs a NYC, NY 8,363,710 27,012 $658,701 3,679,345 DVRPC, PA/NJ 5,499,482 1,477 $300,985 2,548,018 Miami Dade, FL 2,387,170 1,240 $109,939 998,241 Broward, FL 1,759,591 1,449 $74,702 751,629 Oregon METRO, OR 1,600,751 517 $84,866 831,966 Philadelphia, PA 1,448,394 11,254 $70,474 632,755 Sacramento, CA 1,374,724 1,403 $63,517 624,259 Westchester, NY 949,355 2,190 $57,012 411,005 Multnomah, OR 714,567 1,637 $47,457 449,358 Snohomish, WA 669,887 318 $21,656 221,050 Denver, CO 588,349 3,774 $62,817 442,739 Washoe, NV 410,443 64 $19,949 208,318 Sarasota, FL 369,535 635 $7,927 158,001 Collier, FL 315,839 155 $13,682 131,937 Boulder, CO 282,304 386 $17,719 154,367 Loudoun, VA 278,797 534 $15,903 129,253 Napa, CA 133,522 172 $6,940 65,201 Tompkins, NY 101,136 212 $4,116 50,689 Roanoke, VA 90,420 359 $2,946 35,830 Broomfield, CO 53,691 1,951 $3,927 30,517 Routt, CO 21,580 9 $1,474 14,245 a. Retrieved from US Census ( Census, 2011 ) b. Retrieved from IMPLAN ( IMPLAN, 2010 ) Figure 4 2 : Number of cities by type in meta analysis.


! 62 4.2.2 City Characteris t i cs City Characteristics from 21 US Cities : ([\]^! ; 6 lists the 21 US cities of this analysis, along with a few parameters that elucidate the uniqueness' of these cities. Population is as large as 8 million (NYC) and as small as 21,000 (Routt). P opulation density r anges as high as 27,000 people/square mile, to as low as 9 people/square mile. T hese two parameters, among others, a ffect energy use and GHGs in cities To show some of the differences in in boundary energy use across these cities, we c ompu te energy use efficiencies for commercial industrial, r esidential and transportation sectors for the 21 cities. The computed efficiencies are: Comm ercial Ind ustrial : kWh/GDP; therms/GDP; kBTU/GDP (all on annual basis) Resid ential : kWh/cap/mo; therms/cap/mo; kBTU/cap/mo Motorized Surface Transport: VMT/(residents+jobs)/day ([\]^! ; 6 = shows a summary of the most efficient and the most inefficient cities from the sample of 21 cities by parameter W e note the large range in cities by parameter. In terms of commercial industrial energy use, NYC is the most efficient ( 349,235 kBTU/GDP/yr ), and Roanoke is the most inefficient ( 1,214,728 kBTU/GDP/yr ). In terms of residential energy use, Broward is the most efficient (1,788 kBTU/cap/mo), and Roanoke is the most inefficient (4,093 kBTU/cap/mo). Finally, in terms of transportation system efficiency, NYC is the most efficient (5.8 VMT/(res+job)/day), and Sarasota is th e most inefficient (24.3 VMT/(res+job)/day). Table 4 3 : Summary of most efficient and most inefficient cities by energy use effic iency parameter. Most Efficient Most Inefficient Commercial Industrial : [kBTU/GDP/yr ]; {city} [29,103]; {NYC} [101,227]; {Roanoke} Residential : [kBTU/cap/mo]; {city} [1,788]; {Broward} [4,093]; {Roanoke} Transportation : [VMT/(res+job)/day]; {city} [5.8]; {NYC} [24.3]; ( Sarasota )


! 63 The following sets of figures ( 4"_`a^! ; 6 = thru 4"_`a^! ; 6 B ) show how the 21 cities compare to the US average for each of the energy efficiency p arameters. These figures show that although the sample of cities is small, and represents all city counties that had worked with ICLEI as of 2010, the cities' energy efficiency characteristics are widely distributed across the national averages. Commercial Industrial Energy Efficiencies : Figure 4 3 : Commercial Industrial electricity use for 21 cities distributed across the US average. Figure 4 4 : Commercial Industrial natural gas use for 21 cities distributed across the US average.


! 64 Figure 4 5 : Commercial Industrial energy use for 21 cities distributed across the US average. Residential Energy Efficiencies : Figure 4 6 : Residential electricity use for 21 cities distributed across the US average.


! 65 Figure 4 7 : Residential natural gas use for 21 cit ies distributed across the US average. Figure 4 8 : Residential energy use for 21 cities distributed across the US average.


! 66 Transportation System Efficiency : Figure 4 9 : Transportation system efficiency for 21 cities distributed across the US average. 4.2.3 IO Table Data IO tables for the 21 cities were obtained from MIG IMPLAN, Inc. Because IO tables are not originally intended to be used for tracking energy use and GHGs, energy flows emerging from IO tables must be compared to actual energy end use data and corrected for any misma tches. Mismatches in IO tables between monetary flows and energy flows commonly stem from three reasons: 1) residential energy use is downscaled from national data without consideration for local wealth, climate, or urban form; 2) self employed residents o f the city owning and operating energy assets outside of the city; and 3) large corporate headquarters located in the city. Both 2) and 3) present the illusion of highly producing sectors. Thus, mismatches when compared to actual energy end use reported by cities must be corrected where identified. Energy end use data from each of the 21 cities provided by ICLEI allowed calibration for electricity, natural gas, along with gasoline and diesel use within the community. As shown in chapter 3, there were differ ences in IMPLAN projection of local electricity generation compared to EPA eGRID, thus eGRID was used to adjust for the community wide electricity use that is generated in the city ( EPA, 2011 ) Procedures for comparing energy flows, and calibrating IO tables are described in section 3.4 (chapte r 3). The comparisons between


! 67 the unadjusted IO table and GHG inventory reports are presented in Appendix A for each fuel across the 21 cities. Sector Output/Trade was analyzed done to evaluate the top exporting sectors, by dollar output, for each city. Th is step was key towards identifying artificially high producing sectors caused by self employed residents or by large headquarters. In the absence of consistent, more robust databases, an attempt to validate the top sectors was approached through the use o f publicly available data. Seen in ([\]^! ; 6 ; are the top three trade sectors for the 21 cities. Upon review, a few of these make rational sense: Snohomish (Aircraft) supporting operations for Boeing; NYC (Investment services) is a financial hub; and Napa (Wineries) a large producer of wine. Others required additional research to confirm, such as: Loudoun (Telecommunications) was verified through a series of local repor ts ( TAG, 2002 ) ; and Routt (Coal Mining) was also found to have large coal mining operations ( Valley, 2011 ) Meanwhile, those that were found not to agree with the physical case were zeroed out, a s they translate to large GHG emissions


! 68 Table 4 4 : Top three output trade sectors for 21 US cities. City (Type) Description % of Total Total Exports (mill $) Real estate related services 15% Hotels and motel services 4% Amusements and recreation 4% Professional and technical 10% Metal window and door manufacturing 8% Real estate 8% Aircraft manufacturing 50% Aircraft parts and equipment 4% Telecommunications 4% Real estate 12% Telecommunications 9% Hospitals 7% Wholesale trade distribution services 10% Real estate related services 9% Rental services 3% Software publishers 13% Scientific research and development services 7% Pharmaceutical and medicine manufacturing 7% Insurance 16% Computer related services 9% Motor homes 6% Management of companies and enterprises 10% Real estate 9% Telecommunications 7% Telecommunications 40% Air transportation services 8% Computer systems design services 4% Wineries 47% Pharmaceutical and medicine manufacturing 8% Real estate 5% Refined petroleum products 7% Pharmaceutical preparations 6% Wholesale trade distribution services 4% Securities, commodity, investments services 24% Advertising and related services 8% Real estate related services 7% Fundstrustsand other financial vehicles 10% Hospitals 9% Collegesuniversitiesand junior colleges 9% Oil and natural gas 12% Real estate related services 11% Air transportation services 6% Funds, trusts, and other financial services 10% Hotels and motel services 8% Miscellaneous manufactured products 5% Semiconductor and related devices 12% Wholesale trade distribution services 10% Management of companies and enterprises 3% Telecommunications 16% Management of companies and enterprises 7% Pharmaceutical preparations 6% Wholesale trade distribution services 6% Management of companies and enterprises 5% Insurance 3% Wholesale trade distribution services 12% Water transportation services 9% Air transportation services 6% Real estate 19% Coal mining 16% Amusements and recreation 9% Junior colleges, colleges, universities 40% Motor vehicle parts 11% Aircraft engines and engine parts 4% $4,046 $51,060 $6,563 $47,769 $58,535 $1,051 Tompkins (P) $7,959 $6,106 $18,846 $25,675 $58,116 $13,162 $2,990 $32,314 $12,115 $5,915 $154,658 $409,030 $46,284 $48,761 $16,568 METRO (B) Broomfield (B) Multnomah (B) Miami-Dade (B) Routt (P) DVRPC (B) NYC (B) Philadelphia (B) Denver (B) Washoe (B) Boulder (C) Roanoke (C) Westchester (C) Loudoun (B) Napa (B) Collier (C) Sarasota (C) Snohomish (C) Sacramento (C) Broward (C)


! 69 IO data is also verified against two econometric parameters that are generally publicly available for US c ities; GDP/capita and Income/cap. GDP/cap is reported by the BEA nationally, for states, and for metropolitan statistical areas (MSA) ( BEA, 2012 ) W e obtain city specific GDP/cap from IMPLAN, and compare to the GDP/cap for the corresponding MSA retrieved from BEA Meanwhile, BEA does publish estimates for per capita income down to the county scale, thus allowing for a direct comparison to the same retrieved from IMPLAN. ( [\]^! ; 6 C shows t hese comparisons across the 21 cities in this analysis. Note, MSA 's for each of the 21 cities are shown in S4 4. Results from Income/cap between the two datasets (IMPLAN & BEA) show both a re generally in line with each other since both report county level data; h owever, there are some apparent differences in GDP/cap between the two. I n our IMPLAN model NYC is represented by the five county region of Bronx, Kings, New York, Queens, and Richmond counties, constituting a populat ion of 8.4 million, while the comparative MSA is the New York Northern New Jersey Long Island NY NJ PA MSA which has a total population of 18.9 million ( Census, 2011 ) NYC (five counties) may be both a larger consumer and produc er (exporter) of goods/services when compared to its average MSA potentially explaining NYC's larger GDP/cap compared to its MSA Another notable differen ce in GDP/cap is seen in Denver, where IMPLAN estimates yield $106,769 GDP/cap and BEA MSA average is $57,595 GDP/cap The Denver Aurora Broomfield MSA consists of ten counties (Denver, Arapahoe, Jefferson, Adams, Douglas, Broomfield, Elbert, Park, Clear Creek, and Gilpin) with a to tal population of 2.5 million ( Census, 2011 ) As shown in ([\]^! ; 6 < GDP/cap for the total Denver Aurora Broomfield MSA as estimated from IMPLAN is in line with the estimate obtained from BEA. Th erefore we conclude that some of the differences setting Denver (county) apart from some of the other counties is the MSA can be higher emp loyee compensation and/or higher final consumption and exports.


! 70 T able 4 5 : Per capita GDP and Incomes from IMPLAN and BEA, for 21 US cities. GDP/cap Income/cap County IMPLAN BEA (MSA avg) IMPLAN BEA (county) NYC, NY $78,757 $60,965 $51,814 $50,881 DVRPC, PA/NJ $54,730 $50,563 $46,149 $40,914 Miami Dade, FL $46,054 $43,826 $35,852 $32,057 Broward, FL $42,454 $45,847 $42,276 $42,673 Oregon METRO, OR $53,017 $52,122 $40,402 $39,826 Philadelphia, PA $48,657 $51,225 $32,996 $31,288 Sacramento, CA $46,203 $43,489 $35,473 $35,110 Westchester, NY $60,053 $57,879 $64,859 $63,826 Multnomah, OR $66,414 $56,099 $41,913 $41,619 Snohomish, WA $32,328 $58,332 $36,416 $34,960 Denver, CO $106,769 $57,595 $52,017 $51,895 Washoe, NV $48,603 $46,095 $44,888 $44,356 Sarasota, FL $21,450 $34,701 $48,812 $50,033 Collier, FL $43,319 $43,216 $60,001 $63,620 Boulder, CO $62,764 $55,486 $47,624 $46,376 Loudoun, VA $57,042 $67,743 $44,420 $50,009 Napa, CA $51,980 $49,291 $45,519 $45,677 Tompkins, NY $40,698 $33,947 $33,200 $33,902 Roanoke, VA $32,584 $39,643 $41,358 $38,240 Broomfield, CO $73,144 $57,595 $34,788 $38,215 Routt, CO $68,322 $46,938 $43,723 $46,021


! 71 Table 4 6 : per capita GDP and Income for the 10 counties of the Denver Aurora Broomfield MSA. County Population Jobs GDP (mill $) Income (mill $) GDP/cap Income/cap Denver 588,349 646,259 $62,817 $30,604 $106,769 $52,017 Broomfield 53,691 35,618 $3,927 $1,868 $73,144 $34,788 Arapahoe 545,089 423,494 $44,442 $27,105 $81,531 $49,725 Jefferson 529,354 289,807 $22,757 $25,159 $42,991 $47,527 Adams 422,495 202,704 $14,429 $12,598 $34,151 $29,818 Douglas 272,117 110,158 $9,935 $14,418 $36,509 $52,986 Elbert 22,720 6,245 $339 $871 $14,940 $38,353 Park 17,004 3,830 $200 $502 $11,751 $29,502 Clear Creek 8,956 5,198 $420 $520 $46,884 $58,049 Gilpin 5,091 5,271 $431 $186 $84,674 $36,534 Denver Aurora Broomfield MSA TOTAL 2,464,866 1,728,584 $159,697 $113,830 $64,789 $46,181 The modified IMPLAN model was used to analyze in boundary and trans boundary GHGs after correcting for energy end use obtained from the ICLEI database and electricity generation from eGRID. Upon also examining the Toxic Release Inventory (TRI) ( EPA, 2012 ) and the County Business Patterns ( Census, 2012b ) we discovered that cities appear to have good energy "use" data, but do not report GHGs from industrial process (non fossil combustion) activities, since cities are not required to do so. ([\]^! ; 6 G shows some of the in boundary industrial process ac tivities identified through TRI for selected cities that are not reported by cities. Therefore, we proceed with the analysis in this chapter using the modified IMPLAN IO tables, corrected for key parameters, as a model to represent the various city types. We do not assert that the IMPLAN models represent each individual city accurately, but we expect different city types to be well represented with reasonable in boundary energy use and associate d trans boundary supply chains.


! 72 Table 4 7 : Examples of industrial processes located in selected US cities. City In Boundary Industrial Processes Boulder Asphalt paving, C ement Production Heavy Industrial Equipment, Computer P arts Broward Asphalt, Concrete Bricks, S tone Miami Dade Petroleum Refin ing, Cement Production, C hemical Production Napa Crop Production for wine industry Oregon METRO Petroleum Refining, Cement Production, C hemical P roduction Philadelphia Petroleum Refining Routt Coal M ining Sacramento Chemicals, Asphalt, Brick/Tile 4.3 Methods and Results 4.3.1 Rationale for I n frastructure Allocation In this section we develop a rationale for proposing infrastructures that should be allocated to cities based on use. In developing a rationale, t wo plots for each infrastructure sector were compiled to evaluate the relationships between GDP and GHGs across the 21 cities. The first plot in each series shows GDP vs. GHGs from Infrastructure Use (local plus imports), and the second plot shows GDP vs. GHGs embodied in exports, for each respective infrastructure sector Note, infrastructures relate to Water, Sanitation, Energy, Food, Transportation, and Materials for Shelter. The following criteria were proposed to evaluate which sectors may be considere d infrastructures: a. High c orrelation between community GDP vs GHG in community wide (residential commercial industrial) u se of a sector ( R 2 >0.70 ), and b. Weak c orre lation between community GDP vs. GHG embodied in e xport s of that same sector ( R 2 <0.30 ) This combination is considered to represent a strong correlation of infrastructure provisioning on economic development, while also illustrating that exports of these sectors do not significantly contribute to economic development broadly across cities. In othe r words, if a strong correlation (R 2 > 0.70 ) between GDP and GHGs from


! 73 Infrastructure use is observed, it suggests that the respective infrastructure sector is important for economic activity. However, if the exports of the same show a weak correlation (R 2 < 0.30) with GDP across the 21 cities, it suggests the sector in question is likely not a significant creator of GDP across the board. The identical approach is also extended to all other (non infrastructures) goods and services allowing for an evaluation of additional supply chains that may be largely considered as basic for any economy. We present ([\]^! ; 6 E which shows the percent GHGs that each of the forthcoming sectors contribute to national U.S. GHGs. It's noted that the infrastructure sectors alone cover 63% of national GHGs. T he last two columns of the table show the computed regression fits (correlations) for each pair (Use and Exports) for the 21 cities ; mai ntaining US average emission factors for comparison Let us begin with electricity. Most electricity used in cities is generated outside of cities. In our sample of 21 cities, three (or 14%) produce surplus electricity for export. 4"_`a^! ; 6 $K shows a strong correlation between GDP and GHGs in community wide electricity u se (R 2 = 0.84), while GDP vs. GHGs embodied in elec tricity e xport s (R 2 = 0.02) confirms that not many communities are producers of surplus electricity. Indeed this may effectively provide a rationale for considering electricity as a Scope 2 item.


! 74 Table 4 8 : R egression correla tions (R 2 ) from Use and Export from all sectors for 21 cities, along with the % contributing to national (U.S.) GHGs. Sector Category % of National GHGs R 2 from Use R 2 from Export Current Infrastructure Sectors Electricity 39.7% 0.84 0.02 Food Agriculture /Livestock 7.4% 0.73 0.14 Water/WW 4.3% 0.30 0.14 Freight 3.4% 0.46 0.41 Fuel Production 3.1% 0.82 0.13 Air Travel 2.8% 0.70 0.43 Cement 1.6% 0.73 0. 12 Iron/Steel 3.3% 0.77 0.13 Potential Infrastructure Sectors Transport Services (Marine, Rail) 4.5% 0.84 0.38 Mining 3.9% 0.79 0.007 Durable Goods 0.3% 0.87 0.04 Communications 0.5% 0.75 0.27 Alcoholic Beverages, Tobacco 0.1% 0.77 0.01 Natural Gas Production 0.5% 0.19 0.03 Other Sectors Industry/Manufacturing 8.9% 0.93 0.51 Services 4.3% 0.59 0.67 Government Services 4.0% 0.41 0.31 Construction 3.3% 0.34 0.35 Wholesale/Retail 1.0% 0.27 0.65 Food Manufacturing 0.9% 0.93 0.51 Education 0.6% 0.24 0.03 Restaurant/Hotels 0.5% 0.63 0.48 Electronics 0.4% 0.36 0.19 Health 0.3% 0.34 0.28 Recreation 0.2% 0.49 0.10 Other Textiles 0.1% 0.87 0.79 Furniture 0.1% 0.89 0.55 Apparel 0.02% 0.94 0.68


! 75 Figure 4 10 : Electricity GDP vs. GHGs from Use (left) and Export (right). As shown in 4"_`a^! ; 6 $$ Food Agriculture /Livestock also meets the criterion with high (R 2 = 0.73) correlation between comm unity GDP and GHG in use and weak correlation (R 2 = 0.14) between community GDP and GHG embodied in exports Figure 4 11 : Food Agriculture GDP vs. GHGs from Use (left) and Export (right). The GHGs associated with Fuel Refining ( 4"_`a^! ; 6 $. ) shows similar patterns The correlation between community GDP and GHG in use is high (R 2 = 0.82), and the correlation betw een community GDP and GHG embodied in exports is weak (R 2 = 0.11).


! 76 Figure 4 12 : Fuel Refining GDP vs. GHGs from Use (left) and Export (right). The remaining infrastructure correlations are placed at t he end of the chapter in SI (S 4 6, a f) Note that the sample of 21 cities is relatively small, and with a larger set of cities more pronounced patterns might emerge. As for the non infrastructure sectors (all of which are shown in SI : S 4 6, g v ), Services ( 4"_`a^! ; 6 $= ), and Health Care ( 4"_`a^! ; 6 $; ) are examples of sectors where both the Use and Export are correlated to economic development, hence unsuited to allocation. Meanwhile, further investigation reveals that Iron/Steel use is meets our criterion, and may be suited for allocation Figure 4 13 : Services GDP vs. GHGs from Use (left) and Export (right).


! 77 Figure 4 14 : Health GDP vs. GHGs from Use (left) and Export (right). Figure 4 15 : Iron/Steel GDP vs. GHGs from Use (left) and Export (right). 4"_`a^! ; 6 $C shows that Iron / Steel indeed may be a common supply chain s erving all cities, and only one group, the Multnomah/Oregon METRO tandem, which are collocated in the same area, were computed as net exporters of Iron/Steel. Therefore, Iron/Steel production was included as an infrastructure sector in the analysis that fo llows, given the large amounts of Iron/Steel in the built environment. Note, as this analysis uses IO tables to quantify the supply chains of sectors, additional effort is required to identify public data sources that would allow cities to effectively allocate Iron/Steel.


! 78 4.3.2 City Typology, Relationships to Population and Proxies City Typology was covered in depth in chapter 3, and generates a 3 way typology for cities as: Net Consumers, Net Producers, or GHG Trade Balanced. In a trade balanced c ommunity, where G H G M n e t n o n i n f r a << GHG TBIF GHG CBF In a producer community where G H G M n e t n o n i n f r a is a large negative, GHG TBIF > GHG CBF In a consumer community where G H G M n e t n o n i n f r a is a large positive, GHG TBIF < GHG CBF where GHG non infra Mnet are the GHG embodied in net imports of non infrastructure sectors, and is compared to TBIF to determine its relative size. In practice, a community is said to be a Net Consumer if GHG non infra Mnet are >15% compared to TBIF; a Net Producer if < 15% Net Producer; and Trade Balanced if within +/ 15%. Note, we refer to the ratio of GHG non infra Mnet / TBIF as the typology degree. In our sample of sample of 21 cities, 8 are computed as Net Consumers, 11 Trade Balanced, and 2 Net Producers. Among them are 3 exporters of electricity (Routt, Tompkins, and Westchester), and 1 net exporter of cement (Miami Dade). ([\]^! ; 6 B (net producers), ([\]^! ; 6 $K (balanced), and ([\]^! ; 6 $$ (net consumers) present the typology degree for each city. Results show tha t larger cities tend to be balanced, signaling a strong presence of both production and consumption activities in larger US cities. The other balanced communities appear in close proximity to large metros possibly signaling links between their economies ( e.g., Broomfield near Denver; Washoe near Reno; Loudoun near Washington DC). G H G S c o p e s 1 + 2 + 3 T B I F


! 79 Table 4 9 : Degree by which communities are measured as Net Producer, along with alternate metrics for representing typology. Net Producer Community GHG non inf ra Mnet / GHG TBIF Comm Ind kWh/cap Energy Use Ratio (Comm Ind/ HH ) Employment Intensity (jobs/cap) GDP/resident ($/cap) Tompkins, NY 22% 4,811 1.56 0.50 $40,698 Routt, CO 39% 13,271 2.06 0.66 $68,322 net exporter of infrastructure Table 4 10 : Degree by which communities are measured as Trade Balanced, along with alternate metrics for representing typology. Trade Balanced Community GHG non inf ra Mnet / GHG TBIF Comm Ind kWh/cap Energy Use Ratio (Comm Ind/ HH ) Employment Intensity (jobs/cap) GDP/resident ($/cap) Loudoun, VA 15% 7,761 0.92 0.40 $57,042 Napa, CA 11% 4,333 0.94 0.49 $51,980 DVRPC, PA/NJ 10% 6,651 1.36 0.46 $54,730 NYC, NY 9% 4,084 1.19 0.44 $78,757 Philadelphia, PA 8% 5,900 1. 19 0.44 $48,657 Denver, CO 6% 8,704 2.11 0.75 $106,769 Washoe, NV 2% 6,836 1.15 0.51 $48,603 METRO, OR 3% 7,739 1.86 0.52 $53,017 Broomfield, CO 11% 8,326 1.21 0.57 $73,144 Multnomah, OR 13% 8,065 1.74 0.63 $66,414 Miami Dade, FL 7% 5,925 1.00 0.42 $46,054 net exporter of infrastructure Table 4 11 : Degree by which communities are measured as Net Consumer, along with alternate metrics for representing typology. Net Consumer Community GHG non inf ra Mnet / GHG TBIF Comm Ind kWh/cap Energy Use Ratio (Comm Ind/ HH ) Employment Intensity (jobs/cap) GDP/resident ($/cap) Collier, FL 42% 6,568 0.88 0.42 $43,319 Sarasota, FL 34% 5,123 0.79 0.43 $21,450 Snohomish, WA 25% 4,798 0.99 0.33 $32,328 Sacramento, CA 25% 4,262 0.90 0.45 $46,203 Broward, FL 23% 6,132 1.07 0.43 $42,454 Boulder, CO 22% 7,475 1.23 0.55 $62,764 Roanoke, VA 20% 5,738 0.81 0.40 $32,584 Westchester, NY* 23% 3,464 0.86 0.43 $60,053 net exporter of infrastructure Next, cities were organized by type (see ([\]^! ; 6 B thru ([\]^! ; 6 $$ ) and the absolute value of the typology degree was plotted against population to explore patterns. Our


! 80 initial hypothesis was that larger cities (based on population) would tend to be more trade balanced. As seen in 4"_`a^! ; 6 $< smaller communities (lower populations) can be any of the three typologies. However, a convergence towards trade balanced in GHG may be emerging for larger communities (>2,000,000 peop le). Ther efore it may be that larger communities are consistently balanced, hosting both large amounts of production and consumption activities. Recall that the typology degree represents the % of GHG in non infrastructure imports or exports relative to TB IF. Figure 4 16 : Absolute value of typology degree vs population for 21 communities. Larger communities are shown to approach trade balanced. We also explored various parameters that we hypothesized as being reasonable proxies for typology. The three parameters are: total commercial industrial electricity use per capita (Comm Ind kWh/cap) ratio of commercial industrial energy use (in kBTU) to residential energy use (in kBTU) (Comm Ind/ HH ) and employment intensity (Jobs/cap) Note, t he co rresponding US averages for the three parameters are: Comm Ind kWh/cap = 7,704 kWh/cap; Comm Ind/Res = 1.99; Jobs/cap = 0.44. The three parameters are eval uated for correlation with our typology degree in 4"_`a^! ; 6 $G thru 4"_`a^! ; 6 $B While each of the observed trends are as anticipated, the ratio between commercial industrial energy use and residential energy use (Comm Ind/ HH ) appears the best (R 2 = 0.5) suited to serve as a quick proxy for identifying city typology.


! 81 Figure 4 17 : Correlation in community Typology Magnitude versus Total Commercial Industrial Electricity Use per Capita. Figure 4 18 : Correlation in community Typolog y Magnitude versus Commercial Industrial Energy use (in kBTU) per Residential Energy Use (in kBTU). Figure 4 19 : Correlation in community Typology Magnitude versus Employment Intensity.


! 82 4.3. 3 Trans Boundary Supply Chains of Cities The trans boundary supply chain footprints serving cities are computed using each city's calibrated IMPLAN model. We apply 0b`[c"de! = 6 = presented in chapter 3, to compute each city's supply chain GHGs. Recall, G H G C B F = B [ ] T L O [ ] + E F u s e # $ % F [ ] + M F [ ] { } + B [ ] L [ ] M n e t i n f r a # $ + B [ ] L [ ] M n e t n o n & i n f r a # $ !!!!!!!!!!!! = B [ ] T L O [ ] + E F u s e # $ % F [ ] + M F [ ] { } + B [ ] L [ ] M Z + M F + E [ ] Following the equation, GHGs are divided into thre e broad categories: Territorial trans boundary supply chains in Imports, and trans boundary supply chains in Exports. The first term of the equation, (B)(TLO) represents territorial GHGs from the production in serving Local Consumption (F), and Exports (E). Namely, B [ ] T L O [ ] = [ B ] [ L l o c a l ] [ F + E ] Equation 4 1 where, L local is the respective city's local Leontief total requirements matrix. The other element comprising territorial GHGs are from direct energy (natural gas, gasoline, and diese l) use by households. The ir sum equals Territorial GHGs for each city, shown as IB in 4"_`a^! ; 6 .K thru 4"_`a^! ; 6 .. Trans boundary GHGs embodied in imports (IM) either serve local industries and/or homes directly. Moreover, trans boundary GHGs in imports serving local industries can be separated as produc tion for local consumers, and production for exports, by multiplying these imports with the ratio of local outputs for consumers to total local output, { (L)(F)/(L)(TLO) } by sector. These i mports are further separated into infrastructures and non infrastru ctures. Imports for each city are shown in the second bar labeled I M in 4"_`a^! ; 6 .K thru 4"_`a^! ; 6 .. A third bar in the figures shows GHGs embodied in exports for each city. These exports represent the full supply chain including local commercial industrial activities exported, and their as sociated supply chains. In equation form, [ B ] [ L U S ] [ E ] = B [ ] L l o c a l [ ] [ E ] + [ B ] [ L U S ] [ M Z E ] Equation 4 2


! 83 where L US is the US national Leontief matrix, M Z,E are the imports to local commercial industrial users for exports estimated as discussed above and B is the GHG intensity vector. A s IO data does not split imports into production for local consumption versus that for exports, GHGs embodied in the supply chains of exports using 0b`[c"de! ; 6 is compared to those estimated from the Export vector retrieved from the IO data for each city. ([\]^! ; 6 $. shows that our method for separating imports to local commercial industrial users produces reasonable estimates of the full export supply chain as many of the errors are below 15% The few larger errors may in part be attributed to the technique or to L local which may not fully capture local production requirements. Thus we proceed in our analysis using this separat ion for illustrating the supply chains of cities. Table 4 12 : GHGs embodied in Exports. Calculated vs. IMPLAN total Export vector. City [B][L Local ][E] [B][L US ][M Z,E ] [B][L US ][E]: Calculated [B][L US ][E]: IMPLAN % Diff Collier 846,259 1,945,450 2,791,709 2,497,715 12% Sarasota 461,173 2,204,421 2,665,593 1,975,204 35% Snohomish 1,652,523 5,259,992 6,912,515 6,029,561 15% Sacramento 3,960,186 5,508,879 9,469,065 8,484,080 12% Broward 8,463,628 10,587,261 19,050,889 17,486,903 9% Boulder 1,159,811 2,426,372 3,586,183 3,540,732 1% Roanoke 287,268 747,155 1,034,424 1,027,715 1% Westchester 5,744,472 4,040,164 9,784,636 10,113,098 3% Loudoun 2,408,795 2,556,284 4,965,079 4,511,936 10% Napa 537,608 1,391,696 1,929,304 1,994,607 3% DVRPC 34,454,245 42,851,364 77,305,608 69,370,700 11% NYC 36,555,053 29,317,184 65,872,237 81,597,118 19% Philadelphia 7,079,029 8,016,292 15,095,321 15,322,841 1% Denver 3,472,998 9,477,079 12,950,077 12,783,593 1% Washoe 1,883,643 3,648,350 5,531,993 5,693,676 3% METRO 10,868,539 12,154,565 23,023,104 21,480,451 7% Broomfield 269,670 1,239,401 1,509,071 1,544,245 2% Multnomah 8,520,606 9,172,289 17,692,895 18,161,154 3% Miami Dade 26,190,073 10,546,429 36,736,502 34,536,417 6% Tompkins 1,692,002 859,627 2,551,628 2,707,895 6% Routt 1,254,395 239,908 1,494,303 1,424,675 5%


! 84 Figure 4 20 : NET PRODUCERS. GHGs relating to Territorial, Import, and Export supply chains. For net producing cities, 4"_`a^! ; 6 .K illustrates that t erritorial GHGs associated with these cities are larger (>600 mt CO 2 e/GDP) compared to GHGs embodied in imports. We also observe that a large portion of the total footprint is exported For trade balanced cities (seen in 4"_`a^! ; 6 .$ ) it's noted that territorial GHGs are <400 mt CO 2 e/GDP, with the exception being Miami Dade, which is a net exporter of infrastructure supporting both local consumption & exports. GHGs embodied in imports for some trade balanced cities are roughly equal to their territorial GHG (e.g., NYC), while GHGs in imports for others are substantially larger (e.g., Napa) In these trade balanced cities it s consistently observed that GHGs embodied in non infrastructure imports are roughly equal to GHGs embodied in non infrastructure exports. Across net consumers ( 4"_`a^! ; 6 .. ) it s observed that GHGs embod ied in imports are larger than territorial GHGs. Unlike the other cit y types, GHGs embodied in exports are small and larger amounts of territorial GHGs are for local consumption In sum, these figures show that territorial GHGs are highest for net producers and lowest for net consumers. GHGs embodied in imports are reversed, as they are highest for net consumers and lowest for net producers As for GHGs embodied in exports, they are largest for net prod ucers and trade balanced cities.

PAGE 100

! 85 Figure 4 21 : TRADE BALANCED. GHGs relating to Territorial, Import, and Export supply chains.

PAGE 101

! 86 Figure 4 22 : NET CONSUMERS. GHGs relating to Territorial, Import, and Export supply chains.

PAGE 102

! 87 4.3.4 Comparing Coverage of Urban Metabolism by TBIF & CBF Urban metabolism has been long studied within the literature of resource use in cities. The initial work by Abel Wolman was the catalyst as it defined the urban metabolism of cities as "all commodities needed to sustain the city's inhabitants at home, work and at play" ( Wolman, 1965 ) However, due to the extreme data requirements for completing a full metabolic analysis of a city, others have on ly included selected materials, e.g., ( Hanya & Ambe, 1976 ; Liang & Zhang, 2011 ; Newcombe, Kalma, & Aston, 1978 ; Sahely, Dudding, & Kennedy, 2003 ) This is a first of its kind analysis for cities, in that the m etabolic structure for city economies wi th all of their associated supply chains are included. Results are presented in terms of GHGs covered by Territorial (or in boundary) I nfrastructures (TBIF) and C onsumption (CBF). The total footprint refers to all in boundary GHG plus GHG embodied in imports. Total GHG Footprint = GHG IB + GHG IM Equation 4 3 Recall that TBIF accounts for GHG from all territorial activity, along with GHG from allocated infrastructures supporting the city as a whole. On the other hand, CBF accounts for GHGs attributed to final consumption, regardless of production location. CBF also allocate s out the full supply chain of exports. The GHGs that neither TBIF nor CBF account for are those embodied in non infrastructure im ports to local businesses & industries for exports. These are shown in blue cross hatching in 4"_`a^! ; 6 .= O f the total footprint thes e correspond to 12%, 17%, and 12 % for net producers trade balanced, and net consumers respectively on average Comparing the coverage between TBIF and CBF for net producer cities shows that on average, Territorial captures 68% of the total footprint, TBIF captures 75% of the total f ootprint, and CBF only captures 35%. The incremental portion of consumption not captured by TBIF is 13% of the footprint. However, the portion of export production not captured by CBF is 53%. See 4"_`a^! ; 6 .= 4"_`a^! ; 6 .; and 4"_`a^! ; 6 .G

PAGE 103

! 88 For trade balanced cities, Territorial captures 42%, TBIF captures 63%, and CBF capt ures 57% of the total footprint, on average. The incremental portion of consumption not captured by TBIF is 20% of the total footprint. The portion of export production not captured by CBF equals 26% of the total footprint. See 4"_`a^! ; 6 .= 4"_`a^! ; 6 .C and 4"_`a^! ; 6 .E For n et consumer cities, Territorial captures 37%, TBIF captures 62%, and CBF captures 71% of the total footprint, on average. The incremental portion of consumption not captured by TBIF is 26% of the total footprint. The portion of export production not captured by CBF is 17% of the total footprint See 4"_`a^! ; 6 .= 4"_`a^! ; 6 .< and 4"_`a^! ; 6 .B These results show that indeed the amount of the urban metabolism captured greatly depends on the lens with which the analysis is approached. TBIF captures all activity within the boundary along with GHGs embodied in infrastructure imports, although misses the additional non infrastructure consumption to households. Meanwhile, CBF captures all the GHGs embodied in the supply chains relating to consumption, but in the interim misses key aspects of local economies for exports which are larger in balanced an d net producing cities. In total, TBIF captures more than 62% of the total footprint for all three city types. However, the coverage of CBF depends on the city type which on average can be as low as 35% or as high as 71%.

PAGE 104

! 89 Figure 4 23 : GHGs covered by TBIF and CBF, respectively. Shown as the average of cities by typology. Coverage by Infrastructure (TBIF), by Typology : Figure 4 24 : NET PRODUCER % of total footprint GHGs covered by infrastructure supply chain.

PAGE 105

! 90 Figure 4 25 : TRADE BALANCED % of total footprint GHGs covered by infrastructure supply chain. Figure 4 26 : NET CONSUMER % of total footprint GHGs covered by infrastructure supply chain.

PAGE 106

! 91 Coverage by Consumption (CBF), by Typology : Figure 4 27 : NET PRODUCER % of total footprint GHGs covered by consumption supply chain. Figure 4 28 : TRADE BALANCED % of total footprint GHGs covered by consumption supply chain.

PAGE 107

! 92 Figure 4 29 : NET CONSUMER % of total footprint GHGs covered by consumption supply chain. 4.3.5 Metrics for Comparing GHGs Associated with Cities Exploring suitable metrics for comparing GHGs associated with cities builds on the dat a results described above. The key question asked here is whether cities with lower GHG footprints reported by a certain metric (e.g., GHG TBIF or GHG CBF ) truly represent greater urban efficiency. The first step in our method is to construct an urban effici ency index (UEI) that represents the major energy efficiency characteristics of a city. We propose an UEI composed of three key attri butes: 1) production efficiency, 2) household energy efficiency, and 3) transportation system efficiency. Energy use is converted to GHG using national average emission factors for electricity (0.64 kg CO2e/kWh) and natural gas (5.4 kg CO2e/therm). The resulting commercial industrial GHGs are normalized by GDP, while R esidential GHGs are normalized by capita to represent production efficiency and household energy efficiency, respectively. Because a city's transportation serves both private commutes and job related travel the transportation system efficiency is represen ted as motor vehicle miles traveled (VMT) per residents+jobs The composite index for a given city is the sum of the three attributes.

PAGE 108

! 93 Table 4 13 : Attributes and units of the Urban Efficiency Index (UEI) Attr ibute of the U EI Units of Attribute Production Efficiency GHGs from Commercial Industrial Energy Use per GDP (mt CO2e/GDP) Household Energy Efficiency GHGs from Household Energy Use per resident (mt CO2e/capita) Transportation System Efficiency VMT's per residents plus jobs (VMT/(res+jobs)) In sample comparisons are performed by evaluating the correlations between each of the ten metrics presented in ([\] ^! ; 6 $; versus the composite urban energy efficiency index (UEI) for each of the 21 c ities By doing so we aim to answer whether cities with lower GHG indeed represent greater urban efficiency. Tabl e 4 14 : Metrics evaluated for comparing GHGs associated with cities against the UEI GHG Accounting Method GHGs Metric and Units Per GDP (community GDP) Per capita (resident population) Territorial (in boundary) Purely Territorial (GHG IB ) GHG IB /GDP GHG IB /cap Various versions of TBIF Purely Territorial + Electricity Allocated (GHG TBIF Scope 1+2 ) GHG Scope 1+2 /GDP GHG Scope 1+2 /cap Plus Scope 3 w/o Allocating (GHG TBIF Scope 1+2+3 ) GHG Scope 1+2+3 /GDP GHG Scope 1+2+3 /cap Scope 3 w/ Allocating (GHG TBIF, modeled ) GHG Scope 1+2+3 /GDP GHG Scope 1+2+3 /cap CBF Consumption Based GHGs (GHG CBF ) GHG CBF /GDP GHG CBF /cap Scope 3 is now allocated based on use The computed UEI s are shown in ([\]^! ; 6 $C City identities have been suppressed, and column 1 of the table shows each city's unique identifier code. The last column is the composite index, and is t he independent variable used in the evaluation.

PAGE 109

! 94 Table 4 15 : U EI for 21 US cities. City ID Code Production Efficiency Index Household Efficiency Index Transportation System Efficiency Index Composite Index A 0.71 0.78 1.05 2.54 B 0.64 0.71 1.04 2.39 C 1.02 1.12 0.97 3.12 D 0.88 0.99 0.86 2.73 E 0.75 0.84 0.87 2.45 F 1.32 1.50 1.17 3.99 G 1.02 1.39 1.38 3.80 H 1.65 1.32 1.49 4.45 I 0.97 1.08 1.21 3.26 J 0.87 0.97 1.09 2.93 K 1.14 0.94 0.98 3.06 L 1.13 0.75 0.79 2.67 M 0.52 0.78 1.21 2.50 N 0.99 0.92 0.98 2.89 O 1.17 0.82 0.94 2.92 P 1.06 0.80 0.49 2.35 Q 1.50 1.51 1.26 4.27 R 0.97 1.26 0.93 3.17 S 1.18 0.98 1.03 3.18 T 0.44 0.55 0.35 1.34 U 1.07 1.00 0.91 2.98 Correlations between the UEI and each of the ten metrics are used to evaluate suitable metrics for comparing GHGs associated with cities. The progression from 4"_`a^! ; 6 =K towards 4"_`a^! ; 6 == illustrates that GHGs normalized per capita (figures b ) do not capture all activities (production and homes), as observed by consistently low correlations. Meanwhile, the same progression for GHGs normalized by GDP (figures a ) shows that GHGs compared to a city's economic development (GHG/GDP) are more repres entative of all activities located in a city (production and ho mes ). It's observed that a per capita metric correlates significantly better when accounting for consumption based GHGs (R 2 =0.41, 4"_`a^! ; 6 =; ), suggesting that per capita is better suited for representing GHGs from household consumption. T he strong correlation (R 2 =0.76 4"_`a^! ; 6 =C ), between consumption based GHGs and expenditures shows that GHG CBF more directly illustrates the willingness of a city's residents to consume.

PAGE 110

! 95 Figure 4 30 : Metric #1 Correlation of Territorial GHGs vs. U EI a) per GDP, and b) per capita. a) b)

PAGE 111

! 96 Figure 4 31 : Metric #2 Correlation of Scope 1+2 GHGs vs. U EI a) per GDP, and b) per capita. a) b)

PAGE 112

! 97 Figure 4 32 : Metric #3 Correlation of Scope 1+2+3 GHGs vs. U EI a) per GDP, and b) per capita. a) b)

PAGE 113

! 98 Figure 4 33 : Metric #4 Correlation of TBIF GHGs vs. U EI a) per GDP, and b) per capita. a) b)

PAGE 114

! 99 Figure 4 34 : Metric #5 Correlation of Consumption Based GHGs vs. U EI a) per GDP, and b) per capita. a) b)

PAGE 115

! 100 Figure 4 35 : Correlation of GHG CBF /cap vs. Final Consumption Expenditures ($/cap) This analysis shows that territorial GHGs are ineffective for comparing cities w hether per GDP or per capita. GHG Scopes 1+2 remains a production based accounting type of approach; hence GHG/GDP is a good metric to compare production efficiencies of cities. Correlations improve with electricity allocated (R 2 =0.65), to including food agriculture & petroleum refining (R 2 =0.76 ). Additional items had small impact on the correlation. Being that our sample of 21 cities was skewed towards net consumer cities ( only included 2 net producing cities ) it may be that including a larger amount of producing cities would r educe the correlatio n of GHG CBF /GDP with UEI, and increase the correlation of GHG TBIF /GDP with UEI Table 4 16 : Impacts of infrastructures, beyond electricity, having the most impact on urban efficiency. Infrastructure Sector R 2 (w/o allocation) R 2 (w/ allocation) Fuel Refining 0.698 0.661 Iron/Steel 0.676 0.667 Air Travel 0.671 0.669 Food Agriculture 0.665 0.695 Cement 0.664 0.655

PAGE 116

! 101 4.4 Conclusion Th e meta analysis revealed several new insights that can lead to the overall understanding of t he urban metabolism for cities. Upon reviewing energy use and GHGs reported by US ci t i es, we learn that a lack of proto cols may attribute to ci t i es reporting high quality building energy end use and underreporting GHGs from indus trial processes P ollutants from industrial processes are currently only accounted for by national bodies (e.g., TRI) Therefore this research used the modified IMPLAN IO tables as a model for comparing the three GHG emission accounting methods. As a resul t from the four specific objectives in this chapter, the following highlights the major finding s : Objective 1 : A series of correlations show that electricity generation, fuel refining, air travel, and the production of food, cement, and iron/steel, each are infrastructures that are important to economic development of cities. Thus we recommend these infrastructures to be allocated to cities based on "use" to a city's residential commercial industrial sectors. Objective 2 : Using our typology degree r eveals that larger cities tend to approach trade balanced. Thus the world s mega cities may be trade balanced as well in which case TBIF could be used to measure the GHG emissions footprint of these cities It s also s how n that perhaps the ratio of commercial i ndustrial energy use to household energy use can serve as a fast approach for measuring the typology of city. Objective 3 : The size and nature of the supply chains serving cities shows that Territorial GHGs can be as small as 37 % of the total footprint for net co nsumer cities, and as large as 6 8 % for net producers. TBIF is steady capturing between 62% 68 % of the total footprint, while CBF depends on city typology, and captures between 35 % 71 % of the total footprint. Objective 4 : Through an evaluation of ten metrics each compared with an urban energy efficiency index (UEI), we show that Territorial GHG is not suitable for comparing GHG associated with cities by any metric. However, TBIF by GDP is suitable for representing

PAGE 117

! 102 the urban efficiencies of cities, and should be used when comparing production based GHGs. On the other hand, CBF by capita should be used when comparing consumption based GHGs

PAGE 118

! 103 Supplementary Information (SI) Chapter 4 S 4 1 Commercial Industrial Energy Use Efficiencies for 21 cities Commercial Industrial County kWh/GDP/yr therms/GDP/yr kBTU/GDP/yr NYC, NY 51,750 1,727 349,235 DVRPC, PA/NJ 122,185 4,676 884,403 Miami Dade, FL 130,071 308 474,581 Broward, FL 143,410 507 539,977 Oregon METRO, OR 142,592 4,035 889,988 Philadelphia, PA 127,266 3,956 829,744 Sacramento, CA 90,911 1,931 503,276 Westchester, NY 57,591 2,514 447,802 Multnomah, OR 121,085 3,463 759,421 Snohomish, WA 147,025 3,704 871,973 Denver, CO 80,201 3,916 665,176 Washoe, NV 143,502 3,546 844,146 Sarasota, FL 234,714 1,824 983,190 Collier, FL 150,483 536 567,077 Boulder, CO 120,879 4,042 816,561 Loudoun, VA 135,401 1,474 609,383 Napa, CA 81,067 1,823 458,903 Tompkins, NY 118,094 6,209 1,023,756 Roanoke, VA 175,378 6,165 1,214,728 Broomfield, CO 113,769 2,361 624,259 Routt, CO 170,398 3,495 930,815 US 162,380 7,849 1,338,748

PAGE 119

! 104 S 4 2 Residential Energy Use Efficiencies for 21 cities Residential County kWh/cap/mo therms/cap/mo kBTU/cap/mo NYC, NY 144 14 1,928 DVRPC, PA/NJ 326 19 2,962 Miami Dade, FL 443 3 1,822 Broward, FL 518 0.2 1,788 Oregon METRO, OR 303 11 2,108 Philadelphia, PA 205 21 2,827 Sacramento, CA 276 12 2,148 Westchester, NY 216 19 2,612 Multnomah, OR 337 13 2,412 Snohomish, WA 385 11 2,365 Denver, CO 233 20 2,805 Washoe, NV 282 20 2,979 Sarasota, FL 630 1 2,233 Collier, FL 672 0 2,314 Boulder, CO 349 23 3,481 Loudoun, VA 482 15 3,164 Napa, CA 229 13 2,115 Tompkins, NY 242 14 2,227 Roanoke, VA 534 23 4,093 Broomfield, CO 295 21 3,143 Routt, CO 531 23 4,082 US 378 14 2,665

PAGE 120

! 105 S 4 3 Transportation System Efficiencies for 21 cities VMT County VMT/(res+jobs)/day NYC, NY 5.8 DVRPC, PA/NJ 14.9 Miami Dade, FL 17.8 Broward, FL 19.8 Oregon METRO, OR 15.4 Philadelphia, PA 8.0 Sacramento, CA 17.2 Westchester, NY 19.7 Multnomah, OR 16.0 Snohomish, WA 16.8 Denver, CO 14.2 Washoe, NV 16.0 Sarasota, FL 24.3 Collier, FL 22.6 Boulder, CO 15.9 Loudoun, VA 15.2 Napa, CA 17.0 Tompkins, NY 12.9 Roanoke, VA 20.6 Broomfield, CO 14.1 Routt, CO 19.2 US 20

PAGE 121

! 106 S 4 4 Corresponding Metropolitan Statistical Areas (MSA ) for the 21 US City County's in this analysis. County Metropolitan Statistical Area ( MSA ) NYC, NY New York Northern New Jersey Long Island, NY NJ PA DVRPC, PA/NJ Philadelphia Camden Wilmington, PA NJ DE MD Miami Dade, FL Miami Fort Lauderdale Pompano Beach, FL Broward, FL Miami Fort Lauderdale Pompano Beach, FL Oregon METRO, OR Portland Vancouver Beaverton, OR WA Philadelphia, PA Philadelphia Camden Wilmington, PA NJ DE MD Sacramento, CA Sacramento -Arden Arcade -Roseville, CA Westchester, NY New York Northern New Jersey Long Island, NY NJ PA Multnomah, OR Portland Vancouver Beaverton, OR WA Snohomish, WA Seattle Tacoma Bellevue, WA Denver, CO Denver Aurora Broomfield, CO Washoe, NV Reno Sparks, NV Sarasota, FL Bradenton Sarasota Venice, FL Collier, FL Naples Marco Island, FL Boulder, CO Denver Aurora Broomfield, CO Loudoun, VA Washington Arlington Alexandria, DC VA MD WV Napa, CA Napa, CA Tompkins, NY Ithaca, NY Roanoke, VA Roanoke, VA Broomfield, CO Denver Aurora Broomfield, CO Routt, CO N/A. Assumed state averages where needed

PAGE 122

! 107 S 4 5 For 21 US cities, household consumption is 70% of total final consumption, on average. CITY Total HH Total Other FC %HH of FC NYC $372,897 $105,758 78% DVRPC $211,392 $78,646 73% Miami Dade $75,662 $33,215 69% Broward $63,307 $24,845 72% METRO $53,656 $23,142 70% Philadelphia $46,957 $18,715 72% Sacramento $42,152 $36,407 54% Westchester $44,289 $14,073 76% Multnomah $25,668 $12,103 68% Snohomish $20,275 $9,889 67% Denver $26,139 $17,107 60% Washoe $15,656 $5,587 74% Sarasota $15,439 $4,778 76% Collier $15,952 $4,361 79% Boulder $10,554 $8,971 54% Loudoun $9,958 $5,636 64% Napa $4,845 $2,361 67% Tompkins $2,999 $1,189 72% Roanoke $3,136 $795 80% Broomfield $1,526 $713 68% Routt $762 $649 54% Average = 70%

PAGE 123

! 108 S 4 6 Additional correlations between GDP and Use/Exports Remaining Infrastructure Correlations (not shown above) Figure S 4 6a : Air Travel GDP vs. GHGs from Use (left) and Export (right). Figure S4 6b : Cement G DP vs. GHGs from Use (left) and Export (right).

PAGE 124

! 109 Figure S4 6c : Freight GDP vs. GHGs from Use (left) and Export (right). Figure S4 6d : Natural Gas GDP vs. GHGs from Use (left) and Export (right). Figure S4 6e : Water/WW GDP vs. GHGs from Use (left) and Export (right).

PAGE 125

! 110 Figure S 4 6f : Transport Services GDP vs. GHGs from Use (left) and Export (right). Remaining Non Infrastructure Correlations (not shown above) Figure S4 6g : Mining GDP vs. GHGs from Use (left) and Export (right). Figure S4 6h : Construction GDP vs. GHGs from Use (left) and Export (right).

PAGE 126

! 111 Figure S4 6i : Food Mfg. GDP vs. GHGs from Use (left) and Export (right). Figure S4 6j : Alc Bevs/Tob GDP vs. GHGs from Use (left) and Export (right). Figure S4 6k : Other Textiles GDP vs. GHGs from Use (left) and Export (right).

PAGE 127

! 112 Figure S4 6l : Apparel GDP vs. GHGs from Use (left) and Export (right). Figure S4 6m : Industry/ Mfg. GDP vs. GHGs from Use (left) and Export (right). Figure S4 6n : Electronics GDP vs. GHGs from Use (left) and Export (right).

PAGE 128

! 113 Figure S4 6o : Durable Goods GDP vs. GHGs from Use (left) and Export (right). Figure S4 6p : F urniture GDP vs. GHGs from Use (left) and Export (right). Figure S4 6q : Wholesale/Retail GDP vs. GHGs from Use (left) and Export (right).

PAGE 129

! 114 Figure S4 6r : Communications GDP vs. GHGs from Use (left) and Export (right). Figure S4 6s : Education GDP vs. GHGs from Use (left) and Export (right). Figure S4 6t : Recreation GDP vs. GHGs from Use (left) and Export (right).

PAGE 130

! 115 Figure S4 6u : Rest/Hotels GDP vs. GHGs from Use (left) and Export (right). Figure S4 6v : Govt Services GDP vs. GHGs from Use (left) and Export (right).

PAGE 131

! 116 S 4 7 Energy Use Benchmarks for 21 cities. Local and [STATE] Commercial Industrial Residential Road Transport County kWh/job/mo therms/job/mo kWh/HH/mo therms/HH/mo VMT/cap/day Sacramento, CA 771 [918] 16 [59] 748 [580] 33 [34] 2 5.3 [25.2] Napa, CA 719 [918] 16 [59] 714 [1,071] 25 [24] 2 6.0 [25.2] Boulder, CO 1,156 [1,214] 39 [94] 852 [743] 56 [58] 2 4.2 [28.2] Broomfield, CO 1,220 [1,222] 25 [88] 825 [768] 60 [59] 2 2.1 [27.6] Denver, CO 948 [1,222] 46 [88] 546 [768] 47 [59] 25.3 [27.6] Routt, CO 1,470 [1,214] 30 [94] 1,221 [743] 52 [58] 32.3 [28.2] Collier, FL 1,300 [1,187] 5 [13] 1,780 [1,354] 1 [2] 32.3 [30.9] Sarasota, FL 981 [1,173] 8 [13] 1,403 [1,367] 2 [2] 35.3 [31.0] Broward, FL 1,188 [1,187] 4 [13] 1,352 [1,354] 1 [2] 28.5 [30.9] Miami Dade, FL 1,194 [1,173] 3 [13] 1,267 [1,367] 9 [2] 24.9 [31.0] Washoe, NV 1,145 [1,538] 28 [29] 700 [1,022] 50 [34] 23.6 [21.8] Tompkins, NY 799 [892] 42 [37] 564 [554] 33 [46] 19.5 [18.9] Westchester, NY 666 [966] 29 [37] 589 [575] 51 [47] 28.3 [19.7] Multnomah, OR 1,066 [1,423] 30 [49] 793 [1,092] 30 [25] 26.1 [24.2] Philadelphia, PA 1,181 [1,390] 37 [50] 507 [851] 53 [35] 11.0 [23.8] Roanoke, VA 1,202 [1,495] 42 [33] 1,261 [1,247] 54 [23] 28.9 [29.1] Loudoun, VA 1,388 [1,495] 15 [33] 1,472 [1,247] 46 [23] 22.3 [29.1] Snohomish, WA 1,200 [1,513] 30 [36] 994 [1,114] 27 [25] 22.5 [24.3] Oregon METRO 1,212 [1,425] 34 [49] 714 [1,071] 25 [24] 23.9 [26.4] NYC 772 [892] 26 [37] 374 [554] 37 [46] 8.4 [18.9] DVRPC 1,203 [1,279] 46 [52] 842 [851] 48 [50] 21.7 [23.7] US 1,450 70 982 36 27.0 State energy use data retrieved from ( EIA, 2011 ) ; State employment statistics from ( Census, 2011 ) ; State population and households from ( Census, 2011 ) ; State VMT from ( FHWA, 2008 )

PAGE 132

! 117 5. I nternational Application Community wide greenhouse gas (GHG) accounting is confounded by the relatively small spatial size of cities compared to nations, due to which: energy use in essential infrastructures serving cities, e.g., commuter and airline trans port, energy supply, water supply, wastewater infrastructures, etc., often occurs outside the boundary of the cities using them. A Trans Boundary Infrastructure Supply Chain Footprint (TBIF) GHG emissions accounting method, tested in 8 US cities, incorpora tes supply chain aspects of these trans boundary infrastructures serving cities, and is akin to an expanded geographic GHG emissions inventory, covering Scopes 1+2+3. This chapter shows results from applying the TBIF in the rapidly developing city of Delhi India. The objectives of this research are to 1) describe data availability for implementing the TBIF within a rapidly industrializing country, using the case of Delhi, India, 2) identify methodological differences in implementation of the TBIF between I ndian versus US cities, and 3) compare broad energy use metrics between Delhi and US cities, demonstrated by Denver, Colorado USA whose energy use characteristics and TBIF GHG emissions have previously been shown to be similar to US per capita averages. Th is research concludes that most data required to implement the TBIF in Delhi are readily available, and the methodology could be translated from the US to Indian cities. Delhi's 2009 community wide GHG emissions totaled 40.3 million mt CO 2 e, which are normalized to yield 2.3 mt CO 2 e/capita. Nationally, India reports its average per capita GHG emissions at 1.5 mt CO 2 e/capita. In boundary GHG emissions contributed to 68% of Delhi's total, where end use (including electricity) energy in residential buildin gs, commercial/industrial, and fuel used in surface transportation, contributed to 24%, 19%, and 21%, respectively. The remaining 4.3% in boundary GHG emissions were from waste disposal, water/wastewater (WW) treatment, and cattle. Trans boundary infrastru ctures were estimated to equal 32% of Delhi's TBIF GHG emissions, with 5% attributed to fuel processing, 3% to air travel, 10% to cement, and 14% to food production outside the city.

PAGE 133

! 118 5 .1 Introduction Cities are increasingly recognizing their role in globa l greenhouse gas (GHG) emissions. Over one thousand cities have signed onto the ICLEI Cities for Climate Protection (CCP) a framework for engaging local governments in political climate action commitments ( ICLEI, 2009 ) and the Mexico City Pact (MCP) an agreement by more than 140 world mayors to establish GHG emissions inventories and mitigation plans ( WMSC, 2010 ) Outcomes from these efforts include public domain items such as the carbonn # Cities Climate Registry (cCCR) a voluntary online tool where cities are reporting o n their GHG inventories and mitigation commitments ( cCCR, 2010 ) However to date, these tools while very valuable have not incorporated trans bounda ry GHG emissions associated with human activities in cities which have been shown to be quite significant ( Hillman & Ramaswami, 2010 ; Kennedy, et al., 2009 ; Ramaswami, et al., 2008 ) Understanding GHG emissions ass ociated with cities in India, China, and US is important due to their contribution to world totals. A report by the International Energy Agency (IEA) notes that India, China, and US together, constitute 42% of the world's population, and 46% of the world's CO 2 from fuel combustion ( IEA, 2010 ) Moreover, in India, China, and US, 30%, 44%, and 82% of people live in urban areas, respectively ( TWB, 2010 ) With rapid urbanization seen especially in Indian and Chinese cities, quantification of GHG emissions associated with cities becomes important. G HG emissions accounting for cities however, is confounded by the relatively small spatial size of cities compared to nations, due to which: Essential infrastructures commuter and airline transport, energy supply, water supply, wastewater infrastructures, etc. cross city boundaries, hence energy use to provide these services often occurs outside the boundary of the cities using them ( Hillman & Ramaswami, 2010 ; Ramaswami, et al., 2008 ) Significant trade of other goods and services also occurs across cities, with associated embodied GHGs. Two approaches to GHG emissions footprinting (see review in ( Chavez & Rama swami, 2011 ; Ramaswami, et al., 2011 ) ) can be used to alleviate these challenges. The two

PAGE 134

! 119 approaches are the Trans Boundary Infrastructure Supply Chain Footprint (TBIF), and the Consumption Based Foo tprint (CBF). TBIF utilizes the concept of scopes from corporate GHG emissions accounting protocols to include both in boundary and trans boundary GHG emissions associated with city activities; hence it has also been referred to as an expanded geographic i nventory. TBIF recognizes that cities include both producers and consumers, and focuses on infrastructure supply chains that serve the entire community as a whole. The GHG emissions accounted for by the TBIF are a) direct in boundary (Scope 1), b) indirect GHG emissions from generation of purchased electricity (Scope 2), and c) GHG emissions from trans boundary infrastructures serving cities (Scope 3), such as water supply, transportation fuels, airline and commuter travel, and other critical supply chains. Inclusion of trans boundary infrastructures (Scope 3) warrants careful allocation of GHG to avoid double counting ( Ramaswami, et al., 2008 ) For example, infrastructures such as large electric power plants, or oil refineries are easily recognized within city boundaries, and their GHG can be readily allocated based on local demand, thus reducing double counting. TBIF considers the community as a whole, maintaining residential, commercial, and industrial activities together, consistent with the geopolitical definition. Although TBIF captures life cycle GHGs from essential infr astructures serving cities, it does not account for life cycle GHGs of other, non infrastructure goods/services consumed by households, or other non infrastructure supply chains serving local industries because such data are often proprietary. Indeed, inco rporating key industrial supply chains to the TBIF can enhance this method because TBIF includes both consumers and producers in cities. Improved blended metrics that combine GHG/capita and GHG/productivity may be needed. The second approach is a consumpti on based footprint (CBF), which quantifies the full life cycle GHG emissions from economic final consumption in a city defined as household expenditures, government expenditures, and business capital investments. CBF have traditionally been conducted at th e scale of households, using household consumer expenditure surveys (CES) ( Jones & Kammen, 2011 ) with regional/national production matrices, coupled wi th sector specific GHG emission intensities (e.g., Lenzen

PAGE 135

! 120 & Peters 2010). Recent efforts have been made to compute city scale CBF using final consumption vectors reported in sub national input output (IO) tables ( Stanton, et al., 2011 ) While CBF incorporates all trans boundary GHG relating to local household consumption, local production of exports is allocated out. Such allocation alters the definition of a community, w here the geopolitical unit is split in two: local final consumption sectors, and local producers who export goods elsewhere. Both approaches have their advantages and disadvantages, and neither is complete, in that neither fully accounts for all life cycle supply chains serving both producers and consumers in cities. TBIF accounts for life cycle GHGs of essential infrastructures serving cities, but does not account for life cycle GHGs of all other, non infrastructure goods/services consumed by households or those used in industrial production. Also, TBIF recognizes that both, city's production and consumption activities are intrinsically linked, and focuses on publicly managed cross scale infrastructures such as commuter travel, airline travel, freight, and energy & water supply chains that transcend city boundaries and serve the entire community as a whole. In contrast, CBF ignores the in boundary and supply chain impacts of commercial industrial activities that are exported, focusing only on consumption and its supply chains. The utility of TBIF has been described in Ramaswami et al. (2011). In summary, TBIF can be used to quantify a community's GHG emissions by addressing direct energy use and also embodied energy in infrastructures. The method keeps a comm unity's energy use together (residential and business activity), quantifying community GHG emissions as a whole. The method can link in boundary energy use and GHG emissions, to local air pollution and local health impacts, and is able to track the effects stemming from infrastructure policies across scale address buildings energy supply, transportation, water/WW, and waste. By its trans boundary inclusions, TBIF addresses regional cross sector and cross scale infrastructure efficiencies, such as mass trans it, or expanded tele presence aimed at reducing air travel. Lastly, supply chain vulnerabilities impacting local economies as a whole are addressed. ICLEI USA has gathered a group of technical leaders from business, government, and academia to develop a dr aft community scale GHG emissions accounting and reporting

PAGE 136

! 121 protocol ( ICLEI, 2011 ) The protocol recognizes and seeks to address the need for standardized GHG emissions accounting of cities. Four reporting approaches are defined in the draft protocol framework Basic, Expanded, Local Government Focus Area, and Consumption Based. Both, the Basic Reporting Standard (Basic) and Expanded Community Impact Reporting (Expanded) are derived from TBIF. See ([\]^! C 6 $ for a full description of Basic and Expan ded reporting standards. The main objective of this chapter i s to evaluate the TBIF using Delhi, India as the case study. More specifically, this chapter 1) describes data availability for implementing the TBIF within Delhi, a rapidly industrializing city, 2) identifies methodological differences between the implementation of TBIF in Indian versus US cities, and 3) compares broad energy use metrics between Delhi and US cities, demonstrated by Denver whose TBIF per capita has been shown to be similar to US a verages. Table 5 1 : Basic and Expanded reporting frameworks for ICLEI USA community scale GHG emissions accounting and reporting protocol. Methodology Use in ICLEI USA Draft Protocol Trans Boundary Infrastructure Supply Chain Footprint (TBIF) (Ramaswami et al., 2008; Hillman & Ramaswami, 2010; Kennedy et al., 2010; Ramaswami et al., 2011) Basic GHG Emissions Reporting Standard In Boundary Contributions. GHG Emissions from : Combustion of stationary sources (natural gas, LPG, Fuel Oil, etc.) Combustion of mobile sources (gasoline, diesel) Power plant emissions for electricity used in community Landfilling of waste generated in community Other industrial processes (e.g. calcination) Suggested Trans Boundary Contributions : GHG emissions associated with production of fuels used in community, including inputs to electric power plants* Expanded GHG Emissions Reporting Standard As in Basic Reporting shown above, PLUS Suggested Trans Boundary Contributions : Origin Destination or one way allocation of transportation (road, air, freight, maritime) Embodied emissions from trans boundary water pumping and water/WW treatment* Embodied emissions from food production* Embodied emissions from cement production* Assumed that in most cities, these activities occur outside of city boundary, hence trans boundary.

PAGE 137

! 122 5 .2 Trans Boundary Infrastructure Footprint (TBIF) Method Description and Data Needs The TBIF accounts for in boundary GHG emissions from buildings (residential, commercial, and industrial), road transportation, industrial processes (i.e. waste emissions, calcination), plus embodied GHG emissions of a city's trans boundary infrastructure supply chains, e.g. electricity supply, fuel production, water/WW treatment, cement production, spatially allocated airline and freight transport, and production of food consumed in the city, see ([\]^! C 6 $ The method has been tested in the US (e.g., Ramaswami et al. 2008; Hillman & Ramaswami 2010), yielding a convergence in per resident GHG emissions from city to national scale for a set of seven larger US cities, suggesting the inclusi on of these selected trans boundary infrastructures generate scale consistency from city to national levels. ([\]^! C 6 illustrates energy and material uses accounted for by the TBIF, along with appropriate benchmarks, and associated emission factors (EF). ([\]^! C 6 = illustrates the data needs for benchmarking energy and material use described by the TBIF. Table 5 2 : TBIF energy & materials use benchmarks, and EFs Activity Sector In Boundary Energy & Materials Use In Boundary Energy & Materials Use Benchmark Associated EF In & Trans Boundary Buildings Energy Use & Industrial Process Emissions Residential, Commercial, Industrial, Government, and industrial processes (e.g. waste, calcination) Energy Use (residential, commercial, industrial, government): Electricity Natural Gas Cooking Fuels (e.g. LPG) Heating Fuels (e.g. Fuel Oil, Propane) Industrial process emissions : Waste (methane generation) Other (industrial emissions) Residential Intensity : kWh/HH/mo m 3 /HH/mo liter CF/HH/mo liter HF/HH/mo Total kBTU/HH/mo Commercial Intensity : kWh/sm c /yr Other stationary fuels kBTU/sm c /yr Total kBTU/sm c /yr Industrial Process : mt of waste/capita/yr In Boundary EF associated with fuel combustion : EF Elec = kg CO 2 e/kWh EF NG = kg CO 2 e/ m 3 EF Cooking Fu els = kg CO 2 e/liter EF Heating Fuels = kg CO 2 e/liter EF Waste = kg CO 2 e/mt waste Trans Boundary EF associated with fuel production : = kg CO 2 e/mt coal = kg CO 2 e/m 3 = kg CO 2 e/liter = kg CO 2 e/liter E F C o a l P r o d L C A E F N G P r o d L C A E F C o o k i n g F u e l s P r o d L C A E F H e a t i n g F u e l s P r o d L C A

PAGE 138

! 123 Table 5.2 (cont.) Transportation Energy Use Road, Air Travel, Freight, and Rail Energy Use : Gasoline, Diesel, and CNG use in road transport Jet Fuel use in air travel Fuel (i.e. Diesel) use in freight transport Diesel use in rail transport Surface Travel Intensity : VKT/capita/day F leet Fuel Efficiency (VKT/liter of fuel) Air Travel : liter of jet fuel/enplaned passenger Rail : Total Person Kilometers Traveled (PKT) PKT/cap/day Total BTU In Boundary EF associated with fuel combustion : EF Gasoline = kg CO 2 e/liter EF Diesel = kg CO 2 e/liter EF Jet Fuel = kg CO 2 e/liter EF Rail = kg CO 2 e/liter Trans Boundary EF associated with fuel production : = kg CO 2 e/liter = kg CO 2 e/liter = kg CO 2 e/liter Materials Use Water, Food, Cement Use of water, food, and cement. Water : treated wastewater (WW) liters/capita pumped water liters/capita Food : mt food/HH Cement : mt cement/capita In Boundary EF associated with materials : Logic Rules Applied Accordingly # Trans Boundary EF associated with materials production : EF WW = mt CO 2 e/volume treated WW EF water = mt CO 2 e/volume pumped water EF food = mt CO 2 e/mt food EF cement = mt CO 2 e/mt cement City Wide Total local population (capita) Total city area (sq km) Population Density (capita/sq km) Total homes (HH) People per home (capita/HH) Residential floor area (sm r /HH) Total commercial floor area (sm c ) Total floor area per capita (sm/cap) City GDP Emission intensity per unit GDP (GHG/GDP) Emission intensity per resident (GDP/cap) Number of jobs Cities are unique in that most have these trans boundary GHG emissions occurring outside of the community. As in boundary data of these activities become available, the energy use and GHG emissions should be updated accordingly. # These large infrastruc tures are mostly absent in US cities. GHG from infrastructures allocated based on local demand, and are allocated out to avoid double counting. VKT = Vehicle Kilometer Traveled. sm r = residential square meters. sm c = commercial square meters. sm = total s quare meters. HH = Households. GDP = Gross Domestic Product. E F G a s o l i n e P r o d L C A E F D i e s e l P r o d L C A E F J e t F u e l P r o d L C A

PAGE 139

! 124 Table 5 3 : Data needs for benchmarking the TBIF to represent energy & materials use. Activity Sector Data Needs Buildings Energy Use & Industrial Process emissions Total residential floor area (sm r ) Total Commercial floor area (million sm c ) Total floor area per capita (sm/capita) Residential Electricity, Natural Gas, Cooking Fuels, and Heating Fuels use Commercial Industrial Gov ernment Electricity, Natural Gas, Other fuel use Total waste generated in city Transportation Energy Use Allocated daily VKT (VKT/cap/day) Fleet fuel efficiency Volume of Gasoline, Diesel, and CNG used in road transport Number of enplaned passengers at regional airport (Domestic, International) Jet Fuel liters loaded into airplanes % of planes fueling at airport Tons of Long Distance Freight, a nd liters of fuel per ton moved Energy used in Rail transport Materials Use Volume of water used (i.e. pumped) Energy used in pumping water Volume of wastewater treated Energy used in WW treatment % of water used for Residential, Commercial, and Industrial uses Food consumed/used in the community Cement use in the community 5 .3 Socio Economic Profile and Overview of Energy Use & GHGs for Delhi, India India's national population is estimated at 1,155 million people ( TWB, 2010 ) corresponding to about 17% of the world's population. India's GDP is $3,275 billion, roughly 3% of the world's GDP ( TWB, 2010 ) and total primary energy use is estimated at 20 Quad BTU ( EIA, 2010 ) about 4% of the world's total primary energy use. India's annual growth in primary energy use and GDP are 7% and 8.2%, respectively, relative to tren ds for the US of 0.3% and 2.3%, respectively. In Delhi, even greater GDP growth is projected, with annual GDP growth reported at 15.9% ( DES, 2009 ) Delhi is a city state and the capital of India. Home to almost 18 million people, it boasts a vibrant economy which is poised for continued growth. Spurred by an influx of jobs in IT, telecommunication, banking, and manufacturin g, Delhi has become an attractive place for many, generating a per resident GDP that is about twice that of India's ($6,037 vs. $2,835 PPP US$ 2009), see ([\]^ C 6 ; Th e Delhi government is also initiating a wide

PAGE 140

! 125 range of sustainable infrastructure programs addressing energy use and GHG emissions. Two such examples are the new more stringent building codes (Energy Conservation Building Code (ECBC) implemented in 2009, ( DDE, 2009 ) ), and fuel switching of all commercial fleets from gasoline/diesel to natural gas (Indian Supreme Court legislation in 1998, ( Mehta, 2001 ) ), both of which can help reduce carbo n emissions per GDP. Previous research has contributed to some level of energy use and GHG emissions accounting for Delhi. The earliest known GHG emissions research in Delhi was conducted for the baseline year of 1995 by Sharma et al. (2002a d). That research evaluated GHG emi ssions from use of electricity, natural gas, LPG, kerosene, gasoline, and diesel, plus the embodied emissions associated with the production of cement, steel, rice, and milk used in Delhi (Sharma et al. 2002a d). A more recent study inventoried Delhi's 200 7 GHG emissions from in boundary activities only ( Ghosh, 2009 ) Table 5 4 : Comparisons of key demographic and economic variables in USA, India, and Delhi. 2009 U.S. India Delhi Population (million) a 307 1,155 17.6 Annual % change 0.93% 1.4% 2.9% % Urban a 80.8% 29.8% 93.2% % Rural a 19.2% 70.2% 6.8% GDP (billion USD Real); {billion USD PPP} b $14,119 ($1,310); {$3,275} ($42.5); {$106} Annual % change 2.3% 8.2% 15.9% GDP/capita (USD Real/cap); {USD PPP/cap} $45,989 ($1,134); {$2,835} ($2,415); {$6,037} Annual % change 1.4% 8.3% 12.6% Income/capita (USD/cap) c $40,947 $833 $1,965 GHG/capita (mt CO 2 e/cap) d 21.6 1.5 2.4 GHG/GDP (mt CO 2 e/mill $GDP) d, 482 1,317 948 Primary Energy (EJ) e 104 21 0.53 Annual % change 0.3% 7% a. Population statistics sources: U.S. and India = The World Bank (2010); Delhi = DCO (2009) b. Gross Domestic Product sources: U.S. and India = The World Bank (2010); Delhi = DES (2009) c. Per capita income sources: U.S. = BEA (2009); India/Delhi = CSO (2009). d. S ources for GHG estimates: U.S. = EPA (2011); India = MEF (2010); Delhi = Estimated in this study. e. Primary Energy : U.S. and India = International Energy Agency; Delhi = estimated. *. GDP in Real U.S. Dollars (USD).

PAGE 141

! 126 5 .4 Data Sources and Results Indian energy use data are more readily available at the state level versus the city level. Because Delhi is a city state, i.e. considered a union territory among six others in India, energy use data was readily available, which may not be the case for oth er Indian cities. This section first introduces important demographic trends in Delhi. Then, required data sources and their availability for completing the TBIF in Delhi are discussed, with results presented for demographics and then the activity sector c ategories presented above in ([\]^! C 6 and ([\]^! C 6 = : 1) Building Energy Use & Industr ial Processes, 2) Transportation Energy Use, and 3) Materials Use. 5.4.1 Delhi Statistical Handbook A fair amount of the socio demographic and in boundary energy use data for Delhi was obtained through the Delhi Statistical Handbook (DSH), published by the Directorate of Economics & Statistics of the Delhi Government ( DES, 2010 ) Data reported in t he DSH ranges from various economic, demographic, and health parameters, to e lectricity use d in Delhi The edition of the DSH used in this thesis is number 35, and as in all previous issues the DSH depends on primary data supplied by various agencies /ministries thus serving as a conduit for large amounts of data relating specifically to the National Capital Territory ( NCT ) of Delhi. Socio Demographic data retrieved from the DSH and used in this study is population, hous eholds, population density, employment, and GDP, all of which have been supplied by the Dir ectorate of Census Operations ( DCO, 2009 ) Energy use d ata retrieved from the DSH and used in this study is electricity, supplied by the Delhi Electricity Regulatory Commission ( DERC, 2009 ) and the end use of other fuels such as LPG and Kerosene, supplied by the Ministry of Petroleum and Natural Gas ( MPNG, 2009 ) C attle head counts have also been retrieved from the D SH, and supplied by the D irectorate of Animal Husbandry ( DAH, 2010 ) Methodological details pertaining to data collection from the respective ministries have proven to be sparse and difficult to obtain For example, i t is unknown whether energy use data reported through the DSH has been collected through utility sales and revenue

PAGE 142

! 127 shares, or whether it has been estimated throu gh surveys representing average use among sector wise users 5 .4. 2 Demographics Population data for Delhi was obtained from the Directorate of Census Operations (DCO) through the DSH ( DCO, 2009 ) which reports Delhi's 2009 population equal to 17.6 million people. Household counts in Delhi were last reported by DCO in 2001. Thus, using home occupancy as reported in 2001 (4.6 people/HH), we estimated Delhi's homes corresponded with 3.8 million homes. Two estimates of population density were obtained; the first was a 2001 ( DCO, 2009 ) estimate, equal to 9,340 people/sq km, and the second was a 2007 est imate equal to 11,463 people/sq km ( UN, 2010 ) Delhi employment statistics, which were last reported for 2001 ( DCO, 2009 ) illustrate the annualized employment growth from 1981 1991 and 1991 2001 are almost identical, equal to 5% per annum. Applying the assumption of constant employment growth from 2001 2008 yielded an estimated 6 .8 million jobs in Delhi, in 2009. Floor areas for residential, commercial, and industrial units in Delhi were not locally available. A literature search yielded national estimates of average urban residential floor areas equal to 46.8 sq meter/HH ( TOI, 2008 ) and an aggregate India commercial floor area was reported equal to 516 million sq meter ( Satish Kumar, Kapoor, Deshmukh, Kamath, & Manu, 2010 ) While assuming that commercial activity occurs in urbanized places, commercial floor areas were apportioned to Delhi by urbanized population, resulting in an estimate for Delhi equal to 25.7 million sq meter. Industrial floor space is typically difficult to quantify in any community, and was unattainable in Delhi. 5 .4. 3 Buildings Energy Use and Industrial Process Emissions Sector wise electricity use in 2009 was reported by DERC ( DERC, 2009 ) Unlike the US where natural gas is a dominant energy carrier second to electricity, building electricity use in Delhi is followed by a series of ot her fuels that serve end use needs of the

PAGE 143

! 128 community including liquid petroleum gas (LPG), kerosene, and compressed natural gas (CNG). Use of these other fuels were obtained from the Ministry of Petroleum and Natural Gas ( MPNG, 2009 ) and apportioned to end use sectors using ratios previously estimated for Delhi ( Ghosh, 2009 ) (e.g., LPG: 95.9% Residential, 3.5% Commercial, 0.6% Industrial) 5 .4. 3 .1 Re sidential Energy Use Benchmarks Several factors have been shown to contribute to household energy use in India, some of which include home size, home construction material, income, and climatic/weather conditions ( Pachauri, 2004 ; Pachauri & Jiang, 2008 ) Pachauri (2004) notes that on average, direct ener gy use of urban Indian households is two to three times greater than rural households. Electricity use was the dominant end use energy source for Delhi households in 2009, and its monthly use by households is estimated to have been 191 kWh/HH/month ( DERC, 2009 ) Nationally, Indian households use 48 kWh/HH/month ( IEA, 2008a ) This difference in average household electricity use between Delhi and India is in line with Sharma et al. (2002a) who estimated Indian urban electricity use to be is about three times higher than national averages. Delhi households typically do not use natural gas or other fuels ( e.g., propane) for space heating as is done in the US, but do use LPG and kerosene for cooking; any coal or biomass use for cooking was not reported in this research The estimated monthly use of each of the two fuels are LPG = 25.3 liters/HH and Kerosene = 3.4 liters/HH ( MPNG, 2009 ) This compares to 7.8 liters/HH and 4.2 liters/HH, respectively, with national statistics ( IEA, 2008a ) Combining these end uses of energy yields an energy end use intensity (EUI) of Delhi residences, estimated at 1,489 MJ/HH/mo. India's household EUI is reported at 273 MJ/HH/mo ( IEA, 2008a ) These estimates roughly conform to estimates by Pachauri (2004) who notes that urban household energy use is at least triple that of national averages.

PAGE 144

! 129 5 .4. 3 .2 C ommercial Energy Use Benchmarks Com mercial buildings energy use consisted of electricity, LPG, CNG, and diesel. Electricity use, excluding use in treating/pumping water and wastewater, was equal to 5,795 million kWh ( DERC, 2009 ) or 225 kWh/sm c /yr using estimated commercial floor areas from national reports. Nationally, ( Gupta, 2011 ) estimates average commercial electricity use intensity equals 189 kWh/sm c /yr, while estimates provided by ( IEA, 2010 ) reports 93.6 kWh/sm c /yr for India. Total end use of the other fuels in commercial buildings are: LPG = 43 million liters, CNG = 30.6 million cubic meters, and Diesel = 15.8 million liters ( MPNG, 2009 ) Combining these energy end uses yields a EUI for Delhi's commercial buildings equal to 923.8 MJ/sm c /yr. 5 .4. 3 .3 I ndustrial Energy Use Benchmarks Energy statistics report industries in Delhi used 2,991 million kWh in 2009 ( DERC, 2009 ) Other energy end uses by Delhi industries are LPG = 6.9 million liters, CNG = 46.4 million cubic meters, High Speed Diesel (HSD) = 5.9 million liters, Light Diesel Oil (LDO) = 2.1 million liters, and Diesel = 3 million liters ( MPNG, 2009 ) 5 .4. 3 .4 Industrial Process Benchmarks The Delhi Pollution Control Committee (DPCC) estimates Delhi generates 7,310 tonnes of municipal solid waste (MSW) daily ( DPCC, 2010 ) amounting to about 0.16 tonnes/resident/yr, which compares to 0.14 tonnes/resident/yr nationally ( Sharholy, Ahmad, Mahmood, & Trivedi, 2008 ) About 7% of Delhi's waste is diverted in the form of compost. Additionally, there are three on going waste to energy projects in Delh i that promise to divert close to 15% of today's MSW ( DPCC, 2010 ) Releases of untreated wastewater can also be a source of considerable GHG emissions. R ivers, lakes, lagoons, etc., provide anaerobic conditions for untreated wastewater,

PAGE 145

! 130 resulting in methane (CH 4 ) and nitrous oxide (N 2 O) production. It is estimated that Delhi captures and treats 63% of its total produced wastewater ( MUD, 2010 ) Noting that Delhi treated 1,584 million liters of wastewater per day in 2009 ( MUD, 2010 ) we estimate the 2009 releases of untreated wastewater total 339,633 million liters. Among the other industrial processes recognized by the IPCC as contributors to GHG emissions, cement production is the most prominent ( IPCC, 2006a ) The Cement Manufacturers Associati on (CMA) reports no cement production within the boundaries of Delhi, thus providing a basis for incorporating cement as a relevant Scope 3 item. No other industrial process emissions were readily identified within Delhi boundaries. 5 .4. 3 .5 Emissions Fact ors Electricity EF Electricity is generated in Delhi at five power plants; three coal powered and two natural gas powered power plants. Their EF, in kg CO 2 e/kWh are 1.16, 1.52, 1.39, 0.59, 0.36, respectively ( Ghosh, 2009 ) Nationally, India has two power grids. The first grid is the Integrated Northern, Eastern, Western and North Eastern (NEWNE), which has an EF equal to 0.83 kg CO 2 e/kWh. The second is the Southern grid, whose EF is equal to 0.76 kg CO 2 e/kWh. This results in a blended national electricity EF equal to 0.82 kg CO 2 e/kWh ( CEA, 2009 ) previously reported to consist of 90% coal, with the remining 10% being natural gas, oil, and wind ( MEF, 2010 ) The national electricity EF includes transmission and distribution (T&D) losses (including unauthorized connections), which have been estimated to equal ab out 24% across India ( TWB, 2010 ) Because the NEWNE regional grid serves Delhi, its electricity EF was used upon the recommendation of ICLEI SA. Fuel EF Production and Combustion The combustion EFs of fuels used in buildings were obtained from the 2007 national India inventory ( MEF, 2010 ) and are consistent with IPCC 2006. The EF for fuel combustion are: NG = 2.15 kg CO 2 e/cubic meter, LPG = 1.68 kg CO 2 e/liter, and

PAGE 146

! 131 kerosene = 2.7 kg CO 2 e/liter. Production EF of LPG and kerosene were adopted from ( Lewis, 1997 ) since no India specific data was identified. Those production EFs are reported as: LPG = 0.26 kg CO 2 e/liter, and kerosene = 0.22 kg CO 2 e/liter. MSW EF EF from waste landfilling is estimated using IPCC's default methodology ( IPCC, 2006b ) : C H 4 = M S W T M S W F M C F D O C D O C F F 1 6 1 2 R # $ % & ( 1 O X ( ) Equation 5 1 where MSW T is the total waste generated, MSW F is the fraction sent to landfills, MCF is the methane correction factor, DOC is the degradable organic carbon, DOC F is the fraction of DOC dissimilated, F is the fraction of CH 4 in landfill gas with a default value of 0.5, R is the recovery of CH 4 and OX is the oxidation factor with a default value of 0. For the variables requiring specific data relating to Delhi's waste composition, namely MCF, DOC, and DOC F we turn to the lit erature. Both ( Sharma, Dasgupta, & Mitra, 2002b ) and ( S Kumar, Mondal, Gaikwad, Devotta, & Singh, 2004 ) estimate MCF and DOC at 0.4 and 0.15, respectively. However their estimates of DOC F differ, where ( Sharma, et al., 2002b ) report 0.5, and ( S Kumar, et al., 2004 ) report 0.77. Upon substituting these variables into Equation 5 1 we estimate the range of Delhi's EF from landfilling as 0.4 to 0.6, kg CO 2 e/kg waste landfilled and used the average of the two. CH 4 and N 2 O from released WW EF relating to Methane (CH 4 ) a nd Nitrous Oxide (N 2 O) production from released untreated wastewater is consistent with IPCC methodology. The EF for describing the methane production is: E F R i v e r i n e C H 4 = C i n f l u e n t C O D B 0 M C F G W P C H 4 Equation 5 2

PAGE 147

! 132 where, C influent COD is the concentration of chemical oxygen demand (COD) in the influent treated wastewater, which for Delhi has been estimated from ( DPCC, 2010 ) as an average of all Delhi treatment plants, equal to 407 kg COD/million liter. B o is the maximum CH 4 producing capacity, and its default value is 0.25 kg CH 4 /kg COD. MCF is the methane correction factor for rivers and lakes, and its default value is 0.1 GWP CH4 is the methane global warming potential (GWP), equal to 24 kg CO 2 e/kg CH 4 Multiplying the four terms yields 244 kg CO 2 e/million liter of methane from Delhi's untreated released wastewater. The EF for nitrous oxide from untreated wastewater relea ses is adapted from a PNAS study by ( Beaulieu et al., 2011 ) and is written as: E F R i v e r i n e N 2 O = C i n f l u e n t N 2 O E F D e n N 2 O G W P N 2 O Equation 5 3 where, C influentN2O is the concentration of inorganic nitrogen in the influent wastewater, and because Delhi specific data was not available, we assumed the concentration equal's that of another Indian city, Hyderabad, which has been estimated as 52 kg N/million liter ( Miller, 2011 ) EF Den,N2O is the default value of N 2 O emissions from nitrification and denitrification in rivers, 0.005 kg N 2 O/kg N (IPCC 2006). GWP N2O is the nitrous oxide GWP, equal to 298 kg CO 2 e/kg N 2 O. Multiplying the three terms yields 77 kg CO 2 e/million liter of nitrous oxide from Delhi's untreated released wastewater. 5 .4. 4 Transportation Energy Use 5 .4.4 .1 Surface Travel Benchmarks Estimating energy use and GHG emissions from road transport can be challenging in US cities due to the trans boundary movement of vehicles across multiple cities in a commuter shed. For example, in the Denver region consisting of 10 cities (including Denver), 59% of workers commute into Denver, and 33% of Denver residents travel outside for work ( DRCOG, 2007 ) In this study, because Delhi is a mega city, we can

PAGE 148

! 133 assume the administrative boundaries of Delhi and the commuter shed overlap, which significantly simplifies the analysis. The assumption was confirmed by finding that only 3% of Delhi's VKT are trans bo undary, see below. The latest estimates of Delhi's in boundary VKT were obtained from the Central Road Research Institute (CRRI), and are reported at 151 million daily VKT for 2009 ( CRRI, 2009 ) yielding 8.8 VKT/resident/day. The CRRI study also estimated daily vehicle counts entering and leaving Delhi as 431,246 (inbound), and 464,183 (outbound). To estimate the proportion of VKT associated with trans boundary traffic, the average Delhi vehicle trip length, estimated to be 10 km ( IDFC, 2010 ) was applied to eit her inbound or outbound traffic, therefore estimating that only the equivalent of 3% of Delhi's in boundary VKT crosses the city boundary. Thus we hypothesize that in mega cites, VKT's attributed to trans boundary traffic may be negligible due to the high amounts of concurrent in boundary traffic. The other critical component of the vehicular benchmark is fuel efficiencies of vehicles in Delhi. Because data on fuel efficiencies is not currently collected by any Indian government agency ( Roychowdhury, Chattopadhyaya, Sen, & Chandola, 2008 ) estimates of tailpipe emissions from the Automotive Research Association of India ( ARAI, 2007 ) were used as the basis by ( Arora, Vyas, & Johnson, 2011 ) in estimating fuel efficiencies of Indian vehicles (see ([\]^! C 6 C ). We then coupled fuel efficiencies by vehicle type with Delhi's VKT to estimate fuel used in road transport. Upon allocating fuel use of outbound vehicle trips out, we estimate 2009 fuel use in Delhi road transport equal to; Gasoline = 1,547 million liters, Diesel = 1,128 million liters, and CNG = 692 million cubic meters. Estimates of fuel used in road transportation shown above were computed from a number of widely cited and trusted organiz ations As fuel efficiencies are essential in our computations, values by ( Arora, et al., 2011 ) used here were verified, and are in line with estimates published b y ( Bose & Sperling, 2001 ) The aggregate fuel use values above were compared to th ose published by the MPNG which are strictly survey based, as fue l efficiencies are not tracked in India T he ministry reports uses of Gasoline = 1,027 million liters, and Diesel = 1,214 million lite rs, in Delhi for 2009. A number of possible

PAGE 149

! 134 sources could trigger the differences. Under reporting by vendors is the first tangible possibility. Another is that MPNG does not disaggregate by fuel end use, where about 25% of India's diesel is used in non tr ansportation services such as industry, power generation, and others ( Singh et al., 2008 ) The VKT approach adopted here is believed to be of higher quality because it uses data from reliable sources. Moreover, the estimates use transportation and energy/health models strictly used for motorized road transport. Table 5 5 : S urface transport fuel use in Delhi. By fuel type and vehicle type. Fuel Type Vehicle Type Daily VKT (million) a Fuel Efficiency (km/L) b Fuel Use (million liters) Total Fuel Use, by Fuel Type Petrol Car Small 31.1 13.3 825 1,547 million liters Car Big 13.5 13.3 358 Two Wheelers 54.7 53.1 364 Diesel Car Small 8.8 13.5 231 1,128 million liters Car Big 11.6 11.9 346 Bus 0.5 3.6 55 Light Commercial Vehicles (LCV) 3.3 5.2 231 Heavy Commercial Vehicles (HCV) 2.4 2.8 266 CNG Car Small 2.1 15.4 c 48.1 d 692 million cubic meters Car Big 0.8 15.4 c 17.8 d Bus 2.2 2.0 c 393.3 d Auto (Rickshaws) 19.6 30.5 c 232.3 d a. Daily VKT in Delhi retrieved from CRRI (2009). b. Average Fuel Efficiencies within Indian fleet, from Arora (2011) and ARAI (2007). c. CNG fuel efficiencies shown in liters per cubic meter. d. CNG fuel use shown in cubic meters. Fuel efficiency is referred to as fuel economy in the US and reported in equivalent units, miles per gallon. 5 .4.4 .2 Air Travel Benchmarks Jet fuel loaded and passenger traffic at Delhi's IGI airport was obtained directly from the airport. Jet fu el loaded in 2010 is reported as domestic travel = 551 million liters, and international travel = 1,214 million liters, and enplaned passengers are reported as 8.7, and 4.0, million passengers, for domestic and international travel, respectively ( DIAL,

PAGE 150

! 135 2011 ) A passenger survey was conducted at IGI to allocate jet fuel loaded to Delhi based on the proportion of outbound passengers at IGI who were associated with activities in Delhi, either as residents, business travelers, tourists leaving, or visitors of Delhi. Survey results show that 25% of domestic passengers, and 47% of international passengers were traveling through Delhi from another town (see ([\]^! C 6 C ). Thus, 76% of domestic passengers, and 53% of international passengers can be deduced to have Delhi related travel, which was used to allocate jet fuel loaded to Delhi. Allocating jet fuel and passengers to Delhi yields 414 million liters, and 644 million liters for domestic and international travel, respectively. Thus resulting in 56 liters/enplaned passenger, and 275 liters/enp laned passenger for domestic and international travel, respectively. Of the total jet fuel loaded at IGI, only the domestic portion was incorporated into Delhi's TBIF as required by international protocols ( DIAL, 2011 ; UNFCCC, 2006 ) 5 .4.4 .3 Rail Travel Benchmarks We used India's national GHG emissions inventory to determine that emissions from railways constitute 0.4% of the country's GHG emissions ( MEF, 2010 ) mostly diesel combustion. A lack of data and the relatively lower importance in terms of total national GHG emissions guided us to ignore GHG emissions from rail in Delhi at this stage. With new local commuter rail being installed in Delhi, future work m ay incorporate GHG from rail by combining energy use of Indian railways ( IRFCA, 2006 ) rail passenger kilometers traveled (PKT), and goods transported by rail ( TWB, 2010 )

PAGE 151

! 136 Table 5 6 : Results from airport survey conducted at the Delhi International Airport. 111 travelers: 52 domestic travel ers, 59 international travelers. Question Answer Choice % Responses Domestic Terminal (n = 52) International Terminal (n = 59) 1. Are you a resident of Delhi? a. Yes 27% 27% b. No 73% 73% 2. If not a resident of Delhi, are you leaving after a a. Business or work related trip in Delhi 35% 5% b. Holiday or other special occasion in Delhi 6% 0% c. Visited friends or relatives in Delhi 6% 15% d. Sightseeing tour/vacation in Delhi 2% 5% e. None of the above: I am just passing through Delhi from another city or town 25% 47% 3. Where did you initiate your trip ? a. My own home 31% 47% b. H otel in Delhi 6% 12% c. Relatives or friends home in Delhi 10% 15% d. Workplace in Delhi 21% 0% e. Drove into Delhi from outside of Delhi 8% 24% f Flew into Delhi from another city/country, and simply flying through Delhi 25% 2% g Other 0% 0% 4. Will you be willing to share the purpose of your trip ? a. Business of work related trip 65% 47% b. H oliday or other special occasion 13% 19% c. Visiting friends or relatives 13% 15% d. Vacation 6% 12% e. Personal 2% 5% f Other 0% 2% 5. Which mode of transport did you use to come to the airport today ? a. Metro 9% 7% b. Government Bus 5% 3% c. Taxi 68% 62% d. My own car 18% 28% 5 .4.4 .4 Emissions Factors The combustion EFs of fuels used in transportation within Delhi are consistent with IPCC 2006, and equal to those used in India's national GHG inventory ( MEF, 2010 ) The EFs from fuel combustion are: Gasoline = 2.4 kg CO 2 e/liter, Diesel = 2.9 kg CO 2 e/liter, and Jet Fuel = 2.7 kg CO 2 e/liter. EF associated with diesel production has been retrieved from ( Whitaker, 2007 ) who estimated an EF from diesel production in India equal to 0.5 kg CO 2 e/liter. Because the distillation temperature of diesel occurs within a similar range to that of jet fuel kerosene, 200 300 ¡C, it was assumed that both diesel and jet fuel have similar production EF, as has been prev iously assumed ( Hillman & Ramaswami, 2010 ;

PAGE 152

! 137 Kennedy, et al., 2009 ) Production EF for gasoline and CNG in India were not attainable, so the following assumptions were made. F or gasoline, because distillation occurs at lower temperatures than that of diesel, the worst case EF was assumed to be equal to diesel ( 0.5 kg CO 2 e/liter ) ; and CNG was assumed to equal the median value of those reported in ( Kennedy, et al., 2009 ) equal to 0.3 kg CO 2 e/cubic meter 5 .4.5 Embodied Energy of Materials Use Embodied energy incorporated in TBIF includes that for: wastewater (WW) treatment (T)/pumping (P), water treatment (T)/pumping (P), food production, and cement production since these activities are not already counted in in boundary GHG described in previous sections. Although WW and water treatment occurs within Delhi, subtracting these energy uses from the above estimates allows us t o clearly illustrate embodied energy used in WW and water operations. 5 .4.5 .1 Embodied Energy of Materials Benchmarks Wastewater (WW) treatment in Delhi is tracked and reported by the Delhi Jal Board, which treated 1,584 million liters/day of WW in 2009 ( MUD, 2010 ) using a total annual of 40 million kWh (T = 17, P = 23, million kWh) ( DJB, 2011 ) Municipal treated water supply totaled 3,125 million liters/day ( MUD, 2010 ) using a total annual of 266 million kWh (T = 242, P = 24, million kWh) ( DJB, 2011 ) For estimating average food consumption by Ind ian households, ( Miller & Ramaswami, 2011 ) used statistics from the Food and Agriculture Organization (FAO), thus resulting in 3,616 kg food/HH (or 0.78 tonnes food/resident), thereby estimating the 2009 food supply to De lhi as 13.8 million tonnes. This likely under estimates all food used in Delhi, as it excludes food in commercial/tourist establishments which may be a large proportion of the city's economy. Further, there are an estimated 45,285 non milk heads cattle, 45 ,760 milk heads cattle, and 304,655 heads buffaloes within Delhi boundaries ( DAH, 2010 ) which we use in the next sub section for estim ating direct methane emissions from

PAGE 153

! 138 enteric fermentation within Delhi boundaries. These in boundary estimates associated with milk producing cattle are subtracted from the trans boundary GHG emissions from food production. As previously discussed, there is no cement production within Delhi boundaries ( CMA, 2010 ) thus confirming that Delh i cement flows can be treated as trans boundary. Community wide cement use in Delhi was obtained from the Cement Manufacturers Association (CMA), and is estimated at 0.24 tonnes/capita/yr ( CMA, 2010 ) 5 .4.5 .2 Emissions Factors Water and wastewater is supplied from within Delhi, and thus relevant energy use has been subtracted from Delhi' s in boundary en ergy use total, avoiding double count. End use energy intensity used in 2009 for treating and pumping WW and water are obtained from ( DJB, 2011 ) and ( MUD, 2010 ) The resulting ratios of energy to water are, 0.03 Wh/liter of treated WW, 0.04 Wh/liter of pu mped WW, 0.21 Wh/liter of treated water, and 0.02 Wh/liter of pumped water. The Food EF has been retrieved from ( Miller & Ramaswami, 2011 ) who estimated a per unit weight EF from Indian food production (agriculture only). Their food EF qua ntifies direct methane and nitrous oxide emissions from Indian agriculture, eliminating double counting of energy used in processing or transporting food. The food EF, including emissions from cattle is 0.45 mt CO 2 e/mt food. This study also considered dire ct methane emissions from enteric fermentation in Delhi. GHG emissions per cattle in Delhi were retrieved from ( Sharma, et al., 2002b ) who estimated EF from non milk producing cattle as 525 kg CO 2 e/head/yr for milk producing cattle as 966 kg CO 2 e/head/yr, and buffaloes as 1,155 kg CO 2 e/head/yr, yielding GHG emissions from cattle within Delhi boundaries to be 419,855 mt CO 2 e. Lastly, to avoid double counting, we subtract cattle GHG emissions from the above food EF, resulting in a new food EF (less in boundary cattle), equal to 0.42 mt CO 2 e/mt food.

PAGE 154

! 139 Cement EF is well documented, and obtained from the literature. In this study, we applied an Indian EF from cement production equal to 0.93 mt CO 2 e/mt cement ( Hendriks, Worrell, Jager, Blok, & Riemer, 2004 ) 5 .5 Conclusions T his chapter presented the methodology and results from applying the Trans Boundary Infrastructure Footprint (TBIF) GHG emissions accounting approach in the rapidly developing city of Delhi, India. The objectives were to 1) describe data availability for implementing the TBIF in Delhi, 2) identify methodological differences between India and US based implementation of the TBIF, and 3) compare broad energy use metrics between Delhi and US cities, demonstrated by Denver which has previously been shown to be similar to US averages. Multiplying Delhi's 2009 material/energy flows (MFA) with associated emissions factors (EF) resulted in total TBIF GHG emissions equal to 40.3 million mt CO 2 e. Normalizing by population, Delhi's TBIF GHG emissions are 2.3 mt CO 2 e/capita; as expected, they are higher than the 1.5 mt CO 2 e/capita reported nationally ( MEF, 2010 ) since Delhi represents 1.5% of India's population. Of Delhi's 2009 TBIF GHG emissions, in boundary activities represented 68% (or 27.3 million mt CO 2 e) and trans boundary 32% (or 13 million mt CO 2 e). The buildings sector (including residential, commercial, industrial) represented 42% of Delhi's GHG emissions. GHGs from road transportation represented 21%, waste 2.7%, water/WW pumping and treatment 0.5 %, and cattle 1%. See 4"_`a^! C 6 $ The TBIF method was found to be very useful for measuring a comprehensive GHG footprint for Delhi. Most of the required data for app lying the TBIF for Delhi was found to be available, and it's possible that this could have been a result of higher levels of government reporting, as Delhi is a city state. There were some methodological differences that were a result of data constraints, though the method was mostly replicated. In fact, the method applied for Delhi helped to identify clear data needs and

PAGE 155

! 140 knowledge gaps where supplementary primary data collection is needed. For example, in the US, the method has used regional transportation models for allocating jet fuel use to multiple cities served by a single regional airport. In Delhi, the absence of a transportation model required the use of airport surveys to allocate total jet fuel used at the Delhi International Airport, to Delhi. In many ways, implementing TBIF was easier in a large mega city such as Delhi. Trans boundary VKT was found to be a small contribution (3%) of in boundary VKT, thus origin destination allocation of travel between cities in a commuter shed is not needed. Furt her, CMA reports annual cement use in city states such as Delhi, while obtaining this data has been challenging in US cities, and likely will also be the case in other Indian cities that are not city states. Comparing broad metrics across two distinct cities, Delhi, India and Denver, CO, USA, present compelling results. Delhi's per capita GHG emissions are higher than India's (2.4 vs. 1.5; mt CO 2 e/cap), reflecting low urbanization levels in India. Both Ramaswami et al. (2008) and Hillman & Ramaswami (2 010) note that Denver's GHG emissions are fairly close to the US national average, at about 25 mt CO 2 e/cap, due to 80% U.S. urbanization, i.e., 80% of the population lives in urban areas. Delhi's per capita GHG are almost a factor of ten lower than Denver' s, explained by a multitude of factors. For example, Delhi's residential primary energy use of 3,693 MJ/HH/mo is a factor of three lower than Denver (10,551 MJ/HH/mo); and road transportation travel in Delhi (8.8 VKT/cap/day) is about four times below Denv er (39 VKT/cap/day). Most notable are the differences in commercial industrial energy end use which are significantly lower in Delhi (2,064 MJ/capita/yr) compared to Denver (76,166 MJ/capita/yr). Similarly, commercial floor area per capita is much less in Delhi than Denver (1.46 sq meter/capita vs. 36.7 sq meter/capita). Even though the data suggests much less commercial activity in Delhi versus Denver, the economic GHG intensity provides additional insights. Delhi's economic GHG intensity is twice as larg e as Denver's, 948 mt CO 2 e/GDP versus 413 mt CO 2 e/GDP, respectively. Such difference may be attributed to economic structure, where Denver is predominantly a tertiary sector producer, and Delhi a secondary and tertiary sector producer. Other notable differ ences are shown in ([\]^! C 6 G Delhi's GDP/capita is about ten times lower than Denver's (6,037 USD/capita vs. 57,560

PAGE 156

! 141 USD/capita). In terms of population density, Delh i is significantly denser than Denver (9,340 cap/sq km vs. 1,463 cap/sq km). Homes in Delhi are smaller than homes in Denver (46.8 sq meter/HH vs. 102.8 sq meter/HH). As shown, the TBIF method can have important environmental and policy implications for De lhi and other rapidly industrializing cities. The TBIF shows an additional 32% of Delhi's GHG emissions attributed to trans boundary activities, thereby suggesting innovative cross sector strategies towards urban sustainability, particularly in electricity generation, and building materials/cement sectors. Comparing Delhi to Denver, supply chain GHG from cement use in construction contributed 10% to Delhi's TBIF, versus only 2% in Denver (Ramaswami et al. 2008); in contrast, waste/wastewater GHG were a lowe r proportion in Denver (at 1%) versus Delhi (at 3.3%) these data suggest that other construction materials not studied here may also be a significant part of Delhi's TBIF. The TBIF for Delhi shows that both waste management and material exchange symbiosi s can be important in reducing the TBIF of cities in rapidly industrializing countries.

PAGE 157

! 142 Figure 5 1 : TBIF for Delhi, India, 2009 expanded GHG footprint. In Boundary GHG s are represented by solid, and Trans Boundary GHGs are represented in hatched.

PAGE 158

! 143 Table 5 7 : Comparison of broad Energy & Material Use, and Demographic metrics for Delhi, India and Denver, CO, USA. Activity Secto r Metric Delhi, India ( 948 mt CO 2 e/ GDP ) (2.3 mt CO 2 e/cap) Denver, CO, USA j ( 413 mt CO 2 e/GDP ) (25 mt CO 2 e/cap) Buildings Energy Use & industrial process Residential Intensity : kWh/HH/mo 191 a 545 cubic meters/HH/mo n/a 124 liters LPG/HH/mo 25.3 b n/a Liters Kerosene/HH/mo 3.4 b n/a Total MJ/HH/mo (end use) 1,489 # 6,728 Total Primary (MJ/HH/mo) 3,693 # 10,551 # Commercial Industrial Intensity : kWh/GDP/yr 0.21 # 0.15 # Other stationary fuels MJ/GDP/yr 0.13 # 0.78 # Total MJ/GDP/yr (end use) 0.87 # 1.32 # Total MJ/capita/yr (end use) 2,064 # 76,166 # Total Primary (MJ/GDP/yr) 3.5 # 2.4 # Industrial Process : tonnes of waste/capita/yr 0.16 c 1.1 Electricity EF : kg CO 2 e/kWh 0.82 {0.83} 0.75 {0.64} Transportation Energy Use Surface Travel Intensity : VKT/capita/day 8.8 d 38.6 Air Travel : liters jet fuel/enplaned passenger (domestic) 56 e 72 Materials Use, and Demographics Water : treated water/WW (1000 liters/capita/yr) 95 f 560 Cement : mt cement/capita/yr 0.24 g 0.50 GDP/capita ($/capita) $6,037 h $57,560 k Total local population (capita) 17,601,000 i 579,744 Population Density (capita/sq km) 9,340 i 1,463 Total homes (HH) 3,815,104 i 256,524 Residential floor area (sm r /HH) 46.8 102.8 Total commercial floor area (million sm c ) 25.7 21.3 Total floor area per capita (sm/cap) 10.1 # 74.5 Total city area (sq km) 1,886 # 396 a. (DERC 2009) ; b. (MPNG 2009) ; c. (DPCC 2010) ; d. (CRRI 2009) ; e. (DIAL 2010) ; f. (MUD 2010) ; g. (CMA 2010) ; h. (DES 2009) ; i. (DCO 2009) ; j. (Hillman & Ramaswami 2010) ; k. (BEA 2009) ; #. Calculated ; *. Estimated, and may not represent most accurate statistic. Electricity EF : No brackets represent local EF. {brackets} represent national EF.

PAGE 159

! 144 6. Conclusion Contributions, Future Work & Protocols 6.1 Conclusions T his thesis explored mathematical relationships, approximations, implementation challenges, and policy relevance for three city scale GHG emission accounting methods The three methods P urely Territorial, Trans Boundary Infrastructure Supply Chain Footprint (TBIF), and Co nsumption Based Footprint (CBF) each have a unique representation of a city M athematical relationships showed that neither TBIF nor CBF provided a more holistic account ing of trans boundary GHG s and in fact showed that the two methods are linked. These relationships were also used to define a typology of cities defined as net p roducers trade b alanced and net c onsumers in terms of their GHGs embodied in trade Th e typology classification elucidated important differences in the total GHG footprint (territorial plus import supply chains ) for cities. Through a meta analysis of 21 US cities, Territorial GHGs were shown to be as sma ll as 37 % of the total footprint for a net co nsumer city and as large as 68 % for a net producer city T he TBIF was shown to capture 75 % (n=2) of the total footprint for net producer 63 % (n=11) for trade balanced and 62 % (n=8) for net consumer cities. Meanwhile, CBF cap t u res 35 % (n=2), 57% (n=11), and 71 % (n=8) of the total footprint for net producer trade balanced, and net consumer c ities, respectively. In total TBIF captures more than 60% of th e total footprint for all three city types, and CBF coverage is largely dependent on city type. The meta analysis showed that a number of trans boundary infrastructure sectors had high correlation ( R 2 >0.70) between community GDP and GHG in community wide ( residential comm ercial ind ustrial) u se of the same sectors These sectors (e lectricity generat ion, air travel, fuel refining, along with the production of food, c ement and iron &

PAGE 160

! 145 s teel ) might be well suited for allocation to cities based on their use in citywide residential comm ercial ind ustrial activities in the TBIF method Var ious suitable m et r i cs were explored to appropriately compare cities using GHGs computed by the three methods. For territorial GHG neither GHG Territorial /capita nor GHG Territorial /GDP reflected urban efficiency of cities. For the three versions of TBI F evaluated, GHG T BIF /GDP yielded stronger correlations with an u r ban efficie ncy index (UEI) as follows: with only ele ctricity allocated (R 2 =0.62); Scopes 1+2+3 w/o allocating (R 2 =0.75); and Scopes 1+2+3 w/ allocating (R 2 =0.77) Here again GHG TBIF /cap showe d poor correlation (R 2 =0.1) with the UEI as ex pected f r o m production based accounting. In contrast, GHG CBF /cap showed an improved correlation (R 2 =0.4) with the UEI and GHG CBF /cap correlated well (R 2 =0.76 ) with per capita expenditures. These data suggest that GHG TBIF /GDP is the appropriate metric for comparing cit i e s based on their urban efficiency, and that GHG CBF /cap i s a ppropriate for viewing cities f r o m a consu mption perspective For 21 US cities, GHG TBIF /GDP ranged from 154 mt CO 2 e/GDP to 747 mt CO 2 e/GDP, and GHG CBF /capita ranged from 15 mt CO 2 e/cap to 32 mt CO 2 e/capita. This thesis also presented results from the TBIF implemented in Delhi, India. The objectives of this part of the research were to explore issues of data availability and transferability of methods from the US to rapidly industrializing nations. We found that most m ethods translated well from the US to India and that d ata required for completing the TBIF w as reasonably available. In all, f ieldwork showed sufficient availability and adaptability of TBIF methodology from th e US to India yielding GHG TBIF equal to 948 mt CO 2 e/GDP in Delhi vs. 413 mt CO 2 e/GDP in Denver. B road energy use metrics between Delhi and Denver are shown to help describe differences between the two cities. 6.2 Contributions This thesis makes a number of unique contributions to the study and understanding of GHG emissions associated with cities. T he contributions of this thesis are:

PAGE 161

! 146 Presented the f irst side by side comparison of the three methods accounting for GHG emissions associated with cities Derived mathematical relationships between the three methods for cities Clearly articulated a need for studying cities by typology Clearly articulated the Tot al Upstream Footprint of cities and its coverage by the three methods Provided a b etter understanding of metrics for comparing cities on the basis of efficiencies Conducted i nternational field researc h in Delhi, India, leading to the first TBIF for Delhi. 6.3 Future Work & Protocols This thesis answered a number of key questions that are important in the understanding the GHGs associated with cities and number of areas presente d in this thesis can be pursued in future work towards the development of GHG emission accounting protocols. However, cities throughout the world still require additional resources to make progress in the direction of their respective climate goals. In the US and abroad, continued colla bo ration is req ui r ed to develop reasonable protocols t ailored to different city types that allow cities to maximize opportunities for GHG mitigation For example, net producer cities which are shown to have large territorial GHG relative to GHG embodied in imports, can have greater GHG mi tigation impacts by focusing their efforts on greening their local businesses and industries. Although net consumer cities have mitigation opportunities through their local busines ses and industries, consumer awareness campaigns aim ed at lowering consumption (energy and other goods/services consumed by households) may be better suited for these city types. L arge cities can benefit from a production based protocol as CBF may be appro ximated through TBIF, and data for conducting TBIF have been shown to be readily available in many cities. In this thesis we also uncovered some of the data challenges involved with using downscaled input output models for energy and environmental modeling In their current

PAGE 162

! 147 form, downscaled IO models are designed to be used for economic impact analysis and are not designed to be used with high accuracy for energy and environmental analysis Additional collaborations between IO developers and researchers w hich can help flag and correct these mismatc hes between dollar and energy flows can have multiple outcomes for GHG emission accounting and future work. Two potential outcomes are : 1. If IO tables can accurately capture trans boundary vs. in boundary contributions this could become an important feature of such data sets. 2. I mproved IO models at the city scale may capture the entire footprint associated with cities. In the meantime however, some of the other analysis presented in this thesis can have imp ortant implications in protocol development. T he TBIF capture s the majority of the to tal GHG footprint ranging from 62% 75% for 21 US cities TBIF was shown to correlate with the urban efficiency performance of cities on a consistent basis. Implementation and data for TBIF has shown to be readily available for US and Indian cities. TBIF p rovide s a holistic account of GHGs, a nd TBIF is approximately equal to CBF (TBIF CBF) for trade balanced cities, of which large cites may be Lastly as TBIF intrinsically follows the five principles of GHG accounting defined by the WRI (Relevance, Completeness, Consistency, Transparency, Accuracy), TBIF is well suited for international GHG protocols that seek to compare cities

PAGE 163

! 148 Appendix A : Energy Use Data Retr i e ved From GHG Inventories an d IMPLAN = >?!@83,63(*!1(%$*A!B'% 1C1DEFGDGEH Electricity Use % community wide electricity use that is generated locally Unadjusted IMPLAN Community GHG Inventory [state benchmark] Unadjusted IMPLAN EPA eGRID (local generation) Sacramento CA Residential Intensity 721 kWh/HH/mo 748 [580] kWh/HH/mo 78% 61% Total Commercial Industrial Use 7,245 GWh 5,774 GWh Napa, CA Residential Intensity 830 kWh/HH/mo 623 [580] kWh/HH/mo 30% 1% Total Commercial Industrial Use 689 GWh 563 GWh Boulder, CO Residential Intensity 1,138 kWh/HH/mo 852 [743] kWh/HH/mo 60% 48% Total Commercial Industrial Use 2,938 GWh 2,142 GWh Broomfield, CO Residential Intensity 1,078 kWh/HH/mo 825 [768] kWh/HH/mo 0% 0% Total Commercial Industrial Use 629 GWh 447 GWh Denver, CO Residential Intensity 1,284 kWh/HH/mo 546 [768] kWh/HH/mo 91% 19% Total Commercial Industrial Use 11,313 GWh 5,038 GWh Routt, CO Residential Intensity 980 kWh/HH/mo 833 [743] kWh/HH/mo 98% 100% Total Commercial Industrial Use 287 GWh 251 GWh Collier, FL Residential Intensity 1,074 kWh/HH/mo 1,780 [1,354] kWh/HH/mo 24% 0% Total Commercial Industrial Use 2,068 GWh 2,059 GWh Sarasota, FL Residential Intensity 952 kWh/HH/mo 1,403 [1,367] kWh/HH/mo 43% 0% Total Commercial Industrial Use 1,730 GWh 1,861 GWh Broward, FL Residential Intensity 922 kWh/HH/mo 1,352 [1,354] kWh/HH/mo 18% 40% Total Commercial Industrial Use 10,475 GWh 10,713 GWh Miami Dade, FL Residential Intensity 906 kWh/HH/mo 1,267 [1,367] kWh/HH/mo 65% 90% Total Commercial Industrial Use 15,477 GWh 14,300 GWh Washoe, Residential 966 700 [1,022] 43% 10%

PAGE 164

! 149 NV Intensity kWh/HH/mo kWh/HH/mo Total Commercial Industrial Use 3,566 GWh 2,863 GWh Tompkins, NY Residential Intensity 547 kWh/HH/mo 564 [554] kWh/HH/mo 95% 100% Total Commercial Industrial Use 786 GWh 486 GWh Westcheste r, NY Residential Intensity 885 kWh/HH/mo 589 [575] kWh/HH/mo 90% 100% Total Commercial Industrial Use 4,252 GWh 3,283 GWh Multnomah OR Residential Intensity 1,290 kWh/HH/mo 793 [1,092] kWh/HH/mo 90% 53% Total Commercial Industrial Use 12,936 GWh 5,746 GWh Philadelphi a, PA Residential Intensity 1,097 kWh/HH/mo 507 [851] kWh/HH/mo 87% 5% Total Commercial Industrial Use 7,973 GWh 8,969 GWh Roanoke, VA Residential Intensity 1,090 kWh/HH/mo 1,261 [1,247] kWh/HH/mo 99% 0% Total Commercial Industrial Use 758 GWh 517 GWh Loudoun, VA Residential Intensity 1,183 kWh/HH/mo 1,472 [1,247] kWh/HH/mo 24% 0% Total Commercial Industrial Use 2,627 GWh 2,153 GWh Snohomish, WA Residential Intensity 1,402 kWh/HH/mo 994 [1,114] kWh/HH/mo 92% 10% Total Commercial Industrial Use 4,319 GWh 3,184 GWh METRO, OR Residential Intensity 1,208 kWh/HH/mo 714 [1,071] kWh/HH/mo 84% 31% Total Commercial Industrial Use 21,071 GWh 12,101 GWh New York City Residential Intensity 800 kWh/HH/mo 374 [554] kWh/HH/mo 98% 46% Total Commercial Industrial Use 48,316 GWh 34,088 GWh DVRPC, PA NJ Residential Intensity 1,032 kWh/HH/mo 842 [851] kWh/HH/mo 86% 63% Total Commercial Industrial Use 47,360 GWh 36,776 GWh

PAGE 165

! 150 I=EBF=C!J=K BK1 I+)8$+,!J+' B'% ,e[mn`lc^m!&:>/1' -dkk`e"cg!898! &e#^ecdag! !"#$#%& '%()*+$,-. K+.$+L%()4M!D= )^l"m^ec"[]!&ec^el"cg =B!cj^akl O99Okd == /0 & cj^akl O99Okd (dc[]!-dkk^ah"[] 6 &em`lca"[]!,l^ .C=!k"]]"de!cj^akl $.=!k"]]"de!cj^akl I+&+M!D= )^l"m^ec"[]!&ec^el"cg ;
PAGE 166

! 151 R8,)(4L+UM!NF )^l"m^ec"[]!&ec^el"cg $B!cj^akl O99Okd =K 41 & cj^akl O99Okd (dc[]!-dkk^ah"[] 6 &em`lca"[]!,l^ =KK!k"]]"de!cj^akl $<;!k"]]"de!cj^akl XU3,+6%,&U3+M! X= )^l"m^ec"[]!&ec^el"cg =K!cj^akl O99Okd C= /1 & cj^akl O99Okd (dc[]!-dkk^ah"[] 6 &em`lca"[]!,l^ $
PAGE 167

! 152 Appendix B: Sector Trade For Three Communities Routt, Colorado Sector Description Export Value (million $) Real Estate $202 Coal Mining $164 Other amusement gambling and recreation industri es $99 Power generation and supply $61 Food services and drinking places $51 Hotels and motels including casino hotels $42 Cattle ranching and farming $28 Commercial and institutional buildings $24 Hospitals $19 Gasoline Stations $18 Denver, Colorado Sector Description Export Value (million $) Oil & Natural Gas Extraction $5,994 Real Estate $5,437 Air transportation $2,940 Wholesale services $2,751 Telecommunications $2,501 S ecurities, commodity contracts, investments, and related services $2,266 Management of companies $1,483 Legal services $1,416 Advertising services $967 Software $820 Sarasota, Florida Sector Description Export Value (million $) Professional and technical services $600 Metal window and door manufacturing $505 Real estate $461 Telecommunications $289 Insurance agencies $212 Offices of physicians dentists other health $185 Services to buildings and dwellings $181 Employment services $179 Paint and coating manufacturing $133 S ecurities, commodity contracts, investments, and related services $111

PAGE 168

! 15 3 Appendix C: IMPLAN Sector Scheme IMPLAN Sector Code NAICS Sector Code Sector Description 3001 1111A0 Oilseeds 3002 1111B0 Grains 3003 111200 Vegetables and melons 3004 1113A0 Fruit 3005 111335 Tree nuts 3006 111400 Greenhouse, nursery, and floriculture products 3007 111910 Tobacco 3008 111920 Cotton 3009 1119A0 Sugarcane and sugar beets 3010 1119B0 All other crop farming products 3011 1121A0 Cattle from ranches and farms 3012 112120 Dairy cattle and milk products 3013 112300 Poultry and egg products 3014 112A00 Animal products, except cattle, poultry and eggs 3015 113A00 Forest, timber, and forest nursery products 3016 113300 Logs and roundwood 3017 114100 Fish 3018 114200 Wild game products, pelts, and furs 3019 115000 Agriculture and forestry support services 3020 211000 Oil and natural gas 3021 212100 Coal 3022 212210 Iron ore 3023 212230 Copper, nickel, lead, and zinc 3024 2122A0 Gold, silver, and other metal ore 3025 212310 Natural stone 3026 212320 Sand, gravel, clay, and ceramic and refractory minerals 3027 212390 Other nonmetallic minerals 3028 213111 Oil and gas wells 3029 213112 Support services for oil and gas operations 3030 21311A Support services for other mining 3031 221100 Electricity, and distribution services 3032 221200 Natural gas, and distribution services 3033 221300 Water, sewage treatment, and other utility services 3034 230101 Newly constructed nonresidential commercial and health care structures 3035 230102 Newly constructed nonresidential manufacturing structures 3036 230103 Other newly constructed nonresidential structures

PAGE 169

! 154 3037 230201 Newly constructed residential permanent site single and multi family structures 3038 230202 Other newly constructed residential structures 3039 230301 Maintained and repaired nonresidential structures 3040 230302 Maintained and repaired residential structures 3041 311111 Dog and cat food 3042 311119 Other animal food 3043 311210 Flour and malt 3044 311221 Corn sweetners, corn oils, and corn starches 3045 31122A Soybean oil and cakes and other oilseed products 3046 311225 Shortening and margarine and other fats and oils products 3047 311230 Breakfast cereal products 3048 31131A Raw and refined sugar from sugar cane 3049 311313 Refined sugar from sugar beets 3050 311320 Chocolate cacao products and chocolate confectioneries 3051 311330 Chocolate confectioneries from purchased chocolate 3052 311340 Nonchocolate confectioneries 3053 311410 Frozen foods 3054 311420 Canned, pickled and dried fruits and vegetables 3055 31151A Fluid milk and butter 3056 311513 Cheese 3057 311514 Dry, condensed, and evaporated dairy products 3058 311520 Ice cream and frozen desserts 3059 31161A Processed animal (except poultry) meat and rendered byproducts 3060 311615 Processed poultry meat products 3061 311700 Seafood products 3062 311810 Bread and bakery products 3063 311820 Cookies, crackers, and pasta 3064 311830 Tortillas 3065 311910 Snack foods including nuts, seeds and grains, and chips 3066 311920 Coffee and tea 3067 311930 Flavoring syrups and concentrates 3068 311940 Seasonings and dressings 3069 311990 All other manufactured food products 3070 312110 Soft drinks and manufactured ice 3071 312120 Beer, ale, malt liquor and nonalcoholic beer 3072 312130 Wine and brandies 3073 312140 Distilled liquors except brandies 3074 3122A0 Cigarettes, cigars, smoking and chewing tobacco, and reconstituted tobacco 3075 313100 Fiber filaments, yarn, and thread 3076 313210 Broadwoven fabrics and felts

PAGE 170

! 155 3077 313220 Woven and embroidered fabrics 3078 313230 Nonwoven fabrics and felts 3079 313240 Knitted fabrics 3080 313310 Finished textiles and fabrics 3081 313320 Coated fabric coating 3082 314110 Carpets and rugs 3083 314120 Curtains and linens 3084 314910 Textile bags and canvas 3085 314990 All other textile products 3086 315100 Knit apparel 3087 315210 Cut and sewn apparel from contractors 3088 315220 Mens and boys cut and sewn apparel 3089 315230 Womens and girls cut and sewn apparel 3090 315290 Other cut and sew apparel 3091 315900 Apparel accessories and other apparel 3092 316100 Tanned and finished leather and hides 3093 316200 Footwear 3094 316900 Other leather and allied products 3095 321100 Dimension lumber and preserved wood products 3096 32121A Veneer and plywood 3097 32121B Engineered wood members and trusses 3098 321219 Reconstituted wood products 3099 321910 Wood windows and doors and millwork 3100 321920 Wood containers and pallets 3101 321991 Manufactured homes (mobile homes) 3102 321992 Prefabricated wood buildings 3103 321999 All other miscellaneous wood products 3104 322110 Wood pulp 3105 322120 Paper from pulp 3106 322130 Paperboard from pulp 3107 322210 Paperboard containers 3108 32222A Coated and laminated paper, packaging paper and plastics film 3109 32222B All other paper bag and coated and treated paper 3110 322230 Paper and paperboard stationary products 3111 322291 Sanitary paper products 3112 322299 All other converted paper products 3113 323110 Printed materials 3114 323120 Printing support services 3115 324110 Refined petroleum products 3116 324121 Asphalt paving mixtures and blocks 3117 324122 Asphalt shingles and coating materials

PAGE 171

! 156 3118 324191 Petroleum lubricating oils and greases 3119 324199 All other petroleum and coal products 3120 325110 Petrochemicals 3121 325120 Industrial gas 3122 325130 Synthetic dyes and pigments 3123 325181 Alkalies and chlorine 3124 325182 Carbon black 3125 325188 All other basic inorganic chemicals 3126 325190 Other basic organic chemicals 3127 325211 Plastics materials and resins 3128 325212 Synthetic rubber 3129 325220 Artificial and synthetic fibers and filaments 3130 325310 Fertilizer 3131 325320 Pesticides and other agricultural chemicals 3132 325411 Medicines and botanicals 3133 325412 Pharmaceutical preparations 3134 325413 In vitro diagnostic substances 3135 325414 Biological products (except diagnostic) 3136 325510 Paints and coatings 3137 325520 Adhesives 3138 325610 Soaps and cleaning compounds 3139 325620 Toilet preparations 3140 325910 Printing inks 3141 3259A0 All other chemical products and preparations 3142 326110 Plastics packaging materials and unlaminated films and sheets 3143 326121 Unlaminated plastics profile shapes 3144 326122 Plastics pipes and pipe fittings 3145 326130 Laminated plastics plates, sheets (except packaging), and shapes 3146 326140 Polystyrene foam products 3147 326150 Urethane and other foam products (except polystyrene) 3148 326160 Plastics bottles 3149 32619A Other plastics products 3150 326210 Tires 3151 326220 Rubber and plastics hoses and belts 3152 326290 Other rubber products 3153 32711A Pottery, ceramics, and plumbing fixtures 3154 32712A Bricks, tiles, and other structural clay products 3155 32712B Clay and nonclay refractory products 3156 327211 Flat glass 3157 327212 Other pressed and blown glass and glassware 3158 327213 Glass containers

PAGE 172

! 157 3159 327215 Glass products made of purchased glass 3160 327310 Cement 3161 327320 Ready mix concrete 3162 327330 Concrete pipes, bricks, and blocks 3163 327390 Other concrete products 3164 3274A0 Lime and gypsum products 3165 327910 Abrasive products 3166 327991 Cut stone and stone products 3167 327992 Ground or treated mineral and earth products 3168 327993 Mineral wool 3169 327999 Miscellaneous nonmetallic mineral products 3170 331110 Iron and steel and ferroalloy products 3171 331200 Steel products from purchased steel 3172 33131A Aluminum products 3173 331314 Aluminum alloys 3174 33131B Aluminum products from purchased aluminum 3175 331411 Copper 3176 331419 Nonferrous metals (except copper and aluminum) 3177 331420 Rolled, drawn, extruded and alloyed copper 3178 331490 Rolled, drawn, extruded and alloyed nonferrous metals (except copper and aluminum) 3179 331510 Ferrous metals 3180 331520 Nonferrous metals 3181 33211A All other forged, stamped, and sintered metals 3182 332114 Custom roll formed metals 3183 33211B Crowned and stamped metals 3184 33221A Cutlery, utensils, pots, and pans 3185 33221B Handtools 3186 332310 Plates and fabricated structural products 3187 332320 Ornamental and architectural metal products 3188 332410 Power boilers and heat exchangers 3189 332420 Metal tanks (heavy gauge) 3190 332430 Metal cans, boxes, and other metal containers (light gauge) 3191 33299A Ammunition 3192 33299B Arms, ordnance, and accessories 3193 332500 Hardware 3194 332600 Spring and wire products 3195 332710 Machined products 3196 332720 Turned products and screws, nuts, and bolts 3197 332800 Coated, engraved, heat treated products 3198 33291A Valves and fittings other than plumbing 3199 332913 Plumbing fixture fittings and trims

PAGE 173

! 158 3200 332991 Balls and roller bearings 3201 332996 Fabricated pipes and pipe fittings 3202 33299C Other fabricated metals 3203 333111 Farm machinery and equipment 3204 333112 Lawn and garden equipment 3205 333120 Construction machinery 3206 333130 Mining and oil and gas field machinery 3207 33329A Other industrial machinery 3208 333220 Plastics and rubber industry machinery 3209 333295 Semiconductor machinery 3210 33331A Vending, commercial, industrial, and office machinery 3211 333314 Optical instruments and lens 3212 333315 Photographic and photocopying equipment 3213 333319 Other commercial and service industry machinery 3214 33341A Air purification and ventilation equipment 3215 333414 Heating equipment (except warm air furnaces) 3216 333415 Air conditioning, refrigeration, and warm air heating equipment 3217 333511 Industrial molds 3218 33351A Metal cutting and forming machine tools 3219 333514 Special tools, dies, jigs, and fixtures 3220 333515 Cutting tools and machine tool accessories 3221 33351B Rolling mills and other metalworking machinery 3222 333611 Turbines and turbine generator set units 3223 333612 Speed changers, industrial high speed drives, and gears 3224 333613 Mechanical power transmission equipment 3225 333618 Other engine equipment 3226 333911 Pumps and pumping equipment 3227 333912 Air and gas compressors 3228 333920 Material handling equipment 3229 333991 Power driven handtools 3230 33399A Other general purpose machinery 3231 333993 Packaging machinery 3232 333994 Industrial process furnaces and ovens 3233 33399B Fluid power process machinery 3234 334111 Electronic computers 3235 334112 Computer storage devices 3236 33411A Computer terminals and other computer peripheral equipment 3237 334210 Telephone apparatus 3238 334220 Broadcast and wireless communications equipment 3239 334290 Other communications equipment 3240 334300 Audio and video equipment

PAGE 174

! 159 3241 334411 Electron tubes 3242 334412 Bare printed circuit boards 3243 334413 Semiconductor and related devices 3244 33441A Electronic capacitors, resistors, coils, transformers, and other inductors 3245 334417 Electronic connectors 3246 334418 Printed circuit assemblies (electronic assemblies) 3247 334419 Other electronic components 3248 334510 Electromedical and electrotherapeutic apparatus 3249 334511 Search, detection, and navigation instruments 3250 334512 Automatic environmental controls 3251 334513 Industrial process variable instruments 3252 334514 Totalizing fluid meters and counting devices 3253 334515 Electricity and signal testing instruments 3254 334516 Analytical laboratory instruments 3255 334517 Irradiation apparatus 3256 33451A Watches, clocks, and other measuring and controlling devices 3257 33461A Software, blank audio and video media, mass reproduction 3258 334613 Magnetic and optical recording media 3259 335110 Electric lamp bulbs and parts 3260 335120 Lighting fixtures 3261 335210 Small electrical appliances 3262 335221 Household cooking appliances 3263 335222 Household refrigerators and home freezers 3264 335224 Household laundry equipment 3265 335228 Other major household appliances 3266 335311 Power, distribution, and specialty transformers 3267 335312 Motor and generators 3268 335313 Switchgear and switchboard apparatus 3269 335314 Relay and industrial controls 3270 335911 Storage batteries 3271 335912 Primary batteries 3272 335920 Communication and energy wires and cables 3273 335930 Wiring devices 3274 335991 Carbon and graphite products 3275 335999 All other miscellaneous electrical equipment and components 3276 336111 Automobiles 3277 336112 Light trucks and utility vehicles 3278 336120 Heavy duty trucks 3279 336211 Motor vehicle bodies 3280 336212 Truck trailers 3281 336213 Motor homes

PAGE 175

! 160 3282 336214 Travel trailers and campers 3283 336300 Motor vehicle parts 3284 336411 Aircraft 3285 336412 Aircraft engines and engine parts 3286 336413 Other aircraft parts and auxiliary equipment 3287 336414 Guided missiles and space vehicles 3288 33641A Propulsion units and parts for space vehicles and guided missiles 3289 336500 Railroad rolling stock 3290 336611 Ships 3291 336612 Boats 3292 336991 Motorcycles, bicycles, and parts 3293 336992 Military armored vehicles, tanks, and tank components 3294 336999 All other transportation equipment 3295 337110 Wood kitchen cabinets and countertops 3296 337121 Upholstered household furniture 3297 337122 Nonupholstered wood household furniture 3298 33712A Metal and other household furniture (except wood) 3299 337127 Institutional furniture 3300 33721A Wood television, radio, and sewing machine cabinets 3301 337212 Office furniture and custom architectural woodwork and millwork 3302 337215 Showcases, partitions, shelving, and lockers 3303 337910 Mattresses 3304 337920 Blinds and shades 3305 339112 Surgical and medical instrument, laboratory and medical instruments 3306 339113 Surgical appliances and supplies 3307 339114 Dental equipment and supplies 3308 339115 Ophthalmic goods 3309 339116 Dental laboratories 3310 339910 Jewelry and silverware 3311 339920 Sporting and athletic goods 3312 339930 Dolls, toys, and games 3313 339940 Office supplies (except paper) 3314 339950 Signs 3315 339991 Gaskets, packing and sealing devices 3316 339992 Musical instruments 3317 33999A All other miscellaneous manufactured products 3318 339994 Brooms, brushes, and mops 3319 420000 Wholesale trade distribution services 3320 441000 Retail Services Motor vehicle and parts OR BEA ALL RETAIL 3321 442000 Retail Services Furniture and home furnishings 3322 443000 Retail Services Electronics and appliances

PAGE 176

! 161 3323 444000 Retail Services Building material and garden supply 3324 445000 Retail Services Food and beverage 3325 446000 Retail Services Health and personal care 3326 447000 Retail Services Gasoline stations 3327 448000 Retail Services Clothing and clothing accessories 3328 451000 Retail Services Sporting goods, hobby, book and music 3329 452000 Retail Services General merchandise 3330 453000 Retail Services Miscellaneous 3331 454000 Retail Services Nonstore, direct and electronic sales 3332 481000 Air transportation services 3333 482000 Rail transportation services 3334 483000 Water transportation services 3335 484000 Truck transportation services 3336 485000 Transit and ground passenger transportation services 3337 486000 Pipeline transportation services 3338 48A000 Scenic and sightseeing transportation services and support activities for transportation 3339 492000 Couriers and messengers services 3340 493000 Warehousing and storage services 3341 511110 Newspapers 3342 511120 Periodicals 3343 511130 Books 3344 5111A0 Directories and mailing lists 3345 511200 Software 3346 512100 Motion pictures and videos 3347 512200 Sound recordings 3348 515100 Radio and television entertainment 3349 515200 Cable and other subscription services 3350 516110 Internet publishing and broadcasting services 3351 517000 Telecommunications 3352 5181 Data processing hosting ISP web search portals 3353 519100 Other information services 3354 52A000 Monetary authorities and depository credit intermediation services 3355 522A00 Nondepository credit intermediation and related services 3356 523000 Securities, commodity contracts, investments, and related services 3357 524100 Insurance 3358 524200 Insurance agencies, brokerages, and related services 3359 525000 Funds, trusts, and other financial services 3360 531000 Real estate buying and selling, leasing, managing, and related services 3361 S00800 Imputed rental services of owner occupied dwellings 3362 532100 Automotive equipment rental and leasing services 3363 532A00 General and consumer goods rental services except video tapes and

PAGE 177

! 162 d iscs 3364 532230 Video tape and disc rental services 3365 532400 Commercial and industrial machinery and equipment rental and leasing services 3366 533000 Leasing of nonfinancial intangible assets 3367 541100 Legal services 3368 541200 Accounting, tax preparation, bookkeeping, and payroll services 3369 541300 Architectural, engineering, and related services 3370 541400 Specialized design services 3371 541511 Custom computer programming services 3372 541512 Computer systems design services 3373 54151A Other computer related services, including facilities management 3374 541610 Management, scientific, and technical consulting services 3375 5416A0 Environmental and other technical consulting services 3376 541700 Scientific research and development services 3377 541800 Advertising and related services 3378 541920 Photographic services 3379 541940 Veterinary services 3380 5419A0 All other miscellaneous professional, scientific, and technical services 3381 550000 Management of companies and enterprises 3382 561300 Employment services 3383 561500 Travel arrangement and reservation services 3384 561100 Office administrative services 3385 561200 Facilities support services 3386 561400 Business support services 3387 561600 Investigation and security services 3388 561700 Services to buildings and dwellings 3389 561900 Other support services 3390 562000 Waste management and remediation services 3391 611100 Elementary and secondary education from private schools 3392 611A00 Education from private junior colleges, colleges, universities, and professional schools 3393 611B00 Other private educational services 3394 621A00 Offices of physicians, dentists, and other health practitioners 3395 621600 Home health care services 3396 621B00 Medical and diagnostic labs and outpatient and other ambulatory care services 3397 622000 Private hospital services 3398 623000 Nursing and residential care services 3399 624400 Child day care services 3400 624A00 Individual and family services 3401 624200 Community food, housing, and other relief services, including rehabilitation services

PAGE 178

! 163 3402 711100 Performing arts 3403 711200 Spectator sports 3404 711A00 Promotional services for performing arts and sports and public figures 3405 711500 Independent artists, writers, and performers 3406 712000 Museum, heritage, zoo, and recreational services 3407 713940 Fitness and recreational sports center services 3408 713950 Bowling activities 3409 713A00 Amusement parks, arcades, and gambling recreation 3410 713B00 Other amusements and recreation 3411 7211A0 Hotels and motel services, including casino hotels 3412 721A00 Other accommodation services 3413 722000 Restaurant, bar, and drinking place services 3414 8111A0 Automotive repair and maintenance services, except car washes 3415 811192 Car wash services 3416 811200 Electronic and precision equipment repairs and maintenance 3417 811300 Commercial and industrial machinery and equipment repairs and maintenance 3418 811400 Personal and household goods repairs and maintenance 3419 812100 Personal care services 3420 812200 Death care services 3421 812300 Dry cleaning and laundry services 3422 812900 Other personal services 3423 813100 Services from religious organizations 3424 813A00 Grantmaking, giving, and social advocacy services 3425 813B00 Civic, social, and professional services 3426 814000 Cooking, housecleaning, gardening, and other services to private households 3427 491000 US Postal delivery services 3428 S Fed Util Not a unique commodity (electricity from fed govt utilities) 3429 S00102 Products & services of Fed Govt enterprises (except electric utilities) 3430 S00201 Not a unique commodity (passenger transit by state & local govt) 3431 S State Util Not a unique commodity (electricity from state & local govt utilities) 3432 S00203 Products & services of State & Local Govt enterprises (except electric utilities) 3433 S00402 Used and secondhand goods 3434 S00401 Scrap 3435 S00900 Rest of the world adjustment 3436 S00300 Noncomparable foreign imports 3437 S00700 Employment and payroll only (state & local govt, non education) 3438 S00700 Employment and payroll only (state & local govt, education) 3439 S00600 Employment and payroll only (federal govt, non military) 3440 S00500 Employment and payroll only (federal govt, military)

PAGE 179

! 164 REFERENCES Alberti, M., & Susskind, L. (1996). Managing Urban Sustainability. Environmental Impact Assessment Review, 16 (4 6), 213 221. Anand, S., & Sen, A. (2000). Human Development and Economic Sustainability. World Development, 28 (10), 2029 2049. ARAI. (2007). Emission Factor development for Indian Vehicles : Automotive Research Association of India. Arora, S., Vyas, A., & Johnson, L. R. (2011). Projections of highway vehicle population, energy demand, and CO2 emissions in India to 2040. Natural Resources Forum, 35 49 62. Baynes, T., Lenzen, M., Steinberger, J. K., & Bai, X. (2011). Comparison of hous ehold consumption and regional production approaches to assess urban energy use and implication for policy. Energy Policy, 39 7298 7309. BEA. (2008). Benchmark Input Output Data. from jcciNOOooo%\^ [%_d#O"em`lcagO"dp\^ehjk[aq%jck BEA. (2009). GDP by Industry Accounts: Gross Output by Industry in Current Dollars and Price Indexes by Industry., from jcciNOOooo%\^[%_d# BEA. (2011). Industry Economic Accounts. from jcciNOOooo%\^[%_d#O"em`lcagO BEA. (2012). Regional Economic Accounts. from jcciNOOooo%\^[%_d#Oa^_"de[]O Beaulieu, J. J., Tank, J. L., Hamilton, S. K., Wollheim, W. M., Jr., R. O. H., Mulholland, P. J., et al. (2011). Nitrous oxide emission from denitrification in stream and river networks : PNAS. Betsill, M. M., & Bulkeley, H. (2006). Cities and the multilevel governance of global climate change. Global Governance, 12 (2) 141 159.

PAGE 180

! 165 BLS. (2012a). Consumer Expendiure Survey. from jcciNOOooo%\]l%_d#Oh^MO BLS. (2012b). Quarterly Census of Employment and Wages. from jcciNOOooo%\]l%_d#Oh^oO Bose, R., & Sperling, D. (2001). Transportation in Developing Countries: Greenhouse Gas Scenarios for Delhi, India : Prepared for the Pew Center on Global Climate Change. Brown, M. A., Southworth, F., & Sarzynski, A. (2009). The geography of metropolitan carbon foot prints. Policy and Society, 27 (4), 285 304. BTS. (2009). Freight in America. Cabanatuan, M. (2008, December 28, 2008). S.F. fliers may pay their way in carbon usage San Francisco Chronicle cCCR. (2010). carbonn Cities Climate Registry. from jcciNOOh"c"^lh]"k[c^a^_"lcag%da_O CEA. (2009). CO2 Baseline Database for the Indian Power Sector : Government of India. Ministry of Power Central Electricity Authority. Census. (2007). U.S. Census Urban/Rural and I nside/Outside Metropolitan Area Census. (2010). Resident Population Data Retrieved from jcciNOO.K$K%h^el`l%_d#O.K$Kh^el`lOm[c[ Census. (2011). Metropolitan and Micropolitan Statistical Areas Retrieved from jcciNOOooo%h^el`l%_d#Oidi`][c"deOk^cadO Census. (2012a). Census 2000 Geographic Definitions Retrieved from jcciNOOooo%h^el`l%_d#O_^dOoooO_^dpm^fe%jck] Census. (2012b). County Business Patterns. Retrieved March, 2012, from jcciNOOooo%h^el`l%_d#O^hdeOh\iO"em^M%jck]

PAGE 181

! 166 Chavez, A., & Ramaswami, A. (2011). Pr ogress Toward Low Carbon: Approaches for Trans Boundary Greenhouse Gas Emissions Footprinting for Cities. Carbon Management, 2 (4), 471 482. Chavez, A., Ramaswami, A., Nath, D., Ranjan, R., & Kumar, E. (2012). Implementing Expanded Geographic Based Greenhou se Gas Accounting for Delhi, India: Data Availability and Methods. Journal of Industrial Ecology, In Press CMA. (2010). Highlights of Indian Cement Industry. from jcciNOOd]ml"c^%hk["em"[%da_O"em`lcag% jck] CMU. (2008). Economic Input Output Life Cycle Assessment (EIO LCA). from jcciNOOooo%^"d]h[%e^cO CRRI. (2009). Impact of Vehicular Emission Norms on Vehicular Pollution Loads in Delhi : Central Road Research Ins titute. DAH. (2010). Cattle in Delhi According to Livestock Census : Directorate of Animal Husbandary, Govt. of NCT of Delhi. DCO. (2009). Area and Population : Directorate of Census Operations, Delhi. DDE. (2009). Cabinet Decision No. 1552: Energy Conservat ion and Building Code (ECBC) : Delhi Department of Environment (DDE). Denver. (2010). City and County of Denver 2007 GHG Inventory Update : Center for Sustainable Infrastructure Systems, University of Colorado Denver. DERC. (2009). Energy and Electricity : Delhi Electricity Regulatory Commission. DES. (2009). State Domestic Product Estimates : Directorate of Economics & Statistics (DES). Government of NCT of Delhi. DES. (2010). Delhi Statistical Handbook Delhi, India: Directorate of Economics & Statistics, G overnment of National Capital Territory of Delhi, 2010. DIAL. (2011). IGI Airport Fuel and Passenger Data: Delhi International Airport Private Limited, Economic Planning & Research Department.

PAGE 182

! 167 DJB. (2011). Electricity Use in Water and Wastewater Delhi : D elhi Jal Board. DOE. (2009). Transportation Energy Data Book : U.S. Department of Energy. DPCC. (2010). Municipal Solid Waste Management in Delhi : Delhi Pollution Control Committee. DRCOG. (2007). Workforce Commuting Patterns. : Denver Regional Council of Go vernments (DRCOG). EIA. (2010). International Energy Statistics Washington D.C.: U.S. Energy Information Administration (EIA). EIA. (2011). State Energy Data System (SEDS). Retrieved March, 2011, from jcciNOOo oo%^"[%_d#Olc[c^Ol^mlO Eide, M. H. (2002). Life Cycle Assessments (LCA) of Industrial Milk Production. The International Journal of Life Cycle Asessment, 7 (2), 115 126. EPA. (2007). Climate Leaders Program. EPA. (2010). Inventory of U.S. Greenhouse G as Emissions and Sinks: 1990 2008 : U.S. Environmental Protection Agency. EPA. (2011). The Emissions & Generation Resource Integrated Database (eGRID). Retrieved June, 2011, from jcciNOOooo%^i[%_d#Oh]^[e^e^a_gO^e^a_g 6 a^ld`ah^lO^_a"mO"em^M%jck] EPA. (2012). Toxic Release Inventory (TRI). Retrieved March, 2012, from jcciNOOooo%^i[%_d#Oca"O FHWA. (2008). Highway Statistics 2008 Retrieved from jcciNOOooo%fjo[%mdc%_d#Oid]"hgOdji"OjllOjlli`\l%hfk Ghosh, C. (2009). Inventorization of Green House Gases Sources and Sinks in Delhi : University of Delhi, Centre for Environmental Management of Degraded Ecosystem.

PAGE 183

! 168 Greenprint. (2007). Greenhouse Gas Inventory for the City & County of Denver : Greenprint Denver. Grimm, J. (2009). Food Urbanism: a sustainable design option for urban communities Grimm, N., Faeth, S. H., Golubiewski, N. E., Redman, C. L., Wu, J., Bai, X., et al. (2008). Global Change and the Ecology of Cities. Science, 319 (5864), 756 760. Gupta, R. (2011). A critical and comparative evaluation of approaches and policies to measure, benchmark, reduce and manage CO2 emissions from energy use in the existing building stock of developed and rapidly developing countries case studies of UK, USA, and India Hanya, T., & Ambe, Y. (1976). A study on the metabolism of cities. Science Hendr ickson, C., Horvath, A., Joshi, S., & Lave, L. (1998). Economic Input Output Model for Environmental Life Cycle Assessment. Environ. Sci. Technol., 32 (7), 184 191. Hendriks, C. A., Worrell, E., Jager, D. d., Blok, K., & Riemer, P. (2004). Emissions Reducti on of Greenhouse Gases from the Cement Industry. Paper presented at the Greenhouse Gas Control Technologies Conference. Hillman, T., & Ramaswami, A. (2010). Greenhouse Gas Emission Footprints and Energy Use Benchmarks for Eight U.S. Cities. Environ. Sci. T echnol., 44 (6), 1902 1910. Hirst, E. (2000). Expanding US Transmission Capacity : Report to Edison Electric. ICLEI. (2009). Cities for Climate Protection Participants. ICLEI. (2010). Local Government Operations Protocol: For the Quantification and Report ing of Greenhouse Gas Emissions Inventories. ICLEI. (2011). Community Scale GHG Emissions Accounting and Reporting Protocol DRAFT Framework. : ICLEI Local Governments for Sustainability. IDFC. (2010). Multi Modal Tranposrt in a Low Carbon Future : Infra structure Development Finance Company Limited.

PAGE 184

! 169 IEA. (2008a). India Statistics : International Energy Agency. IEA. (2008b). World Energy Outlook : International Energy Agency. IEA. (2010). CO2 Emissions from Fuel Combustion : International Energy Agency (IEA). IMPLAN. (2010). IPCC. (2006a). Volume 3: Industrial Processes and Product Use : Intergovernmental Panel on Climate Change (IPCC). IPCC. (2006b). Volume 5: Waste : Intergovernmental Panel on Climate Change (IPCC). IRFCA. (2006). Energy Consumption by Indian Railways Jacobson, M. Z. (2010). The enhancement of local air pollution by urban CO2 domes. Environ. Sci. Technol., 44 (7), 2497 2502. Jones, C. M., & Kammen, D. M. (2011). Quantifying Carbon Footprint Reduction Opportunities for U.S. Households and Commun ities. Environ. Sci. Technol., 45 (9), 4088 4095. Kennedy, C., Steinberger, J., Gasson, B., Hansen, Y., Hillman, T., Havranek, M., et al. (2009). Greenhouse Gas Emissions from Global Cities. Environ. Sci. Technol., 43 (19), 7297 7302. Kumar, S., Kapoor, R., Deshmukh, A., Kamath, M., & Manu, S. (2010). Total Commercial Floor Space Estimates for India : International Resources Group. Kumar, S., Mondal, A. N., Gaikwad, S. A., Devotta, S., & Singh, R. N. (2004). Qualitative assessment of met han emission inventory from municipal solid waste disposal sites: a case study. Atmoshpheric Environment, 38 4921 4929. Lenzen, M., Pade, L. L., & Munksgaard, J. (2004). CO2 multipliers in multi region input output models. Economic Systems Research, 16 (4) 391 412. Lewis, C. (1997). Fuel and Energy Production Emission Factors

PAGE 185

! 170 Liang, S., & Zhang, T. (2011). Urban Metabolism in China: Achieving Dematerialization and Decarbonization in Suzhou. Journal of Industrial Ecology, 15 (3), 420 434. Lindall, S., Olson, D., & Alward, G. (2005). Deriving Multi Region Models Using the IMPLAN National Trade Flows Model Stillwater, MN: MIG, Inc. Mayors, U. (2010). US Mayors Climate Protection Agreement. Retrieved May, 2011, from jcciNOOooo%`lk[gdal%da_Oh]"k[c^iadc^hc"deO]"lc%[li McGraw, J., Haas, P., Young, L., & Evans, A. (2010). Greenhouse gas emissions in Chicago: Emissions inventories and reduction strategies for Chicago and its metro politan region. Journal of Great Lakes Research, 36 106 114. MEF. (2010). India: Greenhouse Gas Emissions 2007 : Ministry of Environment and Forests (MEF). Government of India. Mehta, R. (2001). History, politics and technology of CNG diesel switch in Delh i. Land Use, Transportation and the Environment. The Transportation Asia Project. MIG. (2004). IMPLAN Pro User Guide Stillwater, MN: MIG, Inc. IMPLAN. Miller, L. (2011). Sustainability Assessment of Wastewater Treatment Plants With Water Reuse for Urban A griculture: A Case Study in Hyderabad, India : University of Colorado Denver. Miller, L., & Ramaswami, A. (2011). Cross Country Review of Food Life Cycle Assessment: From Production to Post Consumption Waste. submitted MPNG. (2009). Petroleum and Natural G as : Ministry of Petroleum and Natural Gas, GOI, New Delhi. MUD. (2010). Improving Service Outcomes 2008 09. Service Level Benchmarking Databook : Ministry of Urban Development. Government of India. Newcombe, K., Kalma, J. D., & Aston, A. R. (1978). The Meta bolism of a City: The Case of Honk Kong. Ambio, 7 (1), 3 15.

PAGE 186

! 171 Ngo, N. S., & Pataki, D. E. (2008). The energy and mass balance of Los Angeles County. Urban Ecosyst, 11 121 139. Pachauri, S. (2004). An analysis of cross sectional variations in total household energy requirements in India using micro survey data. Energy Policy, 32 1723 1735. Pachauri, S., & Jiang, L. (2008). The Household Energy Transition in India and China Laxe nburg, Austria: International Institution for Applied Systems Analysis. Paris, M. d. (2009). Le Bilan Carbone de Paris: Bilan des Žmissions de gaz ˆ effet de serre Parshall, L., Gurney, K., Hammer, S. A., Mendoza, D., Zhou, Y., & Geethakumar, S. (2010). M odeling energy consumption and CO2 emissions at the urban scale: Methodological challenges and insights from the United States. Energy Policy, 38 4765 4782. Peters, & Hertwich, E. G. (2008). CO2 Embodied in International Trade with Implications for Global Climate Policy. Environ. Sci. Technol., 42 (4), 1401 1407. Peters, G. P. (2010). Carbon Footprints and Embodied Carbon at Multiple Scales. Environmental Sustainability, 2 245 250. Ramaswami, A., Chavez, A., Ewing Thiel, J., & Reeve, K. (2011). Two Approac hes to Greenhouse Gas Emissions Foot Printing at the City Scale. Environ. Sci. Technol. 45 (10), 4205 4206. Ramaswami, A., Hillman, T., Janson, B., Reiner, M., & Thomas, G. (2008). Demand Centered, Hybrid Life Cycle Methodology for City Scale Greenhouse G as Inventories. Environ. Sci. Technol., 42 (17), 6455 6461. Routt. (2010). Routt County GHG Emissions Inventory and Sustainable Energy Benchmarking and Actions 2005 : Center for Sustainable Infrastructure Systems, University of Colorado Denver.

PAGE 187

! 172 Roychowdhury, A., Chattopadhyaya, V., Sen, J., & Chandola, P. (2008). Avert the great guzzle: Fuel economy regulations Setting the principles right : Centre for Science and Environment, Delhi. Sahely, H. R., Dudding, S., & Kennedy, C. A. (2003). Estimating the urban m etabolism of Canadian cities: Greater Toronto Area case study. Canadian Journal of Civil Engineering, 30 (2), 468 483. Sarasota. (2008). Sarasota County Greenhouse Gas Emissions Analysis Report Sharholy, M., Ahmad, K., Mahmood, G., & Trivedi, R. C. (2008). Municipal Solid Waste Management in Indian Cities A Review. Waste Management 38 459 467. Sharma, C., Dasgupta, A., & Mitra, A. P. (2002a). Future Scenarios of Inventories of GHGs and Urban Pollutants From Delhi and Calcutta. Paper presented at the IGES/APN Mega City Project, Kitakyushu Japan. Sharma, C., Dasgupta, A., & Mitra, A. P. (2002b). Inventory of GHG's and Other Urban Pollutants From Agriculture and Waste Sectors In Delhi and Calcutta. Paper presented at the IGES/APN Mega City Project, Kitakyushu Japan. Singh, A., Gangopadhyay, S., Nanda, P. K., Bhattacharya, S., Sharma, C., & Bhan, C. (2008). Trends of greenhouse gas emissions from the road transport sector in India. Science of the Total Environment, 390 124 131. Sov acool, B. K., & Brown, M. A. (2010). Twelve metropolitan carbon footprints: A preliminary comparative global assessment. Energy Policy, 38 4856 4869. Stanton, E. A., Bueno, R., Ackerman, F., Erickson, P., Hammerschlag, R., & Cegan, J. (2011). Greenhouse G as Impacts of Oregon's Consumption: Technical Report Somerville, MA: Stockholm Environmental Institute US Center. Stanton, E. A., Bueno, R., Ackerman, F., Erickson, P., Hammerschlag, R., & Cegan, J. (2012). Greehouse Gas Emissions in King County: An Upd ated Geographic Plus Inventory, a Consumption Based Inventory, and an Ongoing Tracking Framework Somerville, MA: Stockholm Environmental Institute US Center.

PAGE 188

! 173 TAG. (2002). Loudoun County, Virginia: Strategic Plan for Telecommunication Facilities : The Atl antic Group. TOI. (2008, November 25). 33% of Indians live in less space than US prisoners Times of India TWB. (2010). World Development Indicators Washington D.C.: The World Bank (TWB). UN. (2007). World Urbanization Prospects: The 2007 Revision : United Nations. UN. (2008). United Nations Demographic Yearbook. New York: United Nations. UN. (2010). Population Density and Urbanization : United Nations (UN) Statistical Division. UNFCCC. (2006). Updated UNFCCC Reporting guidelines on annual inventories : United Nations Framework Convention on Climate Change (UNFCCC). USDOC. (2012). Foreign Trade. from jcciNOOooo%h^el`l%_d#Ofda^"_e 6 ca[m^O Valley, Y. (2011). Yampa Valley Data Parterns Fast Facts. Retr ieved January, 2012, from jcciNOOooo%g[ki[#[]]^gi[ace^al%hdkOf[lcpf[hcl Weber, C. L., & Matthews, H. S. (2008). Quantifying the global and distributional aspects of American household carbon foo tprint. Ecological Economics, 66 379 391. Weber, C. L., & Matthews, H. S. (2010). Food Miles and the Relative Climate Impacts of Food Choices in the United States. Environ. Sci. Technol., 42 (10), 3508 3513. Whitaker, M. (2007). Life Cycle Assessment of Tr ansit Systems in the U.S. and India: Implications for a Carbon Constrained Future. PhD Dissertation : University of Colorado Denver. WMSC. (2010). World Mayor's Summit on Climate Change. Retrieved March, 2011, from jcciNOOooo%oklh.K$K%da_Ocj^ 6 k^M"hd 6 h"cg 6 i[hcO

PAGE 189

! 174 Wolman, A. (1965). The metabolism of cities. Scientific American, 213 (3), 179 190. WRI. (2004). The Greenhouse Gas Protocol: A Corporate Accounting and Reporting Standard Washington, DC: World Resources Institute and World Business Council for Sustainable Development. WRI. (2011). Corporate Value Chain (Scope 3) Accounting and Reporting Standard Washington D.C.: World Resources Institute and World Business Council for Sustainable Developm ent. Wright, L. A., Coello, J., Kemp, S., & Williams, I. (2011). Carbon Footprinting For Climate Change Management in Cities. Carbon Management, 2 (1), 49 60. Wright, L. A., Kemp, S., & Williams, I. (2011). Carbon Footprint: Towards a Universally Accepted D efinition. Carbon Management, 2 (1), 61 72. Zhou, N., Price, L., & Ohshita, S. (2010). A Low Carbon Development Guide for Local Government Actions in China. Paper presented at the US China Workshop on Pathways Toward Low Carbon Cities, Hong Kong.