Urban Growth 2 Growth in Electricity Use (2001-2011) 3 (Kennedy - PowerPoint PPT Presentation

Estimating the Energy Use and CO 2 Emissions of the World’s Cities Shweta Singh 1,2 and Chris Kennedy 3 1. Agricultural & Biological Engineering, Purdue University, USA 2. Environmental & Ecological Engineering, Purdue University, USA 3. Department of Civil Engineering, University of Toronto, Canada International Renewable Energy Agency Roundtable Renewable Energy Deployment in Cities, 29 th October 2015 , Singapore 1

Urban Growth 2

Growth in Electricity Use (2001-2011) 3 (Kennedy et al. 2015)

Modeling Urban Energy Consumption  How would the future urban growth impact energy consumption and related emissions at global scale ?  Objectives  Develop a tool for urban energy consumption estimation.  Application of tool for estimation of global scale impact. 4

22 City Data Set  Data on Electricity (MWh/cap), Heating (GJ/cap) and Transportation (GJ/cap). North/South Asia Bangkok America Denver Shanghai Los Angeles Beijing New York City Jakarta Toronto Amman Chicago Manila Buenos Aires Tianjin Sao Paulo Europe Barcelona Rio de Janeiro Paris-IDF Geneva Africa Dar es Salaam London Cape Town Prague 5

Regression Model for Urban Energy Use  Energy Consumption Categories  Electricity (MWh/cap)  Heating (Gj/cap)  Transportation (GJ/cap)  Energy Consumption Drivers  Urban Density  Per capita space – Inv-Density (ha/cap)  Climate (Heating Degree Days or HDD and Cooling Degree Days or CDD)  HDD  Gross Domestic Product (GDP) 6

Regression Models 7 Ref : Singh and Kennedy, Env Pollution Sp. Issue, 2015

Application at Global Scale Test on 3646 City Data Set No. of Asian Cities : 1783 (approx. 50 %) 8

Data for 3646 Global Cities  Lincoln Institute of Land Policy, Cambridge, MA  Primary focus on land policy, taxation.  Work on urban expansion.  Shlomo Angel et al.  Working Paper : “A Planet of Cities : Urban Land Cover of Estimates and Projections for All Countries, 2000- 2050”  Metropolitan Agglomerations with Populations > 100,000 in 2000  MOD500 : Satellite Based Map of Urban Land Cover  Data  Density : Based on the empirical studies  Population : UN Based projection  Missing Data for Energy Estimation : HDD (Temperature, Weather) 9

Estimating HDD : HDD Proxy  Heating Degree Days estimated based on weather data.  Assumption : Heating is required below 18 deg C.  HDD Estimated (HDD Proxy ) 12 𝐼𝐸𝐸 𝑄𝑠𝑝𝑦𝑧 = 𝑈 𝑏𝑤𝑕 − 18 × 𝑂𝑝. 𝑝𝑔 𝑒𝑏𝑧𝑡 𝑗𝑜 𝑓𝑏𝑑ℎ 𝑛𝑝𝑜𝑢ℎ 𝑗=1 10

Determining T avg for Urban Areas  Climate Data Extracted from Global Historical Climate Network (University of Delaware)  GIS gridded monthly air temperature data  Great Circle Distance method for surface distance calculations.  T avg = mean of stations < 25 miles from urban areas 11

Is the model applicable for Global Estimation ?  Statistical Testing for applicability 12

Extrapolation Check for Global City Data  Based on HDD, 11 % of HDDProxy for global dataset was outset the model dataset HDD.  Based on Inv-Density, 13 % of Inv-Density for global dataset was outside the model dataset. 13

Multidimensional Extrapolation Check : Leverages  Leverage is calculated as the diagonal element (h ii ) of the matrix H. 𝑌 ′ 𝑌 −1 𝑌 ′ 𝐼 = 𝑌 where X is the matrix of Predictor Variables  Interpretation  h ii is “measure for distance of i th case in X from the mean of all the n values”  For the new values of Predictor Variables, leverage is 𝑌 ′ 𝑌 −1 𝑌 𝑜𝑓𝑥 ′ 𝐼 𝑜𝑓𝑥 = 𝑌 𝑜𝑓𝑥  About 13 % of global data set may be under the effect of extrapolation using leverages. 14

Scenarios Simulation at Global Scale Projections of Future Energy and Emissions for Cities 15

Application of Developed Tool for Global Estimation  Global Energy and Emissions Estimation (2000)  Impact of Global Transition to Electric Vehicles.  Population and Density Change Impact.  Climate Change Projection Impact. 16

Global Urban Energy Consumption (2000) Energy Emissions 70 8.35E+03 TWh 60 4.54 E +10 GJ % of Urban in Global 50 5.59E+10 GJ 40 30 5.32 GT-CO 2 20 4.11 GT-CO 2 3.14 GT-CO 2 10 0 Electricity Heating Transportation 17 Urban Share in Global Energy and Related Emissions (2000)

Urban Energy (EJ) 18

Urban Energy (EJ) 19

Urban Energy (EJ) 20

Urban Energy (EJ) 21

What if Urban Areas Switch to Electric Vehicles ?  Carbon Intensity of grid is important for transition to Evs if reducing emissions is the goal.  Grid Intensity at which EVs become carbon competitive : 643 t CO 2 e/GWh (Hertwich et al, LCA of Evs)  𝐻𝐼𝐻 𝐹𝑊 = 𝐻𝐼𝐻 𝐽𝐷𝐹 × 𝑗 643 where i = intensity of electricity grid. Scenario & Results  Scenario 1 : Hold the intensity i constant at year 2000. (Data from UNFCC)  Emissions reduced from 3.14 Gt-CO2 to 2.98 Gt-CO2 (5 % reduction)  Scenario 2 : Countries with i > 643 bring the intensity down to 643.  Emissions reduced from 3.14 Gt-CO2 to 2.60 Gt-CO2 (17 % reduction) EVs as sustainable mode of transportation for emissions reduction only useful if countries reduce grid intensity. 22

Urban Density Change Impact  High Projection : Density decline of 2 % per annum.  Medium Projection : Density decline of 1 % per annum.  Low Projection : Density decline of 0 % per annum.  Density Calculations 𝑞 𝐸 𝑢 = 𝐸 𝑢−1 − × 𝐸 𝑢−1 100 Where D t = Density at t th year D t-1 = Density at t-1 th year p = Rate of density decline 23

Urban Density Change Impact on Energy 2.5E+11 420 % Electricity (MWH) Heating (GJ) Transportation (GJ) 2.0E+11 1.5E+11 213 % 72 % 104 % 72 % 72 % 1.0E+11 89 % 34 % 34 % 34 % 67% 36 % 5.0E+10 308 % 79 % 165 % 33 % 54 % 80 % 0.0E+00 0% 1% 2% 0% 1% 2% 2020 2050 Urban Density Decline Urban Density Decline % Shown Over Year 2000 Consumption 24

Urban Density Change Impact on Emissions 25 346 % Electricity Heating Transportation 20 428 % 193 % Gt-CO 2 15 90 % 101 % 218 % 10 64 % 42.8 % 78 % 78 % 78 % 107 % 92 % 37 % 37 % 37 % 69.7 % 39 % 5 0 0% 1% 2% 0% 1% 2% Urban Density Decline Urban Density Decline 2020 2050 Reference Year for Emissions Increase : 2000 GHG Intensities held constant at year 2000. Ref Electricity Emissions : 5.32 Gt-CO 2 Ref Heating Emissions : 4.11 GT-CO 2 25 Ref Transportation Emissions : 3.14 GT-CO 2

Climate Change Impact on Urban Energy Consumption  Climate change projections from National Center of Atmospheric Research (NCAR) based on IPCC scenarios and Community Climate System Model (CCSM3).  IPCC Scenarios  IPCC-Commit : GHG concentration held constant at year 2000 and models run for 100 years of temperature projection.  B1 : Implements global solutions for economic, social and environmental sustainability issues.  A1B : Very rapid economic growth, growth of population till mid century followed by decline and introduction of more efficient technologies with balanced reliance on both fossil & non-fossil energy sources.  A2 : Regional approach with heterogeneous economic and population growth. 26

IPCC Scenarios : Impact on Electricity Consumption IPCC-Commit SRES-B1 SRES-A1B SRES-A2 4.0E+10 308% 307% 313% 305% 3.5E+10 3.0E+10 162% 163% 2.5E+10 MWh 167% 160% 2.0E+10 76% 75% Approx 80 % 81% 73% 1.5E+10 Approx 53 % Approx 32% 1.0E+10 5.0E+09 0.0E+00 0% 1% 2% 0% 1% 2% Urban Density Decline Urban Density Decline 2020 2050 27

IPCC Scenario Impact on Electricity Emissions 14 IPCC-Commit SRES-B1 SRES-A1B SRES-A2 % Reduction in growth of Emissions over year 2000 12 10 8 6 4 2 0 0% 1% 2% 0% 1% 2% Urban Density Decline Urban Density Decline 2020 2050 SRES A1B Scenario wins for reducing electricity emissions over long term 28

IPCC Scenario Impact on Heating Emissions 40 IPCC-Commit SRES-B1 SRES-A1B SRES-A2 % Reduction in growth of Emissions over year 2000 35 30 25 20 15 10 5 0 0% 1% 2% 0% 1% 2% Urban Density Decline Urban Density Decline 2020 2050 SRES A1B Scenario wins for reducing heating emissions over long term 29

Potential Model Developments for Renewable Energy Deployments  Improved demand estimates using larger data set & region specific adjustments.  More variables such as GDP – move towards more regional models for urban areas.  Resources accessibility can be a major driver for energy consumption  T avg for Urban area can be improved by testing the robustness of GIS methods for distance against monitored data.  Supply Side : Estimate rooftop solar potential for 3646 cities. 30

Acknowledgements  NSERC Accelerator Grant, Canada to Chris Kennedy, University of Toronto  University of Delaware, Center for Climatic Research, Department of Geography – For Climate Data. (http://climate.geog.udel.edu/~climate/)  Shlomo Angel and Collaborators for City Data. (http://www.lincolninst.edu/)  Iain Stewart : Urban Climatologist, University of Toronto 31

Thank You 32