long term mitigation scenarios - energy research … · a long term mitigation scenarios for south...

Long Term Mitigation Scenarios

Technical Appendix

Prepared for: Department of Environment Affairs and Tourism

South Africa

Prepared by: Energy Research Centre

October 2007

The following citation should be used for this report:

Energy Research Centre 2007 Long Term Mitigation Scenarios: Technical Appendix, Department of Environment Affairs and Tourism, Pretoria, October 2007

The suite of reports that make up the Long Term Mitigation Scenario study include the following:

A Long Term Mitigation Scenarios for South Africa

B Technical Summary

C Technical Report

C.1 Technical Appendix

D Process Report

The study was supported by the following inputs:

LTMS Input Report 1: Energy emissions

LTMS Input Report 2: Non-energy emissions: Agriculture, Forestry and Waste

LTMS Input Report 3: Non-energy emissions: Industrial Processes

LTMS Input Report 4: Economy-wide modeling

LTMS Input Report 5: Impacts, vulnerability and adaptation in key South African sectors

Contents

1. Scenario building team members 1

2. Support teams 3

3. Scope of work of research teams 3

3.1 Energy emissions modelling 3

3.2 Non-energy emissions modelling 4

3.3 Macro-economic 4

3.4 Climate change impacts and adaptation 5

4. Energy modeling: more detailed parameters 5

4.1 Agricultural energy demand 6

4.2 Commercial sector demand for energy 6

4.3 Industrial energy demand 7

4.4 Residential energy demand 8

4.5 Transport energy demand 9

4.6 Power generation technologies 11 4.6.1 Coal-fired pulverized fue 11 4.6.2 Fluidized-bed combustion (FBC) 11 4.6.3 Integrated gasification combined cycle (IGCC) 11 4.6.4 Gas-fired open-cycle gas-turbine (OGCT) 12 4.6.5 Gas-fired combined-cycle gas-turbine (CCGT) 12 4.6.6 Nuclear power plants: pressure water reactor (PWR) 12 4.6.7 Nuclear power plants: pebble-bed modular reactor (PBMR) 13 4.6.8 Hydroelectric power and pumped storage 13 4.6.9 Wind 13 4.6.10 Concentrating solar power systems 13 4.6.11 Solar photovoltaic systems 14 4.6.12 Biomass for electricity generation 14 4.6.13 Municipal waste for electricity generation 15

5. Data for analysing industrial process emissions 15 5.1 Introduction 15

5.1.1 General assumptions 15 5.1.2 Methodology 15 5.1.3 Selection of mitigation options 15

5.2 Emission estimates 16 5.2.1 Sector emissions for 1990 16 5.2.2 Update of emission inventory for purposes of this study 17

5.3 Estimates for 2003 emissions 18 5.3.1 Mineral products 18

5.3.1.1 Cement production 18 5.3.1.2 Lime production and dolomite use 18

5.3.2 Chemicals 19 5.3.2.1 Ammonia production 19 5.3.2.2 Nitric acid production 19

5.3.2.3 Carbide production 19 5.3.2.4 Balance of the chemical sector 19

5.3.3 Metals 19 5.3.3.1 Iron and steel industry 19 5.3.3.2 Ferro-alloy production 19 5.3.3.3 Aluminium 20

5.3.4 Mine emissions 20 5.3.4.1 Coal mining 20

5.3.5 Synfuels-specific emissions 20 5.3.5.1 Methane emissions 20 5.3.5.2 Concentrated carbon dioxide streams 20 5.3.5.3 Expansion of synfuels production using natural gas. 20 5.3.5.4 Expanded coal to liquids production. 21

5.3.6 Summary of baseline (2003) non-energy emissions 21

5.4 Emissions for baseline and mitigation scenarios, including cost estimates for reduction 21 5.4.1 Synfuels point-source emissions 21 5.4.2 Iron and steel industry 23 5.4.3 Ferroalloy industry 23 5.4.4 Nitric acid production 23 5.4.5 Balance of chemical industry 23 5.4.6 Cement production 23 5.4.7 Mining 24 5.4.8 Aluminium production 24

6. International experiences on mitigation for agriculture 24

6.1 Introduction 24

6.2 Mitigation opportunities: Increased sinks and reduced emissions 25 6.2.1 Opportunities to increase soil carbon 25

6.2.1.1 Cropland management 26 6.2.1.2 Grazingland and hayland management 26 6.2.1.3 Land-use changes to increase soil carbon 27 6.2.1.4 Total agricultural soil carbon sequestration potential 27 6.2.1.5 Reducing agricultural nitrous oxide and methane emissions 27 6.2.1.6 Reducing nitrous oxide and methane emissions from soils 28 6.2.1.7 Reducing livestock-related methane and nitrous oxide emissions 28

7. Methodology for modelling emissions from livestock enteric emissions 29

7.1 Historical data, assumptions and calculations for enteric fermentation 29

8. Methodology for modelling emissions from livestock manure management 33

8.1 Data, assumptions and calculations of baseline and mitigated emissions for manure management 33

8.2 Assumptions and calculations for mitigation for manure management 34

9. Methodology for modelling emissions from reduced tillage 35

9.1 Historical data, assumptions and calculations for tillage 35 9.1.1 Area under cultivation 35 9.1.2 Carbon storage 36

10. Capital and variable costs requirements to start a no-till system 37

11. Methodology for modelling mitigation from land use changes (fire control and savannah thickening) 40

11.1 Fire control 40

11.2 Savannah thickening 41

12. Data for modelling mitigation from waste sector 43

13. Economic impact analysis 45

13.1 Introduction 45

13.2 Modelling Approach and Scenario Description 47 13.2.1 Objectives 47 13.2.2 Model and Data 48

13.2.2.1 CGE Modelling Overview 48 13.2.2.2 A South African Social Accounting Matrix and Activity Account

Disaggregations 48 13.2.3 Simulation Setup 52

13.2.3.1 Modelling Energy Efficiency and Fuel Switching 52 13.2.3.2 Modelling Structural Shifts 53 13.2.3.3 Modelling the Impact of a CO2 Emissions Tax 54 13.2.3.4 Modelling Investment Requirements 56 13.2.3.5 Additional Modelling Information: Model Closures 56

13.2.4 Final Remarks About the Modelling Approach 57 13.2.4.1 Combined Scenarios: Economic Effects and Modelling 57 13.2.4.2 The Reference Case, Forecasting and Analysis Period 58

13.3 Results and Analyses 59 13.3.1 Start Now and Scale Up 59

13.3.1.1 Simulation Setup 59 13.3.1.2 GDP, Employment and Welfare Effects 60 13.3.1.3 Sensitivity of Results 63

13.3.2 Use the Market 64 13.3.2.1 Simulation Setup 64 13.3.2.2 Modelling Issues 64 13.3.2.3 GDP, Employment and Welfare Effects 65 13.3.2.4 Final Remarks 67

13.4 Analysing the Effects of Individual Mitigation Components (Wedges) 67 13.4.1 Energy Efficiency Wedges 67

13.4.1.1 Overview 67 13.4.1.2 Industrial Energy Efficiency 68 13.4.1.3 Commercial Energy Efficiency 68

13.4.2 Structural Change Wedges 69 13.4.2.1 Overview 69 13.4.2.2 Reference Case 70 13.4.2.3 Nuclear Intensive Scenario and Renewables Scenarios for the

Electricity Sector 70 13.4.2.4 Biofuels Scenario for the Petroleum Sector 74

13.4.3 The Economic Impact of a CO2 Emissions Tax 76 13.4.3.1 Overview 76 13.4.3.2 Prices, Output and Employment 77 13.4.3.3 Concluding Remarks 84

13.5 Additional Tables and Figures 85

References 92

Appendices to the Technical Report – LTMS 1

1. Scenario building team members

Name Company Telephone / cell E-Mail address

Alan Hirsch Presidency 012 300 5200 [email protected] / [email protected]

Alan Munn Sustainable Business Manager, Engen

031 460 2439 / 083 642 2103 [email protected]

Andre van der Bergh BHP Billiton 011 376 3218 / 082 467 7877 [email protected]

Andrew Borraine SA Cities 021 462 1441 [email protected]

Annie Sugrue CURES [email protected]

Bill Rowlston DWAF 012 336 8768 [email protected] / [email protected]

Bob Scholes CSIR [email protected],

Bongi Gqasa DPE 012 431 1126 / 082 451 5591 [email protected]

Bridget Thovhakale DST 012 843 6382 [email protected]

Chris Moseki DWAF 012 336 7867 / 082 801 3532 [email protected]

Dick Kruger Mining 011 498 7275 / 082 569 4417 [email protected]

Dipolelo Elford W.C DEA & DP 021 483 2723 / 084 656 4291 [email protected]

Dr A Paterson Aluminium 011 453 3339 [email protected]

Dr Johan Ledger SESSA 011 680 1553 [email protected]

Dr Joseph Matjila Centre of Excellence Kumba Resources (Iron & Steel)

012 310 3703 / 012 307 4952 [email protected]

Dr Laurrain Lotter CAIA (Chemical) 011 482 1671/2/3/4 [email protected]

Dr Trevor Chorn Engen 021 403 4668 / 083 700 3457 [email protected]

Dr. John Scotcher Forestry SA 033 330 2330 / 083 626 8990 [email protected]

Hassan Mohamed Presidency 012 300 5200 [email protected]

Herman J. van der Walt

SASOL 011 344 0141 / 083 630 0426 [email protected]

Ian Langridge EIUG 011 638 2161 / 083 4599 505 [email protected]

Imraan Patel DST 012 843 6430 / 082 800 5531 [email protected]

Jacob Dikgang DoT 012 309-3243 / 3973; 083 484-2764 ?

[email protected]

Jason Shaeffler 072 4443445 [email protected]

Jeff Subramoney DME 012 317 8662 [email protected]

Joanne Yawitch DEAT 012 310 3666 / 082 571 5337 [email protected]

Judy Beaumont DEAT 012 310 3532 / 082 653 0625 [email protected]

Justice Mavhungu DPE 012 431 1089 [email protected]

Kadri Kevin Nassiep CEF / SANERI 011 280 0421 / 082 460 7804 [email protected]

Kelebogile Moroka DEAT 012 310 3710 [email protected]

Leila Mohamed SEA 021 702 3622 [email protected]

Linda Manyuchi DST 012 843 6521 / 083 634 5216 [email protected]

Litha Mcwabeni Corporate Strategy &

012 431 1087 / 076 903 4890 [email protected]


Name Company Telephone / cell E-Mail address

Structure, DPE

Liza Roussot Treasury 012 315 5071 / 083 306 1638 [email protected]

Lize Coetzee DT 012 309 3428 / 082 909 6222 [email protected]

Lwazikazi Tyani DME 012 317 8227 / 082 465 8077 [email protected]

Lydia Greyling DFA 012 351 1487 / 082 710 4072 [email protected]

Mandy Rambharos Eskom 011 800 6300 / 083 234 1859 [email protected]

Maphuti Legodi DME Via Rabelani above; or Elsa +27 12-317-8216

[email protected]

Maryna Möhr-Swart Chamber of Mines 011 498 7406 / 082 882 7700 [email protected]

MC Moseki Water Resource Planning Systems

012 336 7867 / 082 801 3532 [email protected] (Secretary, Natasha Stoh email)

Mike Edwards Forestry 011 803 3403 / 082 600 7627 [email protected]

Sibusiso Labour [email protected]

Nic Opperman AgriSA 012 322 6980 [email protected]

Nwabisa Matoti Business Unity SA

011 784 8000 / 083 515 0219 [email protected]

Peter Lukey DEAT 012 310 3710 / 083 415 2963 [email protected]

Professor Robin Barnard

Agriculture 012 310 2549 [email protected]

Rabelani Tshikalanke DME 012 317-8434 [email protected]

Richard Worthington Earthlife 082 446 6392 [email protected]

Rod Crompton NERSA 012 401 4003 [email protected]

Russel Baloyi SALGA 084 330 3030 [email protected]

Sakkie van der Westhuizen

Sappi (Paper) 011 407 8367 [email protected]

Sharlin Hemraj Treasury 012 315 5875 / 5071 [email protected]

Simangele Mgquba DME 012 317 8307; 083 263 5247 [email protected]

Stan Jewaskiewitz Envitech Solutions

011 849 9759 / 082 808 0586 [email protected]

Tabby Resane DEAT via Brenda: 012 310 3449 [email protected]

Tony Frost WWF 021 888 2800 [email protected]

Tony Surridge SA Energy 011 280 0448 / 079 499 5062 Fax: 011/280 0577

[email protected]

Tshenge Demana DTI 012 310 9820 [email protected]

Tsietsi Mahema DEAT 012 310 3404 / 082 316 6952 [email protected]

Wendy Poulton Eskom 011 800 2634 / 082 829 7602 [email protected]

Y Stan Pillay Sustainable Development, Anglo Coal

(011) 638 5323 / 082 771 6668 (from list) 6868 (from email)

[email protected]

John Purchase GRAIN SA 082 441 2308 / 056 515 2145 [email protected]

Sonwabo Damba Eskom 011 800 3379 / 083 785 0140 [email protected]

Jeffrey Kgobane DME 082 355 4927 / 012 317 8232 [email protected]

Jonas Mphenya SAWS [email protected]

Note: Bobby Peek of groundWork (033 342 5662 [email protected]) attended one meeting of the SBT, raised specific concerns relating to the cement industry. This should not be taken as endorsement by groundWork of the LTMS as a whole.


Support teams

Project Management: ERC

Dr Harald Winkler

Pierre Mukheibir

Research teams

Harald Winkler – UCT ERC Research Leader

Alison Hughes – UCT ERC Energy modelling

Rina Tavia / Bob Scholes – CSIR Non-energy

Kalie Pauw – UCT Economics Macro-economic

Guy Midgley - SANBI Impacts

Facilitation team

Tokiso Facilitators: Stef Raubenheimer; Edwin Molahlehi

Advisory: Gerrit Kornelius

Tokiso Secretarial:Rachel Mosupye; Yasmin Moola;

Elin Lorimer (support)

Project management team

Joanne Yawitch, DEAT

Peter Lukey, DEAT

Shirley Moroka, DEAT

Smangele Mgquba, DME

2. Scope of work of research teams Each of the four research team leaders gave an indication of their existing work and how it might be useful in the LTMS process. This provided initial information for the SBT. The scope of work of each of the teams has been placed on the closed web-site (see below).

2.1 Energy emissions modelling

ERC modeling group, led by Alison Hughes

• Customise an existing MARKAL MODEL for use in this project

• Include Mitigation action in various energy supply and demand sectors

• Model individual policy cases, including, for example, energy efficiency, nuclear, renewables, biofuels, and others

• Model scenarios as decided by the SBT

• Be available for iterations and refinements of policy cases and scenarios

• Provide outputs for the base case, policy cases and combined scenarios

• On emissions

• On costs of reducing emissions

• Other parameters as needed, particularly those reflecting development goals (e.g. energy access)


• Provide model outputs that can be used by the macro-economic modelling team.

2.2 Non-energy emissions modelling

CSIR, led by Rina Taviv together with Bob Scholes

• Compile new or customise existing databases on emissions for sectors not covered by energy (e.g. agriculture, waste, industrial process emissions, land-use change and others with significant GHG emissions in South Africa).

• The database can be fairly aggregated, but should have sufficient resolution to allow for the major potential mitigation policies to be analysed.

• Produce projections of 'non-energy' emissions (a ‘base case’).

• Produce policy cases showing potential mitigation options, quantifying the total emission reductions, mitigation cost (R/ t CO2 –equiv) and, where possible, other parameters relevant to sustainable development.

• Conduct possible iterations and refinements of policy cases and scenarios as required in the process.

2.3 Macro-economic

Development Policy Research Unit, UCT, led by Kalie Pauw

• Provide a technical report that includes an overview of the macro-economic implications of climate change mitigation scenarios. The overall economic implications shall include the direct (mostly energy-related) costs and benefits, but notably including the indirect effects through the economy. For each scenario, the report shall detail the

• direct and indirect costs to economy;

• overall impact on SA economy;

• impacts on job creation;

• impacts on income - winners and losers by sector and household type.

• Conduct macro-economic analysis to support the above, using a Computable General Equilibrium (CGE) model elaborated for this purpose.

• Disaggregate the underlying Social Accounting Matrix (SAM) sufficiently to allow for consideration of electricity produced from different sources (coal, oil & gas, renewables, imported hydro, nuclear, etc) and different refining options for petroleum products (CTL, GTL, refineries, imports, biofuels, etc.).

• Include various groups of households, including income characteristics.

• Include industrial sub-sectors.

• Calculate the appropriate multipliers for income, employment and GDP in the Social Accounting Matrix and report these separately.

• Define the shock (or perturbation) to the economy, based on information of costs in energy system for various scenarios and / or individual policy cases, in consultation with ERC.

• Run several simulations on a number of mitigation actions and scenarios provided to them through CGE.

• Quantify the overall macro-economic costs and benefits of climate change mitigation actions and scenarios, including but not limited to:

• change in GDP;


• change in employment (at various skills levels, e.g. high, medium, low);

• change in income distribution, in particular:

� impacts on particular industrial sectors;

� any impacts on income of poorer households.

2.4 Climate change impacts and adaptation

Led by Guy Midgley with Barney Kgope, SANBI

Study focusing on:

• Risks, impacts and vulnerability assessment. Including costs estimate of direct impacts and avoidance, based on existing information or simple calculations. The costs can be based on orders of magnitude and should be supported by the methodology used.

• Review of current adaptation activities, policies and programmes.

• Options for future adaptation strategies and activities.

Team member Section

Guy Midgley Barney Kgope

SANBI Biodiversity Ecosystems structure

Mark Tadross Bruce Hewittson

CSAG Climate science and new scenarios

Arthur Chapman CSIR Extreme events Ecosystems & processes (incl fire) Coastal & marine zones

Roland Schulze UDW Hydrology & water resources Summer rainfall agriculture

Stephanie Wand U.Stell Winter rainfall agriculture Gina Ziervogel CSAG Livelihoods Akin Abayomi U.Stell Health and Malaria

3. Energy modeling: more detailed parameters The energy modeling approach (described in the technical report) starts from projections of energy demand. Table 1 shows the fuel use by sector for the ‘growth without constraints’ case, to provide an overiew. This appendix describes demand in for each sector in a little more detailed, followed by the major supply industries, namely electricity generation and liquid fuel supply.

Table 1: Fuel use by sector in the GWC case, selected years

2001 2005 2015 2025 2035 2045 2050

Industry 1 206 1 387 1 962 3 014 4 621 6 576 7 689

Transport 634 720 1 112 1 783 2 693 3 677 4 188

Agriculture 73 76 93 129 179 233 262

Commerce 100 112 156 222 301 380 419

Residential 197 209 231 249 256 260 260

Total 2 209 2 504 3 555 5 397 8 051 11 126 12 818


3.1 Agricultural energy demand

Demands for heat, processing energy, irrigation, tractors, harvesters and other energy needs (all in Peta Joules) are met through various technologies and fuel sources. Technologies using liquid fossil fuels (tractors, harvesters and pumps using diesel or petrol) are able to use a bio-fossil fuel blend. Tractors and harvesters are also able to run on pure bio-ethanol or bio-diesel for a case in which a farmer may be producing his own biofuel for use in farm vehicles. Demand for energy increases in time with respect to the agricultural GDP.

Fuels come from refineries or mines, in the case of coal, and dummy boxes along fuel paths allow for accounting for each specific sector.

3.2 Commercial sector demand for energy

The commercial sector is modelled with demands for cooling, lighting, refrigeration, space heating, water heating and ‘other’ demands that are met by various technologies using a range of energy carriers.

The energy demand in the commercial sector is based on the floor space for a given commercial activity. The increase in energy demand is modelled on an increasing floor space area. Floor space projections are generated using regression analyses with the GDP growth projections for various commercial buildings (warehouses, offices etc). These are then summed up to give the total floor space projection. Table 2 below shows the floor space projections from 2000 to 2030 based on the GDP growth. Figure 1 shows, graphically, the projected floor area growth by commercial building type.

Table 2: Floor space projections for the commercial sector

Year Floor space

(million m2)

Year Floor space

(million m2)

2000 75.2 2015 120.5

2001 77.0 2016 124.9

2002 79.1 2017 129.4

2003 81.6 2018 134.1

2004 84.450 2019 138.8

2005 86.4 2020 143.5

2006 88.5 2021 148.3

2007 91.2 2022 153.1

2008 94.2 2023 157.9

2009 97.9 2024 162.7

2010 102.0 2025 167.4

2011 104.6 2026 172.1

2012 106.9 2027 176.7

2013 110.7 2028 181.3

2014 115.2 2029 185.8

2030 190.3


0

50

100

150

200

250

300

350

1990 2000 2010 2020 2030 2040 2050 2060

Million square meter

Other

Healthcare facilities

Education facilities

Offices

Catering

Accommodation

Shops

Warehouses

Figure 1: Floor space growth projection by type

Since most energy use in the commercial sector takes place during business hours, the time of use is very important for modelling the sector. Much of the energy is used for heating or cooling, thus the seasonal dependence plays an important role in energy demand modelling. The percentage of each demand that occurs in a particular time of use period is shown in Figure 2.

0%

20%

40%

60%

80%

100%

Lighting Cooling Heating Other

Winter night

Winter day

Summer night

Summer day

Intermediate night

Intermediate day

Figure 2: Time of use for the commercial sector

3.3 Industrial energy demand

In the model, the industrial sector is disaggregated into three major sectors, gold mining; other mining and the rest of industry. Industry combines iron and steel; non-ferrous metals; non-metallic minerals; pulp and paper; chemical & petro-chemical; food and tobacco; and other)

End use demands are split up into heating (boilers and process heating), cooling, compressed air, HVAC, facility support, lighting, and a few other end use demands. All these demands, besides


boiler heat, are met with electricity. Boilers are fed with an assortment of fuels such as coal, bagasse, heavy fuel oil, as well as electricity for electrode boilers.

How fuel use changes in industry over time is shown in Table 3.

Table 3: Fuels used in industry in GWC scenario, selected years

2001 2005 2015 2025 2035 2045 2050

Coal 613.41 709.76 1022.78 1592.15 2464.34 3529.68 4136.91

Diesel 18.83 20.97 27.48 39.97 58.91 81.59 94.47

Electricity 412.0358 469.7347 642.4479 961.7504 1447.193 2032.615 2365.092

Gas 8.29 9.6 13.87 21.62 33.5 48.01 56.28

HFO 52.1 60.31 87.01 135.53 209.87 300.7 352.47

HRG 14.62 16.9 24.18 37.46 57.78 82.56 96.68

LPG 0.11 0.12 0.17 0.25 0.38 0.54 0.63

Paraffin 0.41 0.46 0.63 0.95 1.44 2.04 2.38

Bagasse 50.77 58.81 84.93 132.4 205.12 293.97 344.62

Biomass 35.21 40.79 58.91 91.83 142.26 203.89 239.01

3.4 Residential energy demand

The vast range of income in South Africa means that the energy demand of households can differ significantly. Higher income households tend to demand more energy through the use of more electric appliances, whereas lower income households use more traditional energy sources via inefficient means. Whether a household is situated in an urban or rural setting also impacts on the energy use and particularly the type of fuel used to meet energy demands. In many rural areas wood is available whereas a similar economic bracket in the city, may be using coal. In order to capture these differences within the model, the residential sector is divided into six different household types. Table 4 below shows the different housing types and the number of households in each type in 2001.

Table 4: Household type and number of households of that type in 2001

Source: Winkler (2006)

Household Number of households

Share of all households

Notes and assumptions

Urban rich electrified (UHE)

4 074 438 36.4% Virtually 100% of rich urban households are electrified

Urban poor electrified (ULE)

1 255 728 11.2% Remainder of urban electrified households must be poor

Urban poor unelectrified (ULN)

1 349 240 12.0% Rest of urban must be non-electrified

Rural rich electrified (RHE)

1 181 279 10.5% Assume 84% of rich rural households are electrified

Rural poor unelectrified (RLE)

1 095 449 9.8% Remainder of rural electrified must be poor

Rural poor unelectrified (RLN)

2 249 571 20.1% Rest of rural households must be non-electrified; number of households includes the few rich rural not electrified


In this study, ‘poor’ households, with regard to energy consumption, are considered to be those in the bottom two quintiles of income (an annual per capita income of less than R4 033). Households that fall into a ‘middle class’ have been included in the ‘rich’ category (Winkler 2006).

Energy demand in the residential sector is divided into cooking, lighting, space heating, water heating and other electrical demands. Originally refrigeration and laundry were included as separate demands, however national data is not available for such disaggregation. Data collection in the residential sector is a difficult and expensive task thus most of the information used in the model is calculated from census data. Census 2001 gives numbers of households that use a particular fuel to meet a specific demand. From these numbers of households, an energy use is calculated given a fuel use per household. The factor of fuel use per household is an approximation and leads to some inaccuracies. In areas that figures look highly unlikely, an expert (Eugene Visagie, Energy Research Centre, 2006) was consulted and numbers were adjusted, keeping total fuel use similar to what was reported in the DME National Energy Balance for 2001.

Demand for energy increases as population increases since with population growth there is obviously an increase in the number of households. There is also an increase in energy demand as households become richer and thus own more appliances and require more energy. This factor is taken into account with the shifting of household types as people get richer or more urbanization takes place.

3.5 Transport energy demand

The modelling of the transport sector is based on previous work done at the ERC (Vessia 2006). One major difference is that in the older version of the South African national MARKAL model, the demand for transport was given in vehicle-kilometres for specific types of vehicles. This made it very different to simulate change from one mode of transport (for example private cars) to another mode of transport (for example buses or trains). The new model allows for more flexibility by setting the demand to passenger-kilometres for passenger transport and tonne-kilometres for freight. With this method one has to assume an occupancy or tonnage for each type of vehicle. These assumptions are given in Table 5 below.

Table 5: Assumptions for occupancy and load for passenger and freight vehicles

Passenger vehicles Occupancy

(persons/vehicle)

Diesel buses 35

Petrol taxis (minibus) 10

Diesel taxis (minibus) 10

Petrol cars 2.1

Diesel cars 2.1

Hybrid cars (diesel) 2.1

Hybrid cars (petrol) 2.1

SUVs diesel 2.1

SUVs petrol 2.1

Motorcycles 1

Diesel freight vehicles Load

(ton/vehicle)

Light commercial vehicle 3

Medium commercial vehicle 10

Heavy commercial vehicle 15


Petrol freight vehicles

Light commercial vehicle 3

When calculating the efficiency for freight vehicles it is assumed that the vehicle is full for half of the journey (ie half the kilometres) and empty for the other half. Fuel efficiency for diesel vehicles is assumed to be 85% the efficiency of petrol vehicles. New vehicles are assumed to have an efficiency of 90% of the given efficiency to account for city driving versus open-road driving as well as a decrease in efficiency with increased age of a vehicle. This value is confirmed by Kwon 2006 (Kwon 2006). In a study performed in Great Britain, it was concluded that while fuel consumption rates may have improved over time, this was partly offset by an increase in the average engine capacity of vehicles. Thus we use an annual efficiency improvement of 0.9% compared with the potential improvement of 1.1% if there was no change in average engine capacity (Kwon 2006). A recent study in the US showed that households do not consider fuel prices when making decisions about vehicle or gasoline purchases (Turrentine & Kurani in press 2006). The trend of buying vehicles with larger engines and the decoupling of fuel prices from types of vehicles purchased highlights the need for government intervention if energy and emissions savings are to be made in the transport sector.

Another addition to the model is the inclusion of separate categories for sport utility vehicles (SUVs) and hybrid vehicles. The cost for SUVs is averaged from the cost of the following Toyota vehicles for both petrol and diesel: Land Cruiser GX, Land Cruiser Pickup, Land Cruiser Pickup Brutus and Land Cruiser Prado VX. Little data is available for sales of these types of vehicles as they are new to the market. This makes it difficult to predict the growth patterns for these vehicles in the future. Research was done on the penetration rates of SUV and hybrid vehicles into foreign markets to get some idea of future penetration rates in South Africa.

The United States Department of Transport estimated that in 2004 there were 24.3 million SUVs on the road versus 137.6 million ordinary cars. With regards new vehicle sales, an estimated 27% of new vehicle registration in 2002 were SUVs (Plaut 2004). In 2004 hybrid vehicle sales made up 0.52% of the market share and was forecast to 3% in 2011 (de Haan et al 2006).

In South Africa the situation is somewhat different since approximately only 5% of households could afford to buy an SUV or hybrid vehicle1 (SAARF 2005). If each household has two vehicles, 10% of vehicles are owned by the top income households and could potentially be SUV’s or hybrids. Keeping in mind percentage of SUVs and hybrids in new cars sales in the US and the fact that the top 10% of vehicles on the road could be SUVs or hybrids, we assumed that or these top 10% of vehicles, by 2035 40% of them will be SUVs and 10% will be hybrids. This equates to 4% of the total fleet of private passenger vehicles consisting of SUVs and 1% of hybrids. These estimates are inline with original estimates.

Demand for transportation is met through these various technologies using an assortment of energy carriers with liquid fuels such as diesel and petrol being the most dominant. The model has the flexibility to include bioethanol and biodiesel into the transportation fuel mix in any ratio specified. While these fuels are not currently used in South Africa on a large scale, with the growing interest in biofuels and the construction of a bioethanol plant underway, this flexibility allows the model to perform more realistic future scenarios. In the base case it is assumed that ethanol and biodiesel will be made available from 2008 and will be blended with petrol and diesel in ratios of 10% and 5% respectively by 2012. Thereafter the biofuels ratios will remain constant.

1 We assume that households with an income of R18 000 per month or higher are able to afford an SUV. These

households fit into LSM (Living Standards Measure) 10 as described by the South African Advertising Research Foundation.


3.6 Power generation technologies

The general characteristics of new power stations were reported in the Technical Report. Further detail describing the various technologies is provided here.

3.6.1 Coal-fired pulverized fue Conventional pulverized fuel (PF) combustion is common throughout the world and the majority of South Africa’s electricity is generated in these types of plants. Finely ground coal particles of coal are blown into the boiler where they are burnt. Heat from combustion is collected through the water-cooled walls of the boiler and a number of heat-exchangers to produce high pressure steam. This steam passes through a steam turbine which in turn drives an electric generator.

Different configurations of steam plant are possible either for electricity-only or cogeneration (combined heat and power) applications. In South Africa most power plants are electricity only, however in the future there may be development of more cogeneration plants.

The temperature and pressure at which the steam is generated is the key design feature of a conventional PF plant. All PF plants in South Africa use sub-critical boilers (the steam pressure is below the critical pressure of water (approximately 218 atmospheres). Supercritical boilers are proven technology that raise pressure above this, thus increasing the efficiency to about 42% from 38-40% efficiency of sub-boilers. Specialised alloys are required to withstand high-pressure steam which increases the cost for components throughout the power plant. In the future most large coal-fired plants will probably have supercritical boilers.

Emissions control is an important cost factor of all PF plants. Current emissions control in South Africa involves basic particulates control but any future coal-plants built will include flue-gas desulphurisation (FGD). We assume all new coal plants include FGD at over 90% efficiency and bag-house filters. The predominant FGD system consists of a reaction vessel in which sulphur dioxide is absorbed from the flue gas stream by a slurry of limestone or other reagent. These systems add cost and reduce generation efficiency of the power plant, however removal efficiencies are some times higher than 95%. NOx control systems relate to the coal combustion itself and involve the flow of air into the combustion zone and the type of burner used.

3.6.2 Fluidized-bed combustion (FBC) This new technology is proven in many countries, however it has not yet been used in South Africa. Coal is burnt in a ‘bed’ or dense cloud of aerodynamically suspended particles. The airflow suspending the particles is sufficiently strong that a portion of the particles is entrained out of the boiler and recirculated back into it via cyclones. Water is heated in the same way as a conventional power plant and steam is raised to turn turbines and drive electric generators. FBCs have environmental advantages over other coal-fired plants:

• Combustion temperatures are generally lower than in a typical PF plant thus reducing the production of thermal NOx.

• The need for expensive FGD equipment can be avoided by injecting sorbent (for example limestone) directly into the fluidised bed boiler. This allows for fuel flexibility as lower grade (high sulphur content) coal can be used.

In South Africa the use of FBCs is particularly attractive since ‘discard coal’ (low grade currently unusable coal) can be used, however when discard coal is used, the emissions from the FBC are worse than from a PF station using higher grade coal. Another disadvantage is that FBC’s have a lower efficiency than sub- or supercritical PF plants are have a higher capital cost (Van der Riet et al 2005). This may explain why the technology still seems far away for large-scale generation but is perhaps most appropriate for onsite generation of coal mines. In FBCs coal can be supplemented with different types of biomass.

3.6.3 Integrated gasification combined cycle (IGCC) Gasification technology increased the coal power-generation cycle efficiency by combining two or more energy cycles: a high-temperature gas turbine cycle and a steam turbine cycle. In most


applications coal is partially combusted in an oxygen-blown gasifier to yield a synthetic gas (syngas) which is predominantly carbon monoxide and hydrogen. The syngas is cleaned before being burnt in a high efficiency gas turbine to produce electric power. The exhaust gases from the gas turbine are cooled in a heat-recovery steam generator (HRSG) and the steam is sent to a steam turbine for additional electricity generation.

The choice to use oxygen rather than air as a source of oxygen for gasification means that components can be smaller as the volume of source gas is smaller and the heating value of the gas produced is closer to that of natural gas. The gas turbine therefore requires less modification to burn the syngas produced in an oxygen-consuming gasifier. Nevertheless the need for a dedicated cryogenic oxygen production facility adds to the cost of the system.

IGCC has the following benefits:

• Cleaning of syngas can result in very low stack emissions, comparable with natural gas fired power stations.

• Efficiencies of up to 48% by utilising advances gas turbine technologies and combined cycle processes.

• Sulphur removal rates are very high (98%) thus systems can be designed to handle fuels with very high sulphur content. Removed sulphur can also be used in the chemical industry.

• Produces a sintered glassy ash which locks in most chemical components found in fly ash.

• Offers the potential to remove CO2 from the syngas for carbon sequestration.

3.6.4 Gas-fired open-cycle gas-turbine (OGCT) An OCGT power plant is basically a simple gas turbine connected to a generator and auxiliary systems such as the fuel supply system, lube cooling system, fire protection system and the control system. In South Africa all current gas turbine power plants are OGCTs run on liquid fuels such as diesel or kerosene. Of the 662MWe of gas turbine capacity in South Africa, about half are owned by Eskom and half are owned by municipalities. These plants are currently used for emergency power or for peaking power.

3.6.5 Gas-fired combined-cycle gas-turbine (CCGT) A new type of gas turbine plant to be used in South Africa is the CCGT. In a CCGT power plant, the gas turbine is usually run on natural gas and the hot exhaust gases are used to generate steam in a HRSG. The steam is then delivered to a steam turbine for additional power generation. In a CCGT plant, about two-thirds of the electrical power is derived from the gas turbine while the steam turbine contributes the remaining third. The greatest advantage of a CCGT is the very high efficiency (50 - 60%), the low capital costs per kWh and the quick construction time.

The first CCGT in South Africa is under construction in New Castle, KwaZulu-Natal and will produce 15MW electricity and 120 000t/h of industrial steam (Le Roux 2006). The plant is owned and operated by an independent power producer and is scheduled to start production in January. More power plants of this type could prove beneficial to the South African power mix provided that gas supply and gas prices are acceptable.

The type of fuel used by a gas turbine plant determines the emissions. Natural gas has lower emissions than liquid fuels, however both gas and liquid fuels are cleaner fuels than coal. Natural gas has little or no sulphur or particulates.

3.6.6 Nuclear power plants: pressure water reactor (PWR) South Africa’s only nuclear power plant, Koeberg, is situated 30km north of Cape Town and consists of two PWR units. Each unit has a capacity of 920MWe and it cooled by sea water. In this system, water inside a pressurised reactor is heated up by the of uranium fuel in the reactor. High temperature, high pressure water is passed through a heat-exchanger to a secondary water system in which steam is produced. This steam drives turbines that generate electricity. Plans for future plants of this type are underway.


3.6.7 Nuclear power plants: pebble-bed modular reactor (PBMR) A nuclear technology in which South Africa has invested a great deal, is the PBMR. This is small, simple, inherently safe design using helium as the coolant and graphite as the moderator. The fuel consists of uranium surrounded by multiple barriers and embedded in graphite balls or ‘pebbles’. The first demonstration module (165MWe) will go into production in 2013 provided that legal and political approvals are granted. Thereafter 24 modules of 165MWe each will be implemented.

3.6.8 Hydroelectric power and pumped storage Hydroelectricity makes use of natural hydrology and topography. Water at a certain height is trapped (usually in a dam) or diverted to pass through turbines that generate electricity. Being a water-stressed country South Africa does not have vast hydroelectricity resources. There are 665MWe of installed hydroelectric power in South Africa of which most is owned by Eskom. Only two of the hydroelectric stations are over 50MWe – Gariep (360MWe) and Vanderkloof (240MWe). While potential for large hydroelectric schemes is limited, there are possibilities for small- and micro-hydro plants.

Pumped storage is not considered a regular power generation facility since it uses electricity at off-peak times to pump water from a lower reservoir into a higher reservoir. This water is then released during peak electricity demand through pump-turbines to generate power. While these stations are net users of electricity, they are important storage systems for load following. The two large Eskom owned pumped storage stations are Drakensberg (1000MWe) and Palmiet (400MWe) while the Cape Town municipality owns the Steenbras station (180MWe). A new pumped storage scheme is planned for Braamhoek on the Free State/KwaZulu-Natal border which will consist of four 333MWe units.

3.6.9 Wind Wind turbines consist of a rotor, generator, directional system, protection system and tower. Wind spins the rotor blades which drives the generator thus turning mechanical energy into electrical energy. Gearing is some times used to increase the rotation speed for electricity generation. The directional system enables horizontal axis machines to orientate themselves into the wind for maximum power. Modern turbines are usually equipped with protection systems such as variable orientation of blades, mechanical brakes or shut-down mechanisms to prevent damage during excessive wind loads. The tower raises the rotor above the ground to capture the greater windspeeds and avoid turbulence caused by ground-interference.

Until the mid 1980s, wind turbines had typical outputs of less than 100kW and rotor diameters from 10m. In the mid 1990s turbines ranged from 0.5WM – 1.5MW and today commercial prototypes of 3.6MW with greater than 80m rotor diameters are being installed. This increase in size of turbines as well as an economy of scale in many European countries that are installing large on- and offshore wind farms, has led to significant reductions in cost.

Currently in South Africa no electricity on the national grid is generated from wind. Nevertheless wind was important traditionally, and continues to be, for water pumping on farms. An estimated 30 000 systems are currently installed (Banks & Schaffler 2006). There are also about 500 wind turbines on farms that generate direct current electricity, usually at 36V.

In 2003, Eskom installed two 660kWh wind turbines and one 1.7MWe at Klipheuvel in the Western Cape as part of the South African Bulk Renewable Energy Generation (SABRE) programme of demonstration and research. An independent group, Darling Independent Power Producer (Darlipp), is an example of an independent power producer in South Africa. The Darling wind farm project in the Western Cape has a planned initial capacity of 5MW with intentions to expand to 10MW.

3.6.10 Concentrating solar power systems Concentrating solar power (CSP) can be exploited though three different systems: parabolic trough, parabolic dish and power tower. All CSP systems make use of a concentrator which captures and concentrates direct solar radiation and delivers it to the receiver. The receiver absorbs the concentrated sunlight and transfers the heat to a power-energy conversion system. The parabolic


trough uses linear parabolic mirrors to reflect sunlight. The parabolic dish system collects sunlight through a round parabolic solar collector and the power tower employs heliostats (large sun-tracking mirrors) to concentrate solar energy onto a central tower-mounted receiver.

The parabolic trough is the most mature of the technologies however the power tower is looking more attractive with its potentially lower cost and more efficient thermal storage. The dish/engine systems can be used in smaller applications.

CSP systems can also be ‘hybridised’ or operated in combination with conventional fossil fuels. For example parabolic troughs can be combined with gas combined-cycle systems.

In South Africa, as part of the SABRE programme initiated in 1998, a 25kW solar dish with a Stirling engine was installed at the Development Bank of Southern Africa in Midrand in 2002. Eskom is also studying the feasibility of building a 300MWe solar thermal power station near Uppington in the Northern Cape. If built, this station would have three 100MWe units concentrating sunlight via heliostats onto a central power tower in which molten salt would absorb the heat. The salt is able to store heat thus allowing the station to deliver electricity 24 hours a day.

3.6.11 Solar photovoltaic systems Photovoltaic (PV) technology transforms the energy of solar photons into direct electric current using semiconductor materials. When photons enter the photovoltaic cell, electrons in the semiconductor are freed, generating direct electric current. The process of converting sunlight to electricity has very low efficiency: Laboratory tests achieve up to 32% efficiency but in practice it is much lower than this. There are many different solar cell designs but the most common semiconductor materials are single-crystal silicon, amorphous silicon, polycrystalline silicon, cadmium telluride, copper indium diselenide and gallium arsenide. The most important PV cell technologies are crystalline silicon and thin films, including amorphous silicon (NEA et al. 2005).

PV cells are connected to form a PV module or panel. PV modules come in standard sizes ranging from less than a watt to around 100 watts. PV modules can be connected together to form an array. In order to obtain useful electricity from the PV array, a number of other elements such as an inverter, batteries, charge controller are required. PV systems can either be used as stand-alone off-grid systems (often applicable in remote areas when extension of the grid is too expensive or infeasible), grid-connected systems in buildings or large utility-scale systems.

In South Africa no electricity from solar power is generated for the national grid but PV systems are widely used in rural areas. It is estimated that about 70 000 households, 250 clinics and 2 100 schools have PV panels. Programmes are in place to increase the number of these systems (Winkler 2006).

3.6.12 Biomass for electricity generation Much biomass is used in South Africa for heating, lighting and cooking in low-income households. The industrial use of biomass is small but significant. Annually South Africa’s sugar industry produces about two million tons of sugar from about 20 million tons of cane. Approximately seven million tons of bagasse is burnt in boilers to make steam for electricity generation and process heat.

The paper and pulp mills in South Africa also use biomass to generate electricity with an estimated capacity if 170MWe. The mills burn sawdust and bark to make steam for electricity generation and process heat. In chemical pulp mills, ‘black liquor’ is separated from wood fibres after passing through digesters. This black liquor is burnt in recovery boilers to make steam. The pulp and paper industry is expanding and there is room for expansion of generating capacity both for onsite use and for sale to the national grid.

Biofuels from biomass such as ethanol (both liquid and gel) and biodiesel are receiving considerable attention particularly for use in the transport sector (ethanol and biodiesel) and residential sector (ethanol gel). These energy carriers are most appropriate for direct combustion and not for electricity generation.


3.6.13 Municipal waste for electricity generation It has been estimated that South Africa’s total domestic and industrial waste disposed in landfill sites has an energy content of about 11 000 GWh per annum. This could be directly combusted or converted into biogas and methane for electricity production.

A project currently underway in the Durban metropolitan municipality consists of enhanced landfill gas capture from three of the city’s landfill sites and use of this gas to generate up to 10MW of electricity. This project is supported by the World Bank’s Prototype Carbon Fund which will purchase the greenhouse gas reductions of 68 833 metric tonnes CO2 equivalent per annum (DSW & PCF 2006; ENS 2004).

4. Data for analysing industrial process emissions

4.1 Introduction

As part of the Long Term Management Scenarios exercise carried out by the Energy Research Centre at the University of Cape Town for the Department of Environmental Affairs and Tourism, options for the mitigation of GHG emissions from industrial processes are to be investigated. After initial investigations by the CSIR, Gerrit Kornelius of Airshed gathered the data contained in this Appendix. ERC assisted in extrapolating baseline emission and mitigation actions reported in some sections.

4.1.1 General assumptions The non-energy sector consists of a number of diverse activities. To ensure meaningful results from a model, the input data needs to be reliable and consistent across sectors. The output from the model has to be structured in the same format as the outputs from the energy sector model to allow for comparison across all sectors.

4.1.2 Methodology The 1990 GHG emission inventory for the industrial sector was used as a basis. A projection to 2003 or 2005 was made based on growth rates, and data that was not considered for the 1990 inventory is included. Emissions from the mining sector, mitigation of which is to be included in this study, are also summarised. This provides the basis for prioritisation of the sectors or sources for which mitigation should be considered in the context of the total South African inventory. Finally, mitigation options under the different scenarios are summarised, together with mitigation costs estimates based on literature data.

4.1.3 Selection of mitigation options To select mitigation options available local and international literature was assessed. The most relevant studies are described for each sector. The most critical general studies used were:

• the previous South African GHG inventories and the associated country studies;

• IPCC assessments and Working Group documents;

• industry estimates for the mining and synfuels sectors.

In addition, the representatives of sectors that form a part of the LTMS stakeholder group, as well as other sector representatives, were consulted and where possible more recent data was incorporated into the models.

The selection of the areas in which additional research and the acquisition of new data is required was based on the relative importance of the sector and relative importance of the error resulting from the uncertainty associated with the existing calculations.

Some models and calculations were updated in cases when new information became available to allow for more accurate modelling. Modelling methodology, emission factors and assumptions are described for each sector.


4.2 Emission estimates

4.2.1 Sector emissions for 1990 The GHG Inventory provided the assessment of emissions for 1990 summarised in Table 6. The GHG Inventory showed zero GHG emissions for solvents, food and drink and pulp and papers sectors. Similarly, The SA First National Communication shows no emissions from the pulp and paper sector (South Africa 2000). The 2000 SA Country study (Borland et al 2000) does provide an estimate for the pulp and paper industry, but this is linked to the energy use in that sector and therefore dealt with elsewhere in the LTMS process.

Table 6: 1990 Emissions from the industrial sector

Scholes & van der Merwe (1993)

Categories Production

(t)

CO2 emission factor

(t emitted/ t produced)

CO2

(Gg)

CH4 emission factor

(t emitted/ t produced)

CH4

(Gg)

N2O

(Gg)

CO2 eq

(Gg)

Tot. industrial processes 23441.29 3.81 23 528.9

A mineral products 12 261 342 7649.29 7 649.3

Cement production 7 751 520 0.575 4457.12 4 457.1

Lime production 1 862 000 0.8 1489.60 1 489.6

Dolomite use 2 340 000 0.673 1574.82 1 574.8

Soda ash production and use

307 822 0.415 127.75 127.7

B chemical industry 1 621 648 2981.74 3.81 3 069.3

Ammonia production 450 173 1.5 675.26 0.0047 2.12 683.3

Ammonia production 287 816 6.995 2013.27 0.0047 1.35 2 018.4

Nitric acid production 274 659 0.009 2.47 2.47

Carbide production 269 000 1.09 293.21 293.2

Emissions from ethylene 285 000 0.0010 0.29 1.1

Emissions from propylene

55 000 0.0010 0.06 0.2

C metal production 12810.26 12 810.3

Iron and steel production 6 256 961 1.6 10011.14 10 011.1

Ferroalloys production 1 796 725 2698.17 2 698.2

FeSi (75% Si) 60 000 3.9 234.00 234.0

FeCr 1 076 000 1.3 1398.80 1 398.8

Si Metal 37 725 1.2 45.27 45.3

FeMn 390 000 1.6 624.00 624.0

SiMn 233 000 1.7 396.10 396.1

Aluminium production 100.95 101.0

Sóderberg 86 500 0.56 48.44 48.4

Prebaked anode process 89 000 0.59 52.51 52.5


The selection of the areas where additional research and the acquisition of new data is critical was based on the relative importance of the sector and relative importance of the error resulting from the uncertainty associated with the existing calculations. Table 7 below provides some estimates for non-energy sources, based on the industrial data in Table 6 only:

Table 7: Uncertainty associated with sector emissions and accuracy of existing models (based on the total national emissions for 1990 of 347346 Gg CO2 eq.)

South Africa (2000)

Sectors

1990

emissions

(Gg CO2 eq)

% of total

% Mitigation potential

(Gg CO2 eq)

Uncer-

tainty

(%)

Error

(Gg CO2 eq)

Error (% of national emission)

(%)

Mineral products 7 649 2.2

Cement 4 457 1.3 30 50 15 0.00

Chemical sector 3 836 1.1

Nitric acid production 766 0.2 80 50 40 0.01

Metal production 15 572 4.5

Iron and steel Production

10 074 2.9 30 50 15 0.00

4.2.2 Update of emission inventory for purposes of this study Sources requiring updating, or not included in Table 6 are the following:

• Nitrous oxide (N2O) from the nitric acid production plants; this was severely underestimated at the time and as the reduction potential is estimated to be 1 436 Gg CO2-e/a. (SA DNA 2007), the 2003 emission rate is estimated to be 1 595 Gg.

• Perfluorocarbons (PFCs) from the aluminium sector, which were not included at all, and will add approximately 0.073 Gg/a of PFC or 472 Gg CO2-e to the 1990 figure, based on 1990 Australian default emission factors (Australia 2004) .

• The emission factors provided by the SA aluminium sector in 1990 may be in tons C (rather than CO2) per ton of aluminium produced, as the ratio between the factor provided and the IPCC default factor is close to the ratio 44/12. Correcting this error would add 100 Gg/a to the 1990 figures.

• Methane emissions from the synfuels process, then estimated at approximately 47 000 ton per annum, with a carbon dioxide equivalence of close to a million ton or 1000 Gg CO2-e are not included in the table. Subsequent estimates indicate this to be considerably higher at approximately 170 000 ton per annum or 3 570 Gg/a CO2-e (Sasol 2003, Marais 2007).

• The synfuels industry releases approximately 23 million ton carbon dioxide in a concentrated form from gas processing (Goede 2007), mitigation of which will be considered in this report.

• Also to be considered in this report are the methane emissions from mining. For the coal mining sector, an estimate of 0.3 kg methane (6.9 kg CO2-e ) per ton of coal has been made, based on sampling at six mines (de Wit 2006). When non-methane emissions are included, the intensity is 29.1 kg CO2-e per ton of coal, resulting in an estimated total release of 6.55 million ton CO2-e at 2003 production levels. For gold mining, initial measurements on one shaft have been carried out; the data obtained is of too preliminary a nature to extrapolate to the entire gold/platinum group metals sector, but indicates that the concentration is too low, by some orders of magnitude, to utilise or oxidise cost-effectively (Human 2007). Methane emissions from sectors other than coal mining are therefore not addressed in this study.


4.3 Estimates for 2003 emissions

The production quantities and emission factors of the major non-energy sources listed in section 2 were updated to 2003/4 (the latest year for which figures are available) for most of the sources in order to (i) provide a basis for prioritisation in that year and (ii) act as the baseline for emission time series for the sources

4.3.1 Mineral products

4.3.1.1 Cement production

Cement production and emissions figures were provided by the Association of Cementitious Material Producers (ACMP) as given in Table 8 below. It was noted that the CO2 emissions from the cement production process have two main sources: calcination and fuel firing. Calcination is the process whereby CaCO3 dissociates into CaO and CO2 through the application of energy. The IPPC default factor for calcination is 575 kg CO2 per ton of clinker produced from limestone. In addition to CO2 generated through calcination, CO2 from fuel firing also needs to be considered. When CO2 generated from fuel consumption is included, the total CO2 generated per ton of cementitious material produced, was 930 kg per ton. The increased use of alternative sources of pozzolanic materials (slags from the metallurgical industry, fly ash from pulverised coal firing) has lead to a lower clinker content of finished cement and a concomitant reduction of GHG emissions per ton of final product as indicated in Table 8 below.

Table 8: ACMP production and emission values (combustion and dissociation).

Year Total clinker produced (ton)

Total cementitious material produced

Total CO2

produced (ton) Specific emissions kg/ton cementitious material

1990 7 770 000 8 450 000 7 858 500 930

2003 7 346 239 9 511 469 6 798 178 715

The ACMP indicates that they believe that cement production should follow the GDP trend; however, the same growth rate has been used in this case as in the MARKAL energy model for this sector – resulting in an output of around 20% lower by 2050 than if the sector had grown at the assumed GDP growth rate. The sectoral growth rate for the cement sector has been increased for the period up to 2015, based on the ACMP comments.

4.3.1.2 Lime production and dolomite use

Statistics are provided by DME (2003). The report indicates that production and sales are variable, depending somewhat on exports to adjacent territories and on agricultural conditions.

Table 9: Change in lime production and emissions

Year Production (ton) Total CO2 emitted Specific emissions kg CO2/ton lime

1990 1 862 000 1 489 600 800

2002 1 700 000 1 360 000 800

Table 10: Change in limestone and dolomite use

Year Use (ton) Total CO2 emitted Specific emissions kg CO2/ton

1990 2 340 000

dolomite only 1 060 200 453

2002 3 393 000 1 425 060 Approx 420


In the absence of alternative data, these emissions are assumed to follow the same trend as those for cement production.

4.3.2 Chemicals

4.3.2.1 Ammonia production

Implied production in 1994 was 762 000 ton (South Africa 2000) at an emission factor (for production in the synthetic fuels process) of 2.45 ton CO2 per ton ammonia produced. Production increased somewhat (12% in Sasolburg, Sasol 2003) due to the switch to natural gas at Sasolburg. Estimated production in 2003 was therefore approximately 775 000 ton, and emissions at a constant factor 1 892 Gg. CO2

4.3.2.2 Nitric acid production

As given under section 2.1.3, emissions from the four operating plants are estimated to be 1 595 Gg.

4.3.2.3 Carbide production

Production has decreased from the 1990 figure to 70 000 ton per annum in 2006. At a constant default emission factor, this represents emissions of 76.3 Gg/a.

4.3.2.4 Balance of the chemical sector

As the emissions are minor compared to the ammonia/nitric acid streams, these will not be considered further.

4.3.3 Metals

4.3.3.1 Iron and steel industry

Production data for the iron and steel industry is available for 2003 (www.saisi.co.za.) Available on this site are the production figures and predictions for 2003 to 2007. The production of pig iron and direct reduced iron for 2003 is 7.8 million ton/a, which means that the growth in this industry was approximately 1.5%/a since 1990. It is suggested that a constant emission factor (as provided by the 1990 GHG Inventory) should be used for the year 2003, as no material process upgrades occurred between 1990 and 2003.

4.3.3.2 Ferro-alloy production

Comparative production figures are given in the table below.

Table 11: Ferro-alloy production 1990 and 2004

Source:1990 inventory; DME (2006)

1990 2004

Commodity Production (tpa)

CO2 emission (Gg)

Production (tpa)

CO2 emission (Gg)

FeCr 1 076 000 1 399 2 900 000 3 770

FeMn 390 000 624 575 000 865

FeSi 60 000 234 106 000 413

FeSiMn 233 000 396 300 000 510

Si 37 700 45 50 000 60

Totals 2 698 5 618

This represents a compound growth rate in emissions of 5.4% annually over the 14 years. IFM has commissioned portions of a ferrochrome smelter with a final capacity of 267 400 ton per annum in 2006 (Engineering News 23 Feb 2007), while Xstrata’s Lion project added 250 000 ton in 2007 from


a plant that is planned for an eventual 1 000 000 ton per annum. The growth rate has therefore continued at approximately 5.5% per annum between 2004 and 2007.

4.3.3.3 Aluminium

As indicated earlier, in the original 1990 emission inventory, the 1990 CO2 emission were probably underestimated and the figure given in the 2000 First National Communication (249 Gg CO2 only) is probably more correct. The PFC emissions were not calculated for 1990 but probably were of the order of 472 Gg CO2-e. Since then, South African aluminium production has increased considerably, but aluminium cell technology has reduced PFC emissions approximately fourfold (International Aluminium Institute 2006). A comparison of emissions is given in Table 12 below. Thus, although production increased almost fivefold, estimated CO2-e emissions increased at only half the rate.

A further 720 000 ton per annum capacity is being planned for Coega, presently expected to start producing in 2010 (Engineering News 23 February 2007). As this will use the latest Alcan technology, specific emissions can be expected to be low.

Table 12: Aluminium production and emissions

1990 2004

Production ton CO2-e emission Gg (CO2 + PFC)

Production ton CO2-e emission Gg (CO2 + PFC)

175 500 761 865 000 2 010

4.3.4 Mine emissions

4.3.4.1 Coal mining

As indicated earlier, baseline emissions for coal mining for 2003 are estimated at 29 kg/ton CO2-e coal produced in total or 6 550 Gg /a. Of this 6.9 kg/ton CO2-e/ton, or 1 558 Gg/a is in the form of methane.

4.3.5 Synfuels-specific emissions

4.3.5.1 Methane emissions

The Rectisol process used for cleaning the syngas stream of CO2 and sulphur compounds after gasification also captures methane and other light hydrocarbons. The CO2 containing 2-3 mole% of total hydrocarbons is released to the atmosphere after removal of the hydrogen sulphide and contains approximately 178 000 ton per annum of methane, equivalent to 3 738 Gg CO2-e/a(Sasol 2007, Marais 2007)

4.3.5.2 Concentrated carbon dioxide streams

The Rectisol process produces a number of gas streams from its different expansion stages, most of which contain in excess of 95% CO2 and 2-3 mole % total hydrocarbons (see previous section) after removal of the sulphur compounds. The Benfield process, situated in the Synthol gas loop, similarly produces a concentrated CO2 stream. At the Secunda operations, these two streams together release approximately 23 000 Gg CO2.

4.3.5.3 Expansion of synfuels production using natural gas.

The use of natural gas in the synthetic fuel production process was initially intended to increase production at the Secunda facility by approximately 15% but at a much lower specific CO2 emission rate. The possibility exists of using gas that is available (i e not taken up by the industrial fuels market) for further expansion of liquid fuel production (GTL or gas-to liquids). This would replace imported liquid fuels presumably derived from crude oil. It has been shown that the use of GTL fuels, from a life-cycle approach, leads to approximately the same lifecycle carbon emission than crude-derived fuels. (IEA 2004). From a GHG emission point of view, this use of natural gas in GTL would have a neutral effect on GHG emissions and is not considered further here.


4.3.5.4 Expanded coal to liquids production.

Sasol is at present carrying out a pre-feasibility study into building a CTL facility on the Waterberg coal field with a capacity of approximately 80 000 barrels per day (Engineering News 23 February 2007). If the same technology as the Secunda plants is used, this would produce approx. 11 000 Gg/a CO2 as concentrated stream. Carbon capture and storage (CCS) forms part of the investigations. Improvements in the Rectisol process would probably result in much improved capture of the light hydrocarbons, so that methane emissions would be minimal. Sasol have also indicated (van der Walt 2007) that a further five plants of similar capacity could be built without coal and water supply constraints being exceeded.

4.3.6 Summary of baseline (2003) non-energy emissions

Table 13: Baseline (2003) GHG emissions, industrial (non-energy) sector

Gg CO2-e /a Growth rate of commodity production

Cement production 6 798 (incl. fuel) Markal elasticity plus higher emission growth to 2015

Lime production 1 360 Markal elasticity

Limestone/dolomite use 1 425 Markal elasticity

Ammonia production 1 892 Markal elasticity

Nitric acid production 1 595 Zero

Carbide production 76 4.5% pa 2004-2007; thereafter Markal

Iron and steel production 12 494 Markal elasticity

Ferro-alloy production 5 618 5.4% pa 2004-2007; thereafter Markal

Aluminium production 2 010 83% between 2006 and 2011; thereafter Markal

Coal mine methane 6 550 Coal production growth rate

Synfuels concentrated carbon dioxide

23 000. See text 3.5.4.

Synfuels point-source methane

3 738 0

From this table, the main emission sources to be considered for mitigation are those from synfuels production, coal mining, iron and steel, ferro-alloy production, aluminium and cement.

4.4 Emissions for baseline and mitigation scenarios, including cost estimates for reduction

4.4.1 Synfuels point-source emissions The Rectisol concentrated carbon dioxide stream also containing light hydrocarbons presents a number of GHG mitigation possibilities as follows:

• The entire stream could be used for carbon storage or sequestration, the ‘capture’ step of CCS already having been carried out. The possibilities for carbon storage or sequestration in South Africa were summarised by Engelbrecht et al. (2004). The possibilities for capture in geological formation are considered to be unfavourable due to the poor porosity and permeability of even the more promising formations. Lately, the possibility of sequestration in coal seams in Botswana, in conjunction with methane recovery from the formations is being investigated by Sasol (Liebenberg 2007). Should favourable geology be encountered, transportation cost for 5-40 million ton of CO2/a is estimated to cost $1 to $8 per ton (distance up to 250 km), with the storage itself, including monitoring and verification $0.6 to $ 8.3 per ton (IPPC 2006).


Should this possibility be used for Secunda Rectisol and Benfield off-gas, the cost for transport would be $3 to $24 per ton of CO2 (if transport cost is taken to be linear with distance), with storage/sequestration cost probably at the higher side of the scale due to unfavourable geology, thus a total of approximately $33 per ton.

Sasol has made some preliminary cost estimates for partial CCS, subject to their estimate of capacity, at R62.65 (assuming R7/$ exchange rate, real terms, year uncertain) per ton of CO2, and max storage capacity starting at 1 Mt per year in 2007, and increasing to 20Mt per year in 2030 remaining constant after that (Goede 2006). This estimate probably does not include transport cost.

A more detailed estimate can be made using methodology developed by the Institute of Transportation Studies at the University of California at Davis (McCollum & Ogden 2006, Freund and Davison n.d.)

The cost of sequestration of the entire Secunda concentrated GHG stream of approximately 23 million tons CO2e per year into suitable (presently economically unminable) coal seams at 500 m below surface at 400 km from Secunda is as follows (in 2006 US$):

Capital cost compressors/pumps $220 million

Capital cost pipeline $416 million

Capital cost at the well field (for the most pessimistic case about 4000 wells would be required) for drilling $634 million and other equipment $50 million

Running cost side (in 2006 USD per year)

Compressor/pumping power 250 to 300 MW. Other on-site costs at the well field 181 million.

If, however, two million tons per annum is taken to be the realistic maximum for CCS, as this is the maximum that has in practice been achieved in a single field, the economy of scale on the pipeline, which dominates the capital cost, is lost. For two million tons per year, the pipeline cost is approximately $160 million and the cost of pumps and compressors approximately $50 million, both in 2006 values. Running cost would be approximately linear with tonnage.

• The light hydrocarbons in the Rectisol off-gas stream including the methane could be combusted, either as is or with pre-concentration of the hydrocarbons. For combustion of the stream as is, oxygen and possibly supplemental fuel would need to be supplied; alternatively the stream needs to be fed to an existing combustion installation such as a boiler. If pre-concentration proves to be feasible, the hydrocarbon stream could be used to as additional feedstock to the Synthol operation. A suitable technology is however not available for the latter separation.

A preliminary estimate of the cost of regeneratively oxidising the light hydrocarbons in this stream, using technology developed for coal mine ventilation air, is US$19 million in 2002 values. Annual operating cost is given as 10% of this cost. (US EPA 2002) Due to the present uncertainties, the possible credit for steam raised is not included in this number.

Conclusion:

Scenario 1 (maximum development). Future CTL plants will probably include the Selective Rectisol process, which will virtually eliminate the methane emissions. However, each Sasol 2/Sasol 3 size installation (approximately equal to 80 000 barrels per day) will emit approximately 11 million ton per annum CO2-e.

Intermediate scenarios: Capture and storage from the existing Secunda installation would probably not be feasible under these scenarios. New installations expressly designed for CCS and possibly situated near favourable geological formations could have a somewhat reduced cost. Reduction of the present methane emissions could however be feasible under these scenarios.


Scenario 4 (reduction required by science) would probably require sequestration of all CO2 from existing and new plant at the cost given above.

4.4.2 Iron and steel industry The production growth factors for the period after 2003 (from 7.8 million ton pig iron per annum) should be based on the indexed growth rates calculated from the elasticities used for this sector in the MARKAL model. A constant emission factor (as provided by the 1990 GHG Inventory) of 1.6 ton CO2/ton of pig iron should be applied for the baseline scenario between 1990 and 2003.

Modernised processes could increase the energy use efficiency while simultaneously reducing the emission of pollutants. Examples of these processes are: continuous casting technology, recovery and utilization of gas from steel converters (this possibility should be included in the energy sector) and the use of waste heat for preheating. All of these are typically applied to new plant rather than retrofitted to existing plant.

Conclusion

For Scenario 1 and the intermediate scenarios, no material reduction in the emission factor is to be expected; the increased production rate should be multiplied by the constant emission factor. The possibility for using biomass-based reductants is limited by the SA production potential of such material.

For scenario 4, the IPCC estimates (IPCC 2006) that a reduction specific in emissions of between 20 and 33% could be achieved at $20-$50 per ton reduction; this could be phased in from 2020 onwards. The reductions would mainly be in the form of energy efficiency measures and should be allowed for in the energy sector report.

4.4.3 Ferroalloy industry Similar to iron and steel. (The SA DNA has listed an off-gas utilisation project on its website with a potential savings of 600 Gg/a. This represents a significant reduction in terms of energy efficiency; further details are not given).

4.4.4 Nitric acid production The CDM mechanism plays a major role in the mitigation of N2O emissions. The mitigation of N2O emission from nitric acid production is under consideration by all SA manufacturers. AECI, Sasol and Omnia already have registered CDM projects using this technology. Reduction of 80% of N2O emissions from the figures given in Table 13 will be achieved within four years and 95% of N2O by 2020 at essentially zero cost due to the income generated from CDM. No further reductions are expected thereafter.

Members of the Fertiliser Society of SA (van der Linde 2007) indicate that they believe nitric acid production to be largely driven by fertiliser demand, which has been static for some time. New nitric acid production capacity is therefore not anticipated, and emissions expected to remain static.

4.4.5 Balance of chemical industry Potential for reduction in the emissions over the review period is immaterial.

4.4.6 Cement production Modern cement production processes already applied in the local industry include using the dry process with pre-heating for cement production, clinker blending and the use of substitute pozzolanic materials such as blast furnace slag and classified ash from pulverised coal firing in power generation. The cement industry expects to see the clinker content of all cementitious binders used in 2030 to be about 60 %.(Cluett 2006). In South Africa, this approach will reduce the emission factor to 650 CO2 kg per ton cementitious material (compared to 715 CO2 kg/t in 2003) by 2010.

The substitution of coal by suitable wastes (e.g. sludges from the petrochemical sector, scrap tyres etc.) is being actively pursued by the local cement sector. A Korean study (Dong-Woon Noh 2006) indicates that a 30% replacement of bituminous coal by scrap tyres could lead to a further 1.2 %


reduction in overall emissions at a negative cost. The technology has been tested in South Africa and awaits a regulatory framework for implementation. The possibility of using other alternative fuels in South Africa could increase the reduction to 1.5% at minimal additional cost.

Conclusion: The above reductions apply under Scenarios 1 to 3 – reduction in clinker content is assumed to be ‘business as usual’, and is thus not considered to be a mitigation option. A 60% as required under scenario 4 could only be achieved a cutback in production, as the CO2 released by the dissociation of lime cannot be avoided.

4.4.7 Mining The possibility of reducing emissions from the gold and platinum sectors is not sufficiently researched for inclusion as a possibility.

In general, the concentration of methane in coal and in the ventilation air in South Africa is too low to warrant profitable recovery or destruction under CDM, which implies a cost in excess of approximately $25 per ton CO2-e reduced. No reduction is therefore assumed under scenarios 1 and 2. Under scenario 3, some of the higher concentrations could become feasible at the lower end of the estimated costs (say 25% reduction @ $15 per ton) Assuming that an extended programme of identification would be required, start of such an in initiative should not be expected before 2020, with 25% being achieved by 2030. Under scenario 4, 50% reduction could be achieved at the higher end of the cost estimates ($30 per ton CO2-e), again starting in 2020 and achieving full scale implementation in 2030

4.4.8 Aluminium production It is assumed that new capacity will already be equipped with technology to reduce the PFC emissions to a figure of less than 0.001 Gg CO2-e per ton of aluminium (IAI 2006). This is applicable to all scenarios. For existing capacity, a further 45% reduction in emissions is estimated at $15-$30 per ton CO2-e reduced. (US EPA 2006). Cost estimates for this reduction have been scaled down from a CDM project in Argentina: for a reduction of 0.71 Mt CO2-eq, capital costs are R1.2 million, and operating costs are R0.15 million per year.

Conclusion

Under Scenarios 1 to 3, future emissions from existing capacity can be calculated by multiplying 2003 production by the emission factor of 0.00232 Gg per ton of production. Additional production above that baseline can be calculated at 0.00128 Gg/ton aluminium produced.

Under scenario 4, a 45% reduction in CO2-e emissions per ton of 2003 production capacity can be achieved at $15 to $30 per ton. New capacity installed after 2003 can be assumed to have minimal reduction potential.

5. International experiences on mitigation for agriculture

This section summarises the Pew Centre publication, Global Climate Change, ‘Agriculture’s role in Greenhouse gas mitigation’, (Paustian, K. et al. 2006). The US experience described in this article can be used as a benchmark for the role that agriculture can play in GHG mitigation in South Africa.

5.1 Introduction

Agriculture currently contributes substantially to GHG emissions but potentially it has the ability to act as a sink for CO2 as well as to reduce its GHG emissions at a relatively low cost. Overall, land use change (predominantly in the tropics) and agricultural activities globally account for about one-third of the warming effect from increased GHG concentrations (Cole et al. 1997). However, ecosystem processes also act to dampen these GHG increases, primarily through the uptake and storage of CO2 in plants and soil on land and in oceans. These uptake and storage processes - referred to hereafter as carbon ‘sinks’ - play a significant role in the global CO2 cycle, so that only


about one-half of the CO2 emitted from fossil fuels accumulates in the atmosphere. The other half is absorbed by the oceans and terrestrial ecosystems (IPCC 2001). In this report the term mitigation is used to encompass both GHG emission reductions and GHG removals from the atmosphere by sinks.

Over the past decade, US agricultural soils overall have acted as a small net sink of approximately 12 million metric tons (MMT) of carbon per year, mainly due to improved soil management practices and the establishment of conservation reserve lands (USEPA 2006). These practices are helping to sequester about 23 MMT of carbon per year in mineral soils, which make up greater than 99% of annual cropland area.

Cultivated organic soils and agricultural liming contribute substantial GHG emissions - taking into account both soil emissions and sinks the result is a net sink of 12 MMT of carbon per year.

Since 1850 an estimated 160 billion metric tons of carbon from biomass and soils have been emitted worldwide as a consequence of land use and land-use changes (Houghton 2003) compared to their condition under native vegetation. Since the 1940s, as a result of improved productivity and cropping practices, controlled erosion and reduced tillage, organic carbon stocks of many agricultural soils have started to increase resulting in these soils becoming a net sink. Reforestation has also contributed to the present carbon sink.

Current and future trends in the structure of American agriculture will affect both future emissions and opportunities for GHG mitigation. Increased crop yields, along with continued adoption of conservation tillage and maintenance of conservation set-aside programs are likely to support further increases in soil carbon stocks. Higher crop yields also increase the potential for shifting some land from food production to energy crop production. The reduced use of nitrogen fertilizer since 1990 has resulted in emissions from this source remaining constant. Since 1990, a decline in cattle and sheep populations has been counterbalanced by a rise in swine and poultry populations, resulting in roughly stable agricultural methane emissions (USEPA 2006).

The current technical potential to mitigate GHGs through improved agricultural practices over the next 10 to 30 years is substantial. However, the mitigation levels that can be achieved economically are likely to be substantially lower than these technical potentials. This is because a variety of economic and social factors will influence the adoption of alternative practices and production systems, although studies to date suggest that a significant portion of agricultural mitigation practices can be characterized as low-cost options.

What needs to be considered is that changes in land use and management to achieve GHG mitigation can contribute to overall environmental improvements. Hence, a broader consideration of the costs and benefits of improved agricultural practices, beyond the realm of climate change concerns, is merited.

5.2 Mitigation opportunities: Increased sinks and reduced emissions

5.2.1 Opportunities to increase soil carbon Historically, agricultural practices have caused large carbon losses from US cropland soils. If half or more of the original carbon stock of croplands could be regained, tens to hundreds of millions of metric tons of carbon could be stored (i.e., added to and sequestered) in soils annually over the next several decades.

Management practices that favour carbon additions to soil:

• increase of plant residues;

• slowing the rate of soil organic matter decay;

• land-use changes such as conversion of annual cropland to grassland or forest and restoration of degraded lands.


Employing these practices could result in soil carbon increasing for 20 to 30 years, after which it would tend to stabilize (CAST 2004).

5.2.1.1 Cropland management

Carbon inputs to soil can also be increased by:

• increasing the productivity of crops which is largely in line with farmers’ management goals of achieving high productivity;

• using crop rotations with high residue yields;

• reducing or eliminating the fallow period between successive crops in annual crop rotations;

• making efficient use of fertilizer and manure.

On annual croplands, soil carbon losses can be reduced by:

• decreasing the frequency and intensity of soil tillage, in particular through conversion to no-till practices.

Use of high-residue crops and grasses. Annual crops that produce large amounts of residues, such as corn and sorghum, as well as perennial grasses typically result in higher soil carbon. Cereal-hay rotations would therefore serve to increase soil carbon content.

Reduction or elimination of fallow periods between crops. New cropping systems which do not allow for a fallow season, have proved successful in both improving soil moisture and increasing soil carbon (Peterson et al. 1998). If cover crops, such as legumes or annual grasses, are planted during the winter season they not only take up excess soil nutrients (e.g., nitrogen) to reduce leaching or other losses to the environment, fixing atmospheric nitrogen (e.g., legumes), and controlling weeds; but they also serve to augment the input of plant residues, thereby increasing soil carbon content.

Efficient use of manures, nitrogen fertilizers, and irrigation. If more than the optimum input of fertilizer, manure and irrigation are used for high rates of crop production (with attendant carbon input increases), the increases in other GHG emissions, particularly nitrous oxide, can offset part or all of the gains in soil carbon. Tailoring fertilizer and manure applications to satisfy crop nitrogen demands, so that less nitrogen is left behind in the soil, can reduce nitrous oxide emissions while building soil carbon stocks. Efficient use of irrigation water will similarly reduce nitrogen losses including nitrous oxide emissions, and minimize CO2 emissions from energy used for pumping while maintaining high yields and crop-residue production.

Use of low- or no-till practices. Reducing soil carbon losses on croplands is primarily accomplished through reducing the frequency and intensity of soil tillage. Traditional tillage methods, which fully invert the soil, cause the greatest degree of disturbance and consequently tend to cause the most degradation of soil structure and loss of soil carbon stocks. In many areas, the trend over the past several decades has been towards reduced tillage practices that have shallower depths, less soil mixing, and retention of a larger proportion of crop residues on the surface.

No-till, a practice in which crops are sown by cutting a narrow slot in the soil for the seed, and herbicides are used in place of tillage for weed control, causes the least amount of soil disturbance. Ogle et al. (2005) analyzed data from 126 studies worldwide and estimated that soil carbon stocks in surface soil layers (to 30 centimeter [cm] depth) increased by an average of 10 to 20% over a 20-year time period under no-till practices compared with intensive tillage practices. The relative increases in carbon stocks were higher under humid than dry climates and higher under tropical than temperate temperature regimes. Finally, CO2 emissions from machinery use are decreased by 40% for reduced tillage and 70% for no-till, relative to conventional tillage (West & Marland 2002), contributing to further reductions in GHGs from reducing tillage intensity.

5.2.1.2 Grazingland and hayland management

Permanent grasslands used as pastures, rangelands, and hayfields can maintain large soil carbon stocks due to several characteristics. Perennial grasses allocate a high proportion of photosynthetically fixed carbon below ground, maintain plant cover year-round, and promote the


formation of stable soil aggregates. Grassland systems that have been degraded in the past or maintained under suboptimal management conditions are most conducive to sequestering additional carbon with improved land management.Intensive management strategies are usually restricted to more humid regions with high productivity potential or to regions where irrigation is used.

Conant et al. (2001) summarized more than 115 studies of grassland management effects on soil carbon and estimated rates of soil carbon increase ranging from 0.1 to 3 t/ha/a. The highest rates occurred with introduction of deep-rooted African grasses in South American savannas (Fisher et al. 1994).

5.2.1.3 Land-use changes to increase soil carbon

Conversion of annual cropland to grasslands or forest and restoration of severely degraded lands offer significant opportunities to increase soil carbon. Converting cultivated cropland to grassland typically increases soil carbon at rates of 0.3 to 1.0 t/ha/a for a period of a few decades (Lal et al. 1998; Conant et al. 2001).

Highly degraded sites, such as severely eroded areas, reclaimed surface mines and saline soils represent situations with high potential carbon sequestration rates but also higher costs and technical difficulties associated with the reclamation.

Cultivated organic soils represent another land restoration opportunity. These lands are a significant source of agricultural CO2 emissions, with high rates of up to 10 to 20 t/ha/a of carbon (Ogle et al. 2003). Hence, wetland restoration may be a mitigation option. However, restored wetlands may emit methane, which would need to be considered in assessing the overall mitigation potential of this type of restoration.

5.2.1.4 Total agricultural soil carbon sequestration potential

Carbon sequestration rates vary by climate, topography, soil type, past management history and current practices. Various global and national estimates for potential soil carbon sequestration have been made. These estimates are usually based on overall carbon gain for a suite of practices and the available area on which these practices could be applied, resulting in estimates of biological or technical potential.

However there are numerous uncertainties surrounding such estimates of carbon sequestration potential. On the one hand, development of new technologies specifically targeted at increasing soil carbon (through plant breeding or new soil amendments) could increase potentials. On the other hand, rising temperatures due to global warming will likely stimulate soil organic matter decomposition, which may reduce or eliminate the potential to further increase soil carbon stocks.Finally, the amount of carbon sequestration which is actually achieved will depend on economic, social, and policy factors.

5.2.1.5 Reducing agricultural nitrous oxide and methane emissions

Nitrous oxide (N2O) and methane (CH4) emissions result from both crop and livestock operations and account for approximately 80% of U.S. agricultural greenhouse gas emissions on a GWP basis. Despite challenges, there is considerable scope for reducing these emissions.

Nitrous oxide constitutes the largest agricultural source of GHG emissions in terms of warming potential (48%), and almost 70% of total US nitrous oxide emissions are from soils. The best option for reducing these emissions is to use fertilizers more efficiently; adoption of best fertilization practices could reduce agricultural N2O emissions by 30 to 40% (CAST 2004). Livestock are the main source of agricultural CH4 emissions. Increasing the efficiency of production (meat, milk) per animal can decrease these emissions and also reduce costs. Manure management accounts for 25% of U.S. agricultural CH4 emissions; anaerobic (i.e., oxygen-free) digesters that capture and use the methane as an energy source— thereby displacing fossil fuels—offer a nearly ideal solution for these emissions.


5.2.1.6 Reducing nitrous oxide and methane emissions from soils

A characteristic of modern agriculture is the huge increase in nitrogen supplied—not only as mineral fertilizer but also through nitrogen-fixing crops (e.g., alfalfa, clover, and soybeans) and animal manure—to boost crop productivity (Mosier et al. 2001). Methane emissions from agricultural soils are mainly associated with flooded soils such as rice-growing areas and wetlands. Most soils are not a major source of CH4 and, in fact, most non-flooded soils remove some CH4 from the atmosphere.

Nitrous oxide: Unlike the case for CO2 and CH4, there are no significant biological sinks for atmospheric N2O. Since agricultural N2O emissions correlate with the amount of nitrogen available in soils, mitigation rests largely on increasing the efficiency of nitrogen use without compromising crop yields.

Greater than 50% of the major cropland area in the United States is rated as having high nitrogen balances, resulting in soils highly susceptible to losses of N2O to the atmosphere and nitrate (NO3 –) to water bodies (USDA 2003)

Both the application rate and timing are factors in the efficiency of nitrogen use. The application of fertilizer after the start of the growing season provides better synchrony with plant demands. Slow-release fertilizers, such as sulfur-coated urea, which delay the release of fertilizer applied at planting time until plant nitrogen uptake capacity is higher, can also be used. Injecting fertilizer and manure into the soil, near the zone of active root uptake, both reduces nitrogen losses and increases plant nitrogen use, resulting in less residual nitrogen that can be lost as N2O.

5.2.1.7 Reducing livestock-related methane and nitrous oxide emissions

Livestock-related emissions from digestive processes and animal wastes account for 26% of total agricultural emissions. Although enteric (digestive tract) emissions are more significant (70% of agricultural CH4 emissions), emissions from livestock wastes have a greater potential for mitigation. Improving manure-handling facilities, for example by covering animal-waste lagoons and capturing and burning the CH4, can reduce emissions while providing renewable energy and income. Capture and combustion of CH4 from animal wastes also reduces other environmental problems, including odours and nitrate pollution. Overall the best option for reducing digestive process emissions is to increase the efficiency of livestock production.

Manure storage and management. Manure management in the United States currently accounts for 25% of agricultural CH4 and 6% of agricultural N2O emissions. In addition to GHG production, problems associated with odour and nutrient pollution from animal wastes are widespread. Hence, improvements in manure handling that address both GHG reductions and odour and nutrient problems are of great interest.

Manure produced by livestock can emit N2O and/or CH4 during storage and following application to soil. In general, storage under anaerobic conditions (lacking oxygen, such as in waste lagoons) will produce CH4 while N2O emissions will be suppressed. Conversely, piled storage and composting of manure will promote largely aerobic decomposition, suppressing CH4 emissions but promoting N2O emissions. Anaerobic digesters in conjunction with lagoon storage systems offer a nearly ideal option – N2O emissions are suppressed and CH4 can be used as an energy source, thereby displacing fossil fuels.

Opportunities for mitigating N2O emissions from stockpiled or composted manure are relatively limited. Perhaps the most effective measure for reducing manure-related N2O emissions from stockpiled or composted manure is to apply the manure at rates based on crop needs, thus maximizing plant uptake of manure-derived nitrogen.

Enteric fermentation. Methane is produced in the digestive tract of animals, particularly in ruminants such as cows, sheep, goats, and camels. This source of CH4 emissions is termed enteric fermentation. In the United States these emissions amount to about 70% of agricultural CH4 emissions and 20% of total agricultural GHG emissions on a carbon-equivalent basis.

Because CH4 emissions from enteric fermentation are influenced by the feed quality and digestive efficiency of the animals, improving these will reduce CH4 emissions. In simple terms, the more


rapidly food is processed and passed through the rumen (first stomach of ruminants), the less time there is for CH4 production. Where feed quality and digestibility are already at a relatively high level, further improvements from conventional changes in feed rations are likely to be modest. However, where diets are not optimal, improvements in the diet can reduce emissions. One area where substantial improvements are possible is in improving forage quality for grazing animals on smaller livestock operations through better pasture management (DeRamus et al. 2003). Various feed additives such as edible vegetable oils and certain antibiotics can also be used to inhibit the rumen bacteria that produce CH4 (Teather and Forster 1998).

For a given animal type and food quality, CH4 production will be roughly proportional to food intake. Thus, increasing the amount of product (meat, milk) per unit of food consumed will effectively reduce CH4 emissions per unit of product. Ways to increase the production efficiency of individual animals include improved animal genetics (breeding) and animal health.

6. Methodology for modelling emissions from livestock enteric emissions

6.1 Historical data, assumptions and calculations for enteric fermentation

The model for the agricultural sector developed and used for the SA Country Study on Climate Change (Scholes et al. 2000) has been used as a basis for this study. It was updated using latest data from agricultural statistics and extending the calculation for 50 years. Most of the data on livestock population was extracted from Abstract of Agricultural statistics, 2006 (DoA 2006). However, this data does not include the free-range informal cattle. For the total cattle figures the values from the UN Food and Agriculture Organisation were used (FAO 2006). The following livestock figures were used in the model.

Table 14: Historical data for livestock (1990 to 2005)

Source: Own compilation, based on Scholes et al (2000), FAO (2006)

Year Cattle total Dairy Goat Sheep Pigs

Units Million 1000 1000 1000 1000

1990 13.3 1100 2774 29979 1665

1991 13.5 1260 2453 28631 1654

1992 13.5 1090 2285 27448 1653

1993 13.1 1150 2159 25670 1570

1994 12.5 1050 2337 25851 1585

1995 12.6 1130 2369 25481 1707

1996 13 1140 2406 25566 1699

1997 13.4 1100 2394 25010 1736

1998 13.7 1070 2360 25079 1780

1999 13.8 1080 2325 24463 1647

2000 13.6 1370 2355 23586 1678

2001 13.5 1360 2427 22998 1710

2002 13.6 1210 2216 22614 1663

2003 13.5 1070 2160 22693 1663

2004 13.5 1020 2164 22289 1651

2005 13.8 1130 2138 22236 1656


The fluctuations between years are mainly dependant on rainfall and availability of grazing. As no data is available for free-range cattle it was assumed that 15% of the total cattle excluding dairy is in feedlot and the rest is free-range.

Data for poultry is only available for 1988 (51 787 000) and 2002 (185073 000) for commercial farmers (DoA 2006). However other data sources give much higher values. For example USDA FAS Poultry and Products Annual 2006 report for South Africa provides the value of 624 million birds for 2005 (http://www.thepoultrysite.com/articles/671/south-africa-poultry-and-products-annual-2006). The SA Poultry Association reported that in average for 2006 12.5 million birds were slaughtered per week, which is 650 million/a (http://www.sapoultry.co.za/download /broiler_stats.pdf). There were 15.8 million of layer flock (for egg production) in 1999 (www.nda.agric.za/ docs/MarketExtension/9BroilersEggs.pdf). It is increased to 20.5 million in 2006 (www.sapoultry. co.za/download/egg_stats.pdf) In addition there were more than 5 million breeder flock in 2006. It was assumed that the chicken life cycle is 60 days and the number of chicken in the model was corrected by applying the factor of 60days/365 days per year as suggested by IPCC guidelines (IPCC 2006). However the local data suggests that slaughter age for broiler chickens reduced from 45 to 38 days from 1992 to 2002 (Kleyn 2004). To improve the model accuracy the poultry farming need to be split into 3 groups: broiler, layer and breeder and different life cycle and manure management methods should be applied to each.

The enteric methane emissions of livestock are dependant on the type, age and weight of animal, the quality and quantity of food and the energy expenditure of the animal.

The quality of food is very critical and it is expressed as DE, digestibility of the feed in% (e.g. 60%). The assumptions for average mass and DE for different types of livestock are summarised below.

Table 15: Mass and digestive energy for different types of livestock

Source: Scholes et al (2000)

Type Mass (kg)

DE (%)

Free-range 400 50

Dairy (milk) 550 65

Feedlot 250 70

Sheep 30 56

Pigs 70 75

Goats 40 55

The pregnant, lactating and draft (oxen) animals have different energy requirements and for each type of livestock an assumption was made that 30% of herd belongs to these groups. For dairy cattle it was assumed that 87% of herd is pregnant or lactating and none of the feedlot. 3% of draft was assumed for free-range cattle. In order to calculate gross energy intake by livestock (GE, expressed in MJ/d) the following coefficients were used.


Table 16: Energy coefficients for different types of livestock


Energy coefficients (abbrev-iations used in equation)

Free-range

Dairy (milk) Feedlot Sheep Pigs Goats

Feeding energy % (Fenergy) 37 10 0 37 0 37

Weight gain (kg/d) (Wgain) 0.3 0.4 1.2 0.08 0.5 0.08

Milk/day (kg) 2 15 0 0.5 6 0.7

Milk fat % (MilkF) 3 3.5 4 6 3 6

Hours draft work/day(Wh/d) 4 0 0 0 0 0

Resulting GE differ from slightly from values listed in IPCC guidelines, but for this version it was decided to accept model results as more representative for South African conditions.

The gross energy intake (GE expressed in MJ/d) was used to calculate emissions of methane. The emission coefficients used are presented in the table below.

Table 17: Emission coefficients for different types of livestock


Type CH4 emission coefficient

Free-range 0.06

Dairy (milk) 0.06

Feedlot 0.04

Sheep 0.07

Pigs 0.04

Goats 0.07

These emission coefficients represent methane conversion factors (percent of gross energy in feed converted to methane) and were used to calculate the emission factor in kg/head/a.

EFCH4(i)= GEi*CH4prod*365/55.65

Where: 365 – conversion from days to year 55.65 (MJ/kg CH4) is the energy content of CH4

Finally the total emissions were summarised for all livestock types ‘i’ and converted into Gg of CH4 as follows

∑CH4 (Gg/a) = ∑ EFiCH4*Numi /106

Where: Numi is the number of livestock of the type ‘i’ It is divided by 106 to convert units into Gg/a.

Assumptions for baseline and mitigation option

The reduction of enteric emissions of CH4 could be achieved if the herd composition is optimized, the feed improved and cattle is moved from free-range grazing to feedlots.

As a mitigation option, the total number cattle was reduced starting in 2006 from 13.8 Mil heads/a to 9.7 Mil by 5% per year till it reached reduction of 30% by 2011. It was assumed that from 2006 the 5% of free-range herd is moved to feedlot each year till 45% of the cattle will be in feedlots. According to the Department of Agriculture (DoA) (J Classen, pers. communication) with the promotion of emerging farmers this change will be almost impossible to achieve. However, this


assumption was accepted in this version to allow keeping the beef production at the same level, although total number of cattle has eventually been reduced by 30% to achieve significant mitigation level.

A new assumption was added that the number of pigs and chicken will raise according to GDP growth till 2010 and then stabilize (poultry reaches above 250 000 heads/a by 2010).

It is assumed that ship and goat herd sizes are stabilized at 2005 levels. An additional assumption could be added that dairy cattle will grow at a rate of 0.6%/a (J Classen, pers. communication), but it should not exceed 1.5 million.

It is assumed that the feeding, even in free-range will improve and this will reduce the energy required for feeding from 37% to 30% and improve weight gain form 0.3 to 0.5kg.

Similarly, the feeding energy for sheep was reduced to 30% (from 37% - see table 3). It was also assumed that for mitigation option weight gain (kg/d) is increased to 0.1 from 0.08 for sheep.

The most important improvement for mitigation is better digestability. It was assumed that DE will from 50% for free-range cattle to 55% and for sheep to 60%.

The historical data for up to 2005 was replaced in the model and all the calculations extended till 2050. Although a large number of input values and assumptions have been changed, the total mitigation achieved is very similar to mitigation calculated by original model (see Figure in the section on results).

Calculation of costs for baseline and mitigation option and cost efficiency

The cost of production was based on three groups of expenditure: cost of food, veterinary services and fixed costs. The assumptions on the costs and productivity coefficients are summarised below. The new updated productivity rates were provided by the DoA (J Classen, pers. communication). It was assumed that new values are applicable for the period after 2005 in order to keep baseline consistent. It was further assumed that ‘Feed’ cost is proportional to increase in production for post 2005. The rest of the costs were calculated from values in the existing model by applying CPIX index correction.

Production cost in R/head was calculated as follows:

Production costi = (Feedi/AU+Veti/AU+Fixedi/AU)*Prodi/100

Where: AU – animal unit

The calculated cost is divided by 100 to converts from cents to Rands.

Table 18: Costs of production and productivity for different types of livestock (post 2005)

Source: Own compilation, based on Scholes et al (2000); Classen (2006)

Production costs(R per head) Type

Production

(kg/head/y)

Cost

Feed/AU Vet/AU Fixed/AU

Free-range 55 457.6 97 118.1 36.6

Dairy (milk) 2500 81.9 1603 277.8 166.3

Feedlot 150 624.2 720 133.1 83.2

Sheep 45 1020.1 31 95.1 332.7

Pigs 85 364.9 217 16.6 76.5

Poultry 27 398.0 90 0.8 16.6

National cost of production (expressed in R Mil)is calculated as follows:

Prod Cost = ∑Numi*Prodi*Costi/106


Where: Prod(i) is production rate for group ‘i’ expressed in kg/head/a It was divided by 106(to convert to R mil)

The national income was calculated in the same way.

The updated income rates (assumed that applicable after 2005 to keep baseline consistent) were provided by the DoA (J Claase, pers. communication) for some of the categories and for other increase using CPIX index was assumed. The values of products used by updated model presented in the table below.

Table 19: Costs of income (value) for different types of livestock

Source: Own compilation, based on Scholes et al (2000); Claasen (2006)

Type

Value (till 2005)

(c/kg)

Value (post 2005)

(c/kg)

Free-range 800 1331

Dairy (milk) 129 199.7

Feedlot 900 1497

Sheep 935 2012

Pigs 746 1020

Poultry 384 639

7. Methodology for modelling emissions from livestock manure management

7.1 Data, assumptions and calculations of baseline and mitigated emissions for manure management

This section describes how to estimate CH4 produced during the storage and treatment of manure. The emissions associated with the burning of dung for fuel are excluded The decomposition of manure under anaerobic conditions (i.e., in the absence of oxygen), during storage and treatment, produces CH4. These conditions occur most readily when large numbers of animals are managed in a confined area (e.g. dairy farms, beef feedlots, and swine and poultry farms), and where manure is disposed of in liquid-based systems (lagoons). The main factors affecting CH4 emissions are the amount of manure produced and the portion of the manure that decomposes anaerobically. The former depends on the rate of waste production per animal and the number of animals, and the latter on how the manure is managed.

The calculation starts with determining VS, the volatile solid excretion per day on a dry-organic matter basis (expressed in kg VS/day). It is calculated as follows:

VS= GE/18.45*(1-DE/100)*(1-ASH)

Where: GE = gross energy intake (MJ day-1) DE = digestibility of the feed (percent) ASH = the ash content of manure calculated as a fraction of the dry matter feed intake (e.g., 0.08 for cattle). 18.45 = conversion factor for dietary GE per kg of dry matter (MJ kg-1). This value is relatively constant across a wide range of forage and grain-based feeds commonly consumed by livestock.


Then Bo, the maximum methane-producing capacity of the manure is determined. The Bo coefficient, FBo varies by species and diet and expressed in (m

3/kg VS). These coefficient for different types of livestock are provided in the table below.

Table 20: Maximum methane production coefficients (Bo) for manure production by different types of livestock

Type FBo

Freerange 0.1

Dairy 0.13

Feedlot 0.12

Sheep 0.2

Goats 0.2

Swine 0.29

Poultry 0.32

Then the CH4 emissions expressed in (Gg/a) for every type of livestock (‘i’) are calculated by summarizing the emission from each type of manure management system (‘s’)

CH4(i) =VSi*365*Boi*0.67*∑(MSs*MCFs)*Num(i)

Where VSi - volatile solid excretion per day for livestock of type ‘i’ 365 – conversion from days to year Bo - the maximum methane-producing capacity of the manure, varies by species and diet (m3/kg VS). 0.67 = conversion factor of CH4 in m

3 to CH4 in kilograms MCFs = methane conversion factors for each manure management system’s’ MSs = fraction of livestock category, manure handled using manure management system ‘s’(dimensionless) Num(i) is a number of livestock of type ‘i’

The fractions of manure handled by different manure management systems are presented in the table below.

Table 21: MS coefficients for baseline (% manure handled by different types of management system)

MS Freerange Dairy Feedlot Sheep Goats Pigs Poultry

% lagoon 0 50 20 0 0 50 20

% digester 0 0 0 0 0 0 0

%spread 100 50 80 100 100 50 80

7.2 Assumptions and calculations for mitigation for manure management

The mitigation scenario assumes the same growth in the beef, pork and poultry feedlots as in the business-as-usual scenario, but that 40% of the beef, pork and poultry feedlot wastes are anaerobically digested or consumed in a biomass converter. One tenth is treated in open lagoons, and the remainder is dry spread. The differences between baseline and mitigation option are highlighted in red the table below.


Table 22: MS coefficients for mitigation option (% manure handled by different types of management system)

MS Freerange Dairy Feedlot Sheep Goats Pigs Poultry

% lagoon 0 10 10 0 0 10 10

% digester 0 40 40 0 0 50 40

%spread 100 50 50 100 100 40 50

The cost of disposal are calculated by summarizing the costs for each type of manure management system (‘s’)

Cost = ∑Num(i)*VSi*365*∑(MSs*Costs)

Where Num(i) is a number of livestock of type ‘i’ VSi - volatile solid excretion per day for livestock of type ‘i’ 365 – conversion from days to year MSs = fraction of livestock category, manure handled using manure management system ‘s’(dimensionless) Costs = cost of disposal for manure management system ‘s’(R)

8. Methodology for modelling emissions from reduced tillage

8.1 Historical data, assumptions and calculations for tillage

The model for the agricultural sector has been developed and used for the SA Country Study on Climate Change (Scholes et al. 2000) has been used as a basis for this study.

8.1.1 Area under cultivation The area under cultivation was updated using the latest data from the Abstract of Agricultural statistics, 2006 for the period 1970 to 2000 and the latest data (up to 2006) from the Crops Estimates Committee (www.sagis.org.za/Flatpages/Oesskattingdekbrief.htm). The Abstract of Agricultural statistics includes both commercial and developing agriculture, while the Crops Estimates Committee provides data for commercial agriculture and for developing agriculture separately. For the last two years only data for commercial agriculture was provided. An assumption has been made that the areas under developing agriculture amount to 15% of those under commercial agriculture. The area under maize was significantly reduced in 2005/2006 season because of the drop in the price of maize. However according to Crops Estimates Committee, a 73% increase is expected for 2006/7 year. Thereafter the original assumption of 4000 000 ha under maize, was used. The same data sources were used for grain. From 2008 the original assumption of 1300 000 ha under wheat was used.For the year 2007, an average between the areas for 2006 and 2008 was assumed. Dryland grain production is the only form of grain production being considered. Irrigated grain production has been ignored in this model, because carbon storage in irrigated lands differs from that of non irrigated lands. The areas used in the model are presented in the figure below.


0

1000

2000

3000

4000

5000

6000

1980

1985

1990

1995

2000

2005

2010

2015

2020

2025

2030

2035

2040

2045

2050

Years

Area (1000ha)

Maize

Wheat

Figure 3: Area for production of maize and wheat (1000ha)

8.1.2 Carbon storage According to Van der Merwe, M. R. & Scholes, R. J. 1998

The carbon content of South African soils is on average low, for three main reasons:

• the majority of soils are sandy, and therefore stabilise little carbon;

• the temperatures are high, leading to high soil organic matter decomposition rates; and

• the climate is dry (600-800 mm rainfall is the norm in dryland crop areas).

The typical range of virgin soil carbon content within the plough layer (0-300 mm) is 0.3-3.6%, with a modal value around 1.2%. The bulk density of agricultural soils has been assumed to be 1.3 Mg m-3, giving a pre-cultivation carbon density(C0) of 4680g C/m

2 to a depth of 300 mm. The equation describing the change in carbon content is as follows:

Ct = C0 t -0.21

where Ct is the carbon density in year t after commencing cultivation, and C0 is the pre-cultivation carbon density.

In the model, calculations are based on the assumption that, in cultivated lands, carbon storage is reduced to 50% of Corig as a result of tilling. It also assumes that recovery of stored carbon resulting from introducing the no tillage system, is not complete, but reaches 80% of the pre-cultivation level (see table below).


Table 23: Coefficients for calculation of C storage in soil

Description Value

Mean original (pre-cultivation) soil C (Corig) 46.8 MgC/ha

Soil C reduction by till (% reduction) 50 %

Exponent of recovery- α 0.21

Soil C after recovery ( % final) 80 %

If ‘t’ is the number of years after introducing reduced tillage, then Carbon stored in year ‘t’ is calculated as follows:

Cstoredt=Corig*(%reduction/100)+ Corig*((%final-%reduction)/100)*(1-(t+1)-α)

The change in carbon storage during this year is calculated as follows:

∆C = Cstoredt+1- Cstoredt

The total change in carbon stored in all lands under wheat and maize as a result of the introduction of reduced tillage, is as follows:

Total ∆C = Area * tillage adoption*∆C

Finally the total C stored:

Cstored = ∑∆Cstored- area*Corig* %reduction*(yeart – 1970)α- (yeart-1 – 1970)

α

The calculation was updated by extending storage changes to 30 years instead of 10 used by the previous model and by adding calculations up to the year 2050.

9. Capital and variable costs requirements to start a no-till system

Before the cost of a system can be calculated, the assumptions must first be noted.

The following assumptions were made in the calculation of the staring cost of a No-till system

• The farmer must be an above average manager.

• The basic No-till planter must be obtained.

• A good sprayer with sufficient capacity must be obtained to apply the herbicide correctly.

• A planter and sprayer can handle 500 ha per year.

• A herbicide application program must be in place.

• In the first year a cover crop must be planted to supply the stubble for the No-till system.

• Sorghum is used as the cover crop.

• The Roundup ready system will be used as the basis for maize production.

• A maize price of R1000 per ton is used.

• In the first year, the maize yield will be 80% of the conventional system crop yield.

• In the second year, the maize yield will be 95% of the conventional system crop yield.

• In the second year, the maize yield will be equal to the conventional system crop yield.

• Budgets for each year and crop must be compiled.

• The effect of inflation will not be included.

• The fixed cost per ha will not be included.


• The starting cost of the No-till system will be the

• capital lay out of a planter and sprayer;

• direct cost of the cover crop;

• loss of income between the conventional roundup ready system and the No-till year one crop systems.

Table 24 gives an estimate of the capital required to buy the planter and sprayer.

Table 24: Capital layout for planter and sprayer

No-till planter Metasa No-till planter R300 000

Sprayer Tecnoma Galaxy sprayer R150 000

Total R450 000

According to the assumptions, a planter and sprayer can handle 500ha of maize per year. Taking this into account, the capital cost for the planter and sprayer will be R900 per ha. In year one the farmer will not be able to sell his old equipment as he must still plant with the old equipment. If all the lands are switched over to No-till, there will be some equipment that can be sold. Normally this old equipment doesn’t have a market value and will be sold for next to nothing.

In the table below, the direct cost and gross margins for a 3.5 ton maize yield conventional roundup ready and No-till systems, are given. The effect of the lower production was taken into account and the yields were lower.

Table 25: Production cost for different maize systems

Production year 2006/2007 Product price (R000/ton)

System Roundup ready system

No-till system

No-till system

No-till system

Year Year 1 Year 1 Year 2 Year 3

Yield (to/ha) 3.50 2.80 3.33 3.50

Gross production value (R/ha) 3 500.00 2 800.00 3 325.00 3 500.00

A: Direct allocated variable cost (R/ha)

Seed 412.40 412.40 412.40 412.40

Fertiliser 555.41 555.41 555.41 555.41

Lime 51.00 51.00 51.00 51.00

Fuel 501.99 250.41 250.41 250.41

Repairs 395.48 281.84 286.14 287.58

Lubricant 25.10 12.52 12.52 12.52

Herbicides 124.92 242.90 242.90 242.90

Pesticides 147.30 147.30 147.30 147.30

A: Total Direct allocated variable cost (R/ha) 2 213.60 1 953.77 1 958.08 1 959.51


Production year 2006/2007 Product price (R000/ton)

System Roundup ready system

No-till system

No-till system

No-till system

B: Other allocated variable cost (R/ha)

Crop insurance (R/ha) 133.00 106.40 126.35 133.00

Part time labour 45.40 45.40 45.40 45.40

Production interest (R/ha) 132.82 117.23 117.48 117.57

B: Total other direct allocated variable cost (R/ha) 311.22 269.03 289.23 295.97

C: Total allocated variable cost (A+B) (R/ha) 2 524.81 2 222.80 2 247.31 2 255.49

Gross margin (R/ha) 975.19 577.20 1 077.69 1 244.51

According to Table 25 the gross margin of no-till in year one is lower than the conventional roundup ready system. In year 2 and 3 the gross margins of the no-till systems are higher than the roundup ready system. The difference between the roundup ready system and the No-till year 1 system is R397.99. This will be the opportunity cost for changing from the roundup ready system to the No-till system.

In Table 26 the direct allocated cost for the production of sorghum is shown. In the calculation the cost of a 5 ton per ha sorghum was used to produce enough stubble for the No-till system. The cost to produce 5 ton of stubble will be R 1099 per ha.

Table 26: The direct allocated cost of sorghum for cover crop

Production year 2006/2007

Yield (ton/ha) 5.00

A: Direct allocated variable cost (R/ha) 67.20

Seed 239.67

Fertiliser 397.79

Fuel 277.22

Repairs 117.38

Herbicides 1 099.26

A: Total direct allocated variable cost (R/ha)

Table 27 summarises the cost to start one hectare no-till maize. The total cost to change from a roundup ready system to a no-till system will be R2 397.25 per ha. In effect this means that a farmer must arrange for R2397.25 extra production credit to start this action. The payback time will differ from farmer to farmer and is not included in this calculation. If the assumption is made that the farmer can cover 50 % of the capital layout with the selling of old equipment it will take the farmer nine years to break even. This project must be done with good financial planning.

Table 27: The cost of start one hectare no-till maize

Capital layout for planter and sprayer per ha (R) R900.00

Direct allocated cost of sorghum as cover crop R1099,25

Opportunity cost for year 1 in no-till R379.99

Total capital, direct and opportunity cost for no-till system R2 397.25


10. Methodology for modelling mitigation from land use changes (fire control and savannah thickening)

10.1 Fire control

Fires in grasslands, savannas, fynbos and plantation forestry in South Africa are modelled. Six land cover types were considered by the model: Fertile savanna; infertile savanna; sweet grassland; sour grassland; fynbos; plantation. Although a large quantity of CO2 is generated as result of fires, it is not a net emission (assuming that it is re-absorbed in plants in the next growing season) and only CH4 and N2O emissions were calculated. The emissions for each land cover (lc) are calculated as follows:

Elc =∑area(km2)/fire return frequency *fuel load (kg/ha)*combustion completeness*emission

factor(g/kg)

The parameters used for this calculation are provided in the table below.

Table 28: Data used for calculation of emission from fire

Land cover

Historical fire return frequency

(Yr)

Present fire return frequency

(Yr)

Area

(ha)

Fuel load

(kg/ha)

Complete frac (combustion completeness)

Fertile savanna 6 10 28 285000 1000 0.9

Infertile savanna 3 5 1 2122100 2500 0.95

Sweet grassland 4 4 14 411340 1100 0.9

Sour grassland 2 3 9 607560 3000 0.95

Fynbos 20 15 46046 20000 0.7

Plantation 100 200 1 241300 30000 0.4

The N2O emissions are calculated using the same equation with different emission factors.

Some frequency of fires is necessary in these vegetation types (other than plantations) in order to maintain their ecological health. Furthermore, the fires are to a degree inevitable, given the seasonally-dry climate in South Africa (see Figure 4). Nonetheless, the return frequency of fires can be reduced significantly below their current frequency without causing ecological damage, while at the same time realizing savings in loss of life, livestock, grazing and infrastructure. The costs of complete fire prevention are unaffordable and it is an unrealistic and unnecessary, but fire frequency reduction is an attainable target. For this model mitigation by 50% reduction is assumed.


(A)

y = 1.3026x - 400.09

R2 = 0.716

000

100

200

300

400

500

600

700

800

900

1000

200 300 400 500 600 700 800 900 1000 1100

Mean rainfall in preceding 2 years (mm)

Area burnt (ha x 1000)

Figure 4: The relationship between mean annual rainfall over the preceding years and the extent of fires in the Kruger National Park

Source: van Wilgen et al (2004)

Unit Cost/ha (UC) of achieving 50% fire reduction is calculated by summarising different components of costs for all land cover (lc) types as follows:

UClc =∑(UC detection*AC detection + UCequip*Cequip+ Num people*AC people+ Num people*cost of kit

Most of the unit costs are expressed as cost per 1000ha

Unit Cost/ha of damaged (UD) caused by fire is also calculated by summarising different components of costs for all land cover (lc) types as follows:

UD = Dvegetation* fuel load*probability of Dv + D livestock* probability of Dl+ D infrastructure *probability of Di

And finally the total control and damage costs are calculated as follows:

Control costs =∑area*UC* fire return frequency/(0.5*historical fire return frequency)

Damage cost =∑area*UD/fire return frequency

10.2 Savannah thickening

It has been widely observed that the woody biomass in savannas (‘bushveld’) has increased over the historical period. This phenomenon has been noted in Africa, Australia and America. The main reason is a reduction in fire frequency and intensity. Frequent, intense fires formerly restricted the recruitment of woody plants. With the introduction of domestic livestock in large numbers, an increasing fraction of the grass production is grazed rather than burned, allowing the trees to become established. Once the trees mature, they further suppress grass growth, leading to the downward spiral known as ‘bush encroachment’.


This process has negative economic consequences for graziers, but positive consequences for carbon sequestration, since densely wooded savannas store more carbon, both as trees and in the soil, than open savannas.

Increase in basal/woody area is considered for two land cover types – fertile and infertile grasslands. It is assumed that original woody area in this grasslands are 6 and 8 m2/ha (Areao )

Firstly the Max woody area is calculated as follows:

MaxWAlc= Max Ba* fire return frequency /(RT50+ fire return frequency )

Where: Max Ba is maximum tree basal area which is a function of rainfall RT50 is 50%ile of the fire return frequency

Increase in woody area is calculated as follows:

∆WA (lc) = R* WAreao *((MaxWA- WAreao)/MaxWA))

Where: R is coefficient of savanna growth (assumed 0.04)

Then woody area in year ‘t’ is calculated as follows

WAreat = WAt-1+ ∆WA

Finally the increase in CO2 sequestration (in GgCO2/y) is calculated as follows:

∆CO2=∑(∆WAlc* Conv factor*Arealc*0.4*44/12/1000

Where: Conv factor is conversion for sequestration of C (5.2 Mg/ m2 ) 44/12 is conversion from C to CO2 0.4 represents assumption that only 40% of savanna area would exhibit thickening

The loss of grazing was calculated from an equation derived in Australia, relating grass biomass to tree basal area. It summarises losses for both types of grassland(lc)

Loss grazing (Rmillion/a) =∑ (Normal CC -(Grass C+( Grass C +LSU/ha)^(Grass K* WAreat))*Income/ha *Area (lc) )/1000000

The Normal CC is calculated as follows:

Normal CC = Grass C+(LSU/ha + Grass C )^( Grass K *Normal BA)

Where: Grass C is the amount of grass production (expressed in animal carrying capacity units) with maximum tree cover Grass K is the amount of grass production (expressed in animal carrying capacity units) without any tree cover Normal BA is tree basal area LSU/ha is livestock units per ha

It must be noted that LSU/ha should be higher for fertile than infertile savanna given the same rainfall, but averaged over SA the infertile savannas have higher rainfall. Therefore it is assumed 0.1 for fertile and 0.15 for infertile savanna


Table 29: Parameters used for calculation of loss of grazing

Land cover Grass k Grass C Normal BA

Fertile savanna 0.25 0.01 6

Infertile savanna 0.25 0.01 8

11. Data for modelling mitigation from waste sector The most comprehensive national study describing the waste sector in South Africa was the baseline study in 2001 in preparation for the National Waste Management Strategy (DWAF 2001). It classified and quantified the waste generated and disposed of in South Africa. The results of the baseline data collection are presented in the table below.

Table 30: National information on general waste generation in 1997

Source: DWAF (2001: Table 9)

Information from receipts at landfills

Information on generation from questionnaires

Province Waste disposed Waste collected

Eastern Cape 571 000 441 000

Free State 782 000 482 000

Gauteng 4 297 000 1 963 000

KwaZulu-Natal 1 811 000 410 000

Mpumalanga 481 000 353 000

Northern Province 153 000 199 372

NorthWest 354 000 290 000

Northern Cape 262 000 147 000

Western Cape 1 487 070 423 000

Total 10 245 070 4 699 503

The discrepancy between waste collected and waste disposed shows low accuracy of the available information. While government has implemented a national waste information system to collect regular data on waste disposal to landfill, it will be some time before accurate national data are available.

The estimation of waste received at landfills is inaccurate. Many landfills do not have weighbridges and they are basing their estimations on guesses or on density estimations, which may an order of magnitude out. Many of the landfill sites base their estimates on volumetric measurement on the vehicles coming into the site both in a loose (open trucks) and compacted form (rear end loaders), hence difficult to tie up with estimated densities. There are also periodical insitu topographical surveys of the landfill, but this form of waste estimation also has problems with respect to densities and sometimes the incorrect method of volume calculation due to ongoing settlement in the landfill (S Jewaskiewitz, pers communication 2007).

CH4 from landfills is produced in combination with other landfill gases (LFGs) through the natural process of bacterial decomposition of organic waste under anaerobic conditions. The LFG is generated over a period of several decades and it can start 6 to 9 months after the waste is put in place. CH4 makes up 40-50% of LFG. The remaining component is CO2 mixed with trace amounts of volatile fatty acids (VFA), hydrogen sulphide (H2S), mercaptans (R-SH) and ammonia/amines (R-NH2). The mercaptan and amine compounds have particularly strong and offensive odours even at low concentrations.


Typical landfill gas, if permitted to accumulate in low lying, enclosed or confined spaces, may produce an atmosphere that is both explosive and hazardous to life. The CO2 and VFA components of landfill gas are highly aggressive to concrete, brick mortar and mild steel. Landfill gas will displace oxygen from enclosed spaces making entry to them extremely hazardous. CH4 is explosive in air concentrations of 5 – 15% by volume.

Landfills are engineered sites designed and operated to employ waste management practices, such as mechanical waste compacting and the use of liners, daily cover, and a final capping. As the landfill uses a soil cover (biocover) in its operations, a portion of the CH4 is oxidized as it passes through these soil layers and converted to CO2.

A lot of international research is currently underway looking into biocovers for landfills. This is particularly important with respect to landfills that do not have landfill gas extraction and management systems. There are many such landfills in South Africa including many of the so-called open or controlled dumps or even the smaller permitted landfills where the production of landfill gas is deemed to be too low for the consideration of gas extraction systems. Many of the open dumps are now either being closed or are being encouraged to register with DWAF with the view to being permitted (licensed). In these cases these landfills will become producers of landfill gas. Biocovers are used to oxidize the methane on its way through the capping of the landfill. This can therefore also be considered as mitigation measure.

The existing landfills are running out of available space, therefore new landfills have to be identified and approved through a lengthy, rigorous and frequently contested EIA processes. The provincial governments are reluctant to approve the creation of new landfills in built-up urban areas, which means that landfills will have to be constructed in peri-urban or rural areas. The consequence thereof is that waste has to be transported over long distances, resulting in high energy and resource inputs with associated high costs and increase in emission from transport.

Waste minimization and recycling is an upstream intervention. According to DST (2006):

In an attempt to reduce an amount of waste going into a landfill, South African Government and related stakeholders have pledged to grow the recycling industry by 30% in 2012. In 2003, 52% waste paper was recycled. South Africa recycles about 20% of glass containers produced per annum and it is estimated that 30% of plastic used for packaging is recyclable. DEAT estimates 85% of beverage cans are recovered and recycled annually in South Africa. It is also approximated that 2% of electronic waste is being recycled in South Africa.

The proposed Waste Management Bill (DEAT, 2006) has placed the waste hierarchy within a life-cycle assessment approach as it states: ‘Every person who undertakes a recovery, re-use or recycling activity must, before undertaking that activity, ensure that the recovery, re-use or recycling of the waste uses less natural resources than disposal’. This implies that only viable, sustainable options to recovery, re-use or recycling of the waste will be considered in future, which are e.g. less resource intensive that e.g. landfilling. The life-cycle assessment is also the approach used in the UK and Europe.

In South Africa only large cities have engineered landfills, with smaller cities, towns and villages having controlled or open dumps. The large sites (with input greater than 30 000 t/a) were studied to determine potential for power generation (DME 2004). It was found that 20 large sites yielded 41 MWe of energy, which is close to 70% of the potential energy for all of the landfills studied. According to the DME study, a total of 57 sites could be considered for power generation out of a recorded 453 sites. These 57 sites were estimated to produce 502 Mm3 of LFG in 2005.

It can be assumed that for the smaller sites and controlled dumps, which generate remaining 30% of the landfill gas, these emissions can be mitigated by flaring (destroying) the gas, using the gas in thermal applications or oxidising the methane through the use of biocovers.

A comprehensive pre-feasibility study was recently completed by DST (DST 2006) on energy recovery from municipal waste, which suggested some innovative approaches to energy from waste projects. However, energy recovery from LFG is not an optimal solution. There is a need to put mechanisms in place to divert organic waste from landfills (e.g. into composting) as a long-term


solution, with energy recovery from landfills a short-term solution, to deal with the current LFG generation.

The problem with the diversion of organic wastes away from landfill is that broadly speaking the organic wastes can be divided into two classes: that which can be easily separated (garden wastes etc) and that which is mixed into other wastes. The garden wastes are generally sent to composting operations whereas the mixed wastes can be treated biologically – Mechanical Biological Treatment (MBT), prior to being landfilled. In this way the generation of landfill gas is avoided.

Utilisation of LFG not only produces energy and prolongs the life span of the waste sites, but also allows carbon credits under the CDM to be claimed.

12. Economic impact analysis

12.1 Introduction

Climate change mitigation actions have various important implications for the economy. This section reports on results from a series of model simulations examining the economic impacts of mitigation scenarios. The scenarios were developed by the Scenario Building Team (SBT) of the Long-Term Mitigation Scenarios (LTMS) project. The economic analysis follows directly from the energy modelling part of the study, i.e. the implications of various mitigation scenarios for the South African energy system were explored using the MARKAL energy model. Results from this model was then used to inform various policy shocks in an economy-wide model. Thus, the economic model is linked to the energy modelling in a ‘top-down’ fashion, using key outcome variables from the energy model to define ‘shocks’ for the economic model.

Given the complexity of scenarios it is necessary to employ a comprehensive economic framework that models interactions between a variety of economic agents, including productive sectors, factors of production (capital and labour), households, incorporated business enterprises, government and the rest of the world. Econometric models, while generally more suitable for making longer term predictions given their focus on trends in economic variables, are usually unable to deal with highly detailed interaction effects. Since policy makers are keen to understand impacts on economic institutions and agents that are too disaggregated for an econometric model, this study opts for a computable general equilibrium (CGE) model. CGE models are economy-wide models that take into account aspects of microeconomic behaviour of producers and households, while maintaining macroeconomic constraints fundamental to economic accounting systems. These models incorporate representations of all markets, including commodity markets, factor (labour) markets and international trade.

CGE models are, however, not very suitable for predictions over long periods of time, given model complexity, and various restrictive assumptions regarding rigidities in the structure of the economy and relationships between economic agents. Instead, these models are useful for showing how one state of the economy may differ from another state in terms of numerous economic variables. Sadoulet and De Janvry (1995:287) write about input-output models that they are “more useful as

guidelines to potential linkage effects … [rather] than as predictive models”. This also applies to CGE models, which fall in the same class as input-output models. In this study we adopt an approach whereby we report on changes in various key economic variables – in particular, gross domestic product (GDP), employment and household welfare levels – relative to a so-called ‘business as usual’ reference case. Following the MARKAL model time horizon, results are generated for selected years over the period 2005 to 2050. However, we only report the results for 2005, 2010 and 2015 given that CGE models are fairly restrictive as far as long term prediction is concerned. More detail about the modelling approach is supplied in section 12.2.3

Most mitigation strategies require large investments in the economy that may have a number of spin-off effects. Ignoring financing requirements for the moment, the immediate effect of an increase in investment is observed as an increase in demand for investment-type goods such as machinery and equipment, or investment-related services such as building and construction works. Those sectors


supplying these investment-type goods and services are the immediate beneficiaries of increased investment demand. These sectors react by increasing production in order to satisfy this additional demand, subject of course to production constraints in the economy and also the share of newly demanded investment goods that are imported. The increase in demand for investment goods is only observable during the ‘current’ economic accounting period or the ‘construction phase’. For this reason the first-round or direct effect of an initial investment demand shock is regarded by economists as having only a ‘temporary’ demand-driven impact on production and employment levels, and mainly in those sectors that supply the investment-type goods and services. Given inter-industry linkages in the economy some downstream effects can also be observed, for example in those sectors that supply intermediate inputs to the primary beneficiary sectors.

The long term effects associated with investments only become observable once investments have been converted to installed production capacity or improved production processes. This effect is very different in nature from the direct short-run effects associated with the initial investment in that the structure of production is now altered, either due to changes in production relationships (e.g. production technology changes or efficiency enhancements) or due to changes in the levels of capital stock employed in productive sectors (production capacity changes). Therefore, while the short-run investment effects are simply modelled as changes in current-period investment demand, the long term effects, require slightly more complex modelling, i.e., production relationships in the model have to be altered. Experience has shown that shifts in current investment levels generally only have small compositional effects in the economy. The real interesting effects are caused by the permanent long-term structural effects (see for example Van Seventer and Davies, 2006).

While we do account for the short-run effects of changes in investment levels in the simulation setup, we are more interested in two types of long-term effects that ultimately overshadow the investment results. The first is increased energy efficiency in the mining, manufacturing, commercial and transport sectors. Energy efficiency can be understood as a special type of production efficiency, and is modelled as a reduction in the use of ‘energy inputs’, including coal, petroleum (liquid fuels), gas or electricity, per unit of output. In some of the scenarios considered, fuel switching also takes place.2 For example, a shift towards electrified transport may reduce petrol or diesel use in the transport sector, but electricity use per unit of output increases. In a fuel switching scenario any efficiency gains in terms of the use of one type of energy input will be offset by increased usage of another, which in a modelling sense would appear like increased inefficiency. Ultimately, overall efficiency changes in a sector will depend on whether production costs under energy efficiency and fuel switching scenarios increase or decrease (see section 12.2.3.1).3

A second long-term effect relates to changes in production capacity. In the context of energy research we are concerned here with investments that lead to increased production capacity in production processes that are more environmentally friendly. For example, under a strategy whereby electricity supply from renewable sources or nuclear power is increased, the production capacity, measured in terms of capital stock employed in these two electricity sub-sectors (plant, machinery and equipment) are likely to increase. Increased production capacity in nuclear and renewables implies not only an increase in output from these sectors, but also a relative reduction in output from coal power stations, i.e. the output share of coal-fired electricity plants declines. In a comparative static modelling context, such as our CGE model, we model relative production capacity shifts rather than absolute changes in production capacity. In this study we refer to this as structural shifts in the energy output mix.4 The reason for this approach is explained further in section 12.2.3.2.

2 When referring to ‘energy efficiency’ scenarios or ‘wedges’, we imply that this may include fuel switching as

we ll. A more apt description would perhaps be ‘changes in intermediate input requirements per unit of output’. 3 Section 12.4.1 summarises findings from earlier CGE analyses of the energy efficiency ‘wedges’, and

specifically considers effects of industrial and commercial energy efficiency in terms of coal and electricity use in mining, manufacturing and commercial sectors.

4 Section 12.4.2 reports on selected findings from earlier analyses (using a SAM multiplier model) of various structural change ‘wedges’, in particular renewables and nuclear intensive scenarios for the electricity sector, and a biofuels scenario for the petroleum sector.


Three mitigation scenarios are modelled: ‘Start Now’ (initial wedges), ‘Scale Up’ (extended wedges) and ‘Use the Market’ (economic instruments with increased energy efficiency).5 These are the same scenarios that are simulated in the MARKAL energy model. Selected MARKAL model results are used to inform shocks for the economic model. These results include changes in the energy supply mix (structural shift scenarios), changes in energy efficiency or fuel switching, and capital outlay or investment requirements of alternative scenarios. Both the Start Now and Scale Up

scenarios are made up of combinations of energy efficiency gains and structural shifts energy supply sectors, while the Use the Market scenario adds the use of economic instruments, in particular a tax on CO2 emission.

By assumption, energy efficiency gains are achieved in virtually all the economic sectors in the economy, grouped here into mining, manufacturing, commercial and transport sectors. As far as structural shifts in energy output mix is concerned, we focus on two sectors, namely the electricity and petroleum sectors. In the case of the electricity sector, we consider structural shifts between electricity supplied in coal-fired plants, nuclear plants, electricity from renewable sources and from gas turbines. For the petroleum sector we allow for structural shifts in the output mix of crude oil refineries, coal-to-liquids (CTL) plants, gas-to-liquids (GTL) plants and from biofeuls. In each instance we compare simulated results for various selected years against a ‘business as usual’ reference case in a comparative static fashion.

The results from the Start Now and Scale Up scenarios are reported separately from the Use the

Market results. This is done for a number of reasons. Firstly, the Use the Market scenario considers rather extreme structural shifts in output mix. Hence results from this scenario are at a different scale in terms of percentage changes, and therefore not directly comparable with the moderate switching in the other scenarios, which in itself offers policy makers some insight into possible adjustment processes. Secondly, as far as the energy efficiency components are concerned, the scenario is also rather different in that it assumes the future availability – most likely via imports – of natural gas as a viable alternative to coal. It therefore includes substantial fuel switching. Thirdly, the use of a tax instrument makes this scenario very different from the others, and hence their economic impacts are expected to be dissimilar.6

The reference case against which outcomes under the mitigation scenarios are compared, also assumes some structural shifts in output supply over time that are required in order to meet future energy demand. These shifts obviously involve much less ‘decarbonisation’ than the mitigation scenarios. We therefore adjust the reference case to reflect these structural shifts in the energy sector. MARKAL model results are further used to determine how the mitigation scenarios differ from the reference case in terms of investment requirements (capital outlay), energy efficiency gains or fuel switching, and, in the case of the Use the Market scenario, tax instruments.

This report is structured as follows. Section 12.2 introduces the modelling approach, giving a more detailed description of CGE models and the type of results that they generate, as well as the mitigation scenarios and how these are simulated. Section 12.3 analyses the CGE model simulation results of the Start Now, Scale Up and Use the Market scenarios. Section 12.4 is a summary of results from simulation runs done for previous SBT meetings, specifically various ‘individual wedges’ (see footnotes 3 and 4). Additional figures and tables are included in section 12.5, which serves as an appendix to this part of the study.

12.2 Modelling Approach and Scenario Description

12.2.1 Objectives

The objective of this analysis is to develop a better understanding of the likely impact that various mitigation options may have on the economy in terms of GDP, employment and household welfare.

5 These scenarios were also previously called the Start Now, Scale Up and Use the Market scenarios. 6 As part of the background analyses we also looked at the pure economic effects of a CO2 tax in the absence of

structural shifts and energy efficiency. Results from this analysis is reported on in section 12.4.3.


As noted in the introduction, outcomes of three mitigation scenarios, Start Now, Scale Up and Use

the Market, are evaluated. These mitigation scenarios are combinations of different degrees of energy efficiency that can be achieved, structural shifts in energy output mix and, in the case of the latter scenario, economic instruments used to reduce emissions. Results are compared in comparative static fashion against a ‘business as usual’ reference case called growth without constraints (GWC). This remains a scenario analysis, and by no means can we claim that results are necessarily an accurate reflection of the true outcome. Given the long time horizon and the multitude of economic variables and parameters that may change over time and impact on each other, not to mention factors external to the South Africa economy that cannot be controlled, it is unwise to have too much confidence in results. However, the exercise remains useful. We are upfront about the limitations, the assumptions and the methods used to arrive at results, and given these, the scenario analysis provides a useful starting point for policy discussions around possible outcomes under various different mitigation scenarios for South Africa.

12.2.2 Model and Data

The study makes use of the Standard General Equilibrium (STAGE) model developed by McDonald (2006). This model is calibrated with a Social Accounting Matrix (SAM) for South Africa with base-year 2000, which was compiled by the Western Cape Department of Agriculture (PROVIDE, 2006). This section contains detailed descriptions of the CGE model and SAM data, with a specific focus on adjustments made to the SAM in order for mitigation actions to be analysed more accurately.

12.2.2.1 CGE Modelling Overview

The STAGE model is a member of the class of single country CGE models that are descendants of the approach to CGE modelling described by Dervis et al. (1982). The model adopts the SAM approach to modelling (see Pyatt, 1998). CGE models combine the productive sectors or activities with commodity and factor markets, and also draw linkages between these markets, domestic institutions (households, government and incorporated business enterprises) and the rest of the world. Essentially, CGE models are an extension to simpler IO or SAM-multiplier models. The main differences are the introduction of flexible prices and a variety of substitution mechanisms that allow for a more realistic or accurate representation of economic behaviour in response to relative price changes as opposed to the strict ‘linearities’ and fixed prices found in multiplier models.

What further makes CGE models unique is that they are macroeconomic or economy-wide models that are based on neoclassical microeconomic foundations. Agents optimise behaviour subject to constraints; for example, households (or consumers) maximise utility subject to prices and a budget constraints, while producers (or activities) maximise profits subject to a production technology constraint. Equilibrium is reached when supply equals demand in all the commodity and factor markets simultaneously, subject to various macroeconomic constraints: aggregate demand equals aggregate supply, total investment equals total savings, government and household budgets balance (revenue or income equals expenditure plus savings or deficit), and the foreign account is also balanced (balance of payments).

CGE models are set up with a range of flexible macro adjustment or closure rules. These define the way in which various of the macro-equilibriums are reached, based on beliefs or assumptions about how the economic system operates. The closure rules for this particular study are discussed in section 12.2.3.5, while a more detailed description of CGE models and their limitations is available in the technical appendix of the SBT 5 report.

12.2.2.2 A South African Social Accounting Matrix and Activity Account Disaggregations

When economic agents are involved in transactions with each other, financial resources exchange hands. The objective of a SAM is to capture all these financial resource flows in the economy that take place in a certain period (usually a year or representative year). As such a SAM provides a snapshot picture of the economic and social structure of an economy over that period. It also provides the statistical basis for economic models (King, 1985). A SAM contains information about productive activities in the economy, along with information from non-productive institutions such as factor markets, capital markets, households, government and the rest of the world.


A detailed explanation of a SAM structure, the underlying accounting principles and details of the accounts used in this study is included in the technical appendix of the SBT 5 report, while Table 39 in section 12.5 of this report includes a full listing of the commodity, activity, factor and household accounts. For the purpose of understanding the simulation exercises in this study a proper understanding of how information on the productive sectors, represented in a SAM by so-called activity accounts, are captured. Any firm produces goods by combining intermediate inputs (various commodities that form inputs into the production process) and factors of production. Factors of production may include land, capital and labour, and the combined contribution (in terms of production values) of these factors are called value added. An example of how the production structure is modelled in a CGE model appears in Figure 6 in section 12.2.3.1. The SAM captures information on each activity’s use of intermediate inputs and employment of factors of production.

One of the important mitigation options explored in this study is shifting the energy output mix from carbon-based processes towards more environmentally friendly processes. Large shifts are particularly expected to take place within the electricity and petroleum sectors. In the original SAM these two sectors are not disaggregated any further into different types of production processes. However, in order to effectively explore structural shifts in the energy output mix, it is necessary to specify sub-processes. Hence, four different types of liquid fuels production activities and four types of electricity generation processes are created, which are now essentially sub-activities of the original petroleum and electricity activities respectively:

• Petroleum sector: (1) crude oil refineries, (2) gas-to-liquids (GTL), (3) coal-to-liquids (CTL) and (4) biofuels.

• Electricity sector: (1) coal-fired power plants, (2) nuclear power, (3) renewable energy sources and (3) gas turbines.

In order to disaggregate the petroleum and electricity activity accounts it is necessary to understand precisely how the production processes differ in terms of their use of intermediate inputs, the returns to capital, the labour intensity of production and the skills composition of employment. The intermediate input use and employment data in the PROVIDE SAM is based on Statistics South Africa’s Supply and Use Tables for 2000 (SSA, 2003) and the Labour Force Survey for September 2000 (SSA, 2002). In both these datasets the electricity and petroleum industries are separate sectors. No further disaggregation is available in any of the national statistical databases. Hence, information had to be sourced from industry associations and firms that operate within these industries.7 Unfortunately, firms were generally unable or reluctant to release data and a range of assumptions had to be applied where data was unavailable.

The main corporations in the South African petroleum industry are members of the South African Petroleum Industry Association (SAPIA). Data from this association was used to obtain liquid fuel output shares for crude oil refineries, CTL plants, GTL plants and biofuels. Figure 5 shows the output shares for petroleum (and electricity – see discussion further below) that was applied to the SAM. SAPIA, however, does not have the mandate to divulge firm-level information on intermediate input usage and employment within each of these processes. Consequently we had to rely on firms’ Annual Reports and other information sources. Very few firms report on intermediate input usage in the way that it is captured in a SAM, and hence for this part of the data disaggregation we had to rely on a number of assumptions. Fortunately, however, the various production technologies are relatively easy to deduce in terms of the main inputs they use. Crude oil refineries consume all the crude oil in the economy, the CTL process (represented by Sasol) consumes virtually all the coal demanded by the petroleum sector, while GTL (represented by PetroSA) consumes virtually all the gas used in the petroleum industry.8 Finally, all agricultural inputs (field

7 It is a consistency requirement that the disaggregated accounts based on this new information add up to the

original aggregate electricity and petroleum accounts which is achieved by means of appropriate scaling. 8 Natural gas is included under ‘other mining products’ in the SAM.


crops in particular) demanded by the petroleum industry is allocated to biofuels.9 The remainder of intermediate inputs, which only account for a small portion of overall expenditure on intermediate inputs, was allocated across the sub-industries using the output shares shown in Figure 5.

Figure 5: Energy Output Shares in Base SAM: Electricity and Petroleum Sectors

Coal-fired plants93%

Gas turbines1%

Renewable sources1%

Nuclear power5%

Crude oil refineries68%

Biofuels1%Gas to liquids

6%Coal to liquids

25%

Source: South African SAM 2000

Since no information could be obtained on differences in labour intensity across the various sub-industries in the petroleum sector, total employment figures are assumed to be directly proportional to the output shares. As far as skills composition is concerned, external data sources were consulted. In particular, drawing on research published by the International Labour Organisation (ILO) on employment and refinery performance, information on skills breakdown at large and small refineries could be found.10 The occupation categories in this study were mapped to four skills groups used in the SAM to obtain an indication of the skills composition for typical refineries.11 For CTL technologies the SASOL Annual Report for 2006 was used.12 The social performance section in the report provides a breakdown of employment by skill levels, and these are then mapped to the SAM labour accounts to obtain an estimate of the skills composition in CTL technologies. This same skills composition is assumed for GTL technologies as PetroSA was unable to provide the relevant information. Finally, for biofuels, we assume that employment is biased towards lower skill levels relative to the other sub-industries in the petroleum sector.13 Table 31 shows the resulting skills composition in the sub-sectors of the petroleum account.

Estimates of total value added were obtained by multiplying the employment levels with the average wages in the petroleum sector as reported in the original SAM. This assumption implies equality of

9 Biofuels as a fuel source was virtually non-existent in 2000, which is the base year of the SAM. As a result

expenditure on agricultural goods by the petroleum sector was very low and had to be increased if we wanted to assume a 1 per cent output share for biofuels. Hence, some data manipulation was required to produce realistic agricultural input data associated with a 1 per cent biofuels share. However, this did not alter the total output of the agricultural in any meaningful way.

10 See www.ilo.org/english/dialogue/sector/techmeet/tmor98/tmorr.htm. 11 Labour is disaggregated into high-skilled, skilled, semi-skilled and unskilled workers. Table 39 in section 12.5

provides a detailed listing of all the accounts in the SAM. 12 Available online at www.sasol.co.za. 13 This assumption was made after discussions with two experts, Bamikolo Amigun and Harro von Blottnitz, from

UCT’s Department of Chemical Engineering.


wage rates across the four petroleum sub-sectors. Returns to capital (GOS) was allocated across sub-industries using output shares.

Table 31: Skills Compositions for Petroleum Account Disaggregation and Original SAM

Crude oil CTL GTL Biofuels SAM (apetro)

High skilled 30% 29% 29% 17% 30%

Skilled 23% 15% 15% 13% 21%

Semi-skilled 44% 47% 47% 50% 45%

Unskilled 3% 9% 9% 19% 5%

100% 100% 100% 100% 100%

Source: Author’s calculations and South African SAM (2000)

Data from Eskom was used as the primary source of information for the disaggregation of the electricity sector. Eskom supplies over 90 per cent of electricity in South Africa. It was not possible to obtain disaggregated data on intermediate input use, and hence a number of assumptions had to be made about the main types of inputs under each of the four electricity production processes. We consider coal-fired plants, nuclear plants, renewables and gas turbines. All the coal used by the original electricity activity in the SAM was allocated to coal-fired electricity plants. The imported nuclear fuel used at Koeberg is captured under petroleum products account (this is in line with the standard industrial classification scheme used in South African production statistics), hence almost all the expenditure on ‘petroleum’ is allocated to nuclear power plants. Natural gas is included under ‘other mining products’ in the SAM, and all the industry’s expenditure on other mining is allocated to gas turbines. The remainder of intermediate inputs was allocated across the sub-industries using output shares obtained from production capacity of various types of electricity plants owned by Eskom (see Figure 5 above).

Information on labour intensity or ‘jobs per megawatt installed capacity’ is obtained from a report by AGAMA (2003), which draws on Eskom for employment figures in various plants in 2003 and other energy statistics. We use the AGAMA study ‘operational’ employment estimates. In particular, the study finds that there are on average 0.93 jobs/MW in coal-fired plants, 0.54 jobs/MW in nuclear plants, 1 job/MW in renewable energy14 and 0.13 jobs/MW in gas turbines. These direct employment multipliers were used to estimate total employment in each of the four electricity sectors given known output shares from each process. Data on the skills composition within these sub-industries was not readily available. Since electricity from coal makes up 93 per cent of total electricity supplied in South Africa, the skills profile in this industry was assumed to be very similar to overall skills profile of the industry. Information from a nuclear skills study in the United Kingdom was used to arrive at a plausible skills distribution in the nuclear power industry.15 For the renewables sector a skills mix of 65:35 (high skilled to low-skilled) was assumed, whereas for gas turbines this ratio was 70:30. A balancing procedure was applied, resulting in the final skills composition shown in Table 32.

Estimates of total value added were obtained in the same way as before, i.e. by multiplying the employment levels in the respective industries with the average electricity industry wage as reported in the original SAM. Also, returns to capital (GOS) was allocated across sub-industries using output shares. The fully disaggregated petroleum and electricity accounts are shown in Table 40 in section 12.5.

14 The study reports separate employment multipliers for hydro, pumped storage and solar energy sources. A

simple weighted average employment multiplier is derived and used in the calculations here. 15 Nuclear Skills Study 2002, UK DTI, www.dti.gsi.gov.uk


Table 32: Skills Compositions for Electricity Account Disaggregation and Original SAM

Coal Nuclear Renewables Gas turbines SAM (aelec)

High skilled 24% 54% 33% 36% 25%

Skilled 17% 20% 17% 18% 17%

Semi-skilled 43% 19% 33% 32% 42%

Unskilled 16% 8% 17% 15% 15%

100% 100% 100% 100% 100%

Source: Author’s calculations and South African SAM (2000)

12.2.3 Simulation Setup

In this section we describe how each of the individual ‘building blocks’, also called ‘wedges’, of the mitigation scenarios, that is, energy efficiency, structural shifts in output mix, CO2 taxes and investment requirements, are modelled in a CGE model. We also discuss the so-called ‘closure rules’ assumed for this study. The formation of the combined scenarios is described in the results and analysis part of this paper (section 12.3).

12.2.3.1 Modelling Energy Efficiency and Fuel Switching

Increased energy efficiency in production refers to a situation where productive sectors such as mining, manufacturing, commercial and transport sectors improve their production processes in such a way that they require less energy inputs per unit of output produced. In a CGE modelling context this is modelled as a reduction in the intermediate input use coefficient. This coefficient is included as a parameter in the model and specifies the fixed proportion of a given input used per unit of output. Figure 6 shows the production structure used in a typical CGE model. QX is quantity produced in a sector or activity, defined in the CGE model as a CES or Leontief function of aggregate intermediate inputs (QINTA) and value added (QVA). QVA is a CES function of primary factors of production (capital, labour, land etc.), as denoted by F1, … , Fn. QINTA is a Leontief function of various commodities used as intermediate inputs in production (C1, …, Cm).

Figure 6: Two-tier Production Function in a Standard CGE Model

QX

QINTA QVA

F1 F2 Fn…C1 C2 Cm…

CES/Leontief

CESLeontief

Increased energy efficiency, therefore, implies that a producer needs less energy (say, Ce) per unit of output (QX). Increased energy efficiency causes production costs to decline, which has various


downstream effects. Since the standard setup of this class of models assumes perfect competition, output prices are expected to decline with lower production costs.16 Other producers using output from that industry will benefit from lower prices, which enable them to also lower costs. Also, end-use consumer will increase demand due to lower prices, which causes further economic gains to be realised, both in terms of output, employment and general welfare gains for households. The downside, however, is that decreased demand for energy causes a decline in output and employment in those sectors that supply energy-related goods and services, i.e. coal, petroleum and electricity. With impacts pushing in opposite directions, both positive and negative, the use of an economy-wide modelling framework such as a CGE model is important, as it gives an indication of what the overall outcome is likely to be in terms of economic activity, employment and household incomes.

QX in the figure above represents the output level in a given economic sector. The MARKAL model produces results on savings in electricity, coal, gas or liquid fuels across a variety of economic sectors for each mitigation scenario relative to the reference case. These savings are assumed to be implemented in a ‘costless’ way and used as a proxy for energy efficiency, a valid assumption given that output levels in the various scenarios are assumed to be the same as in the reference case. Results are obtained for various mining and manufacturing sectors, the commercial sector and the transport sector. These percentage savings are applied directly to the appropriate input-use coefficients in the CGE model, i.e. by sector and for specific energy commodities.

Of course, in some instances fuel switching may take place under a mitigation scenario. For example, if a transport mitigation action includes a combination of energy efficiency for normal combustion engines as well as modal shifts towards electric- or hybrid motor vehicles, one may expect an increase in electricity use relative to the reference case. Such an increase in energy use per unit of output is modelled in exactly the opposite way as increased efficiency, and hence may appear like increased inefficiency in terms of the results generated. In our modelling framework both increases and decreases in input-use coefficients are considered. Details about the specific changes applied to the input use coefficients are discussed in sections 12.3.1 and 12.3.2 where the setup of the combined scenarios are described.

12.2.3.2 Modelling Structural Shifts

Various mitigation actions are associated with specific structural shifts in output mix. Currently, the South African petroleum and electricity industries are highly dependable on coal and crude oil as intermediate inputs. Coal in particular is associated with high emissions. Therefore, the more progressive a mitigation action, the larger the substitution away from coal is likely to be.

As with energy efficiency, the structural change simulations are set up on the basis of outcomes from the MARKAL energy model. Producers’ output levels are demand-driven in a CGE model, thus under normal circumstances an increase in productive capacity would not lead to an increase in output unless there is demand for the good. One way, however, to induce the model to shift away from the base-level output mix is by reallocating capital stock so that the capacity in targeted industries increases. At the same time capacity in those sectors for which relative output is expected to decline, is decreased, again assumed to be a costless exercise.

The CGE model is set up with a so-called commodity aggregation function, which allows the commodity market to ‘choose’ the sector from which it wishes to source a particular commodity. Suppose, in the electricity commodity market, electricity can be sourced from either coal-fired power plants or from renewable sources. When capacity is increased in renewables and decreased in coal-fired plants, it becomes more expensive for coal-fired plants to produce at the original output level, since they would now be producing beyond capacity. Renewables therefore becomes a relatively cheaper option, and since spare capacity now exists, the cost of taking up this spare capacity is favourable. This leads to a shift along the commodity aggregation function whereby coal-fired electricity is substituted for renewables. This is the desired result of the mitigation action. The

16 Imperfect competition is ignored here as we assume that even with a limited number of market players

(suppliers) there is effective regulation that ensures that lower production costs are indeed translated in lower production prices.


degree of substitutability, or the ease with which the model can substitute between different processes, is determined by the elasticity of substitution parameter, which is set exogenously by the modeller. The degree of substitutability may have important implications (as we see later on, especially in the CO2 emissions tax scenarios) for energy production costs.

This approach of reallocating capital stock between alternative producers rather than simply increasing production capacity in the growing sector is crucial in this modelling context. When increasing total supply of a commodity (say petroleum or electricity) by increasing the production capacity exogenously without adjusting the demand side, market prices will fall. This can simply be explained as supply and demand forces at work (see Figure 7). This is not a desirable outcome as we work from the assumption that petroleum and electricity suppliers will always aim to meet demand, not to exceed it. Therefore, the options in a comparative static framework are to either increase demand in line with increases in supply, or as is the case in these specific simulations, to consider a relative change in the composition of energy supply from different sources, but keeping the total supply of energy unchanged. The latter approach is preferred, as it avoids the need to forecast demand, not only for energy, but for all other commodities as well.

Figure 7: Shifting Supply in an Industry

S1: Base-level supply curve

S3: Horizontal supply curve

D1: Base-level demand

Price

QuantityQ1 Q2 Q3

P1

P2

A

B

C

S2: Exogenous shift in supply

When interpreting the simulation results of structural shifts in output mix, it is important to bear in mind that the simulations represent outcomes under a relative shift in output composition and not an absolute increase in production capacity. Thus, we are interested here in how the structure of the economy might be altered in line with mitigation scenarios, and how this in turn may affect employment and income distribution patterns in the economy. Given that results are reported throughout relative to the reference case this should be straightforward to interpret.

12.2.3.3 Modelling the Impact of a CO2 Emissions Tax

One of the proposed economic instruments that could be used to reduce emissions is a tax on CO2 emissions, something that can be readily analysed in a CGE model. The STAGE model does not make provisioning for CO2 emissions as a measured ‘by-product’ of production, hence a tax on CO2 emissions cannot be modelled directly. The standard approach to modeling CO2 emissions in CGE models is to include emissions as a by-product of production by including it in the production structure. Some form of substitution away from CO2 emissions is then allowed for if the ‘price’ of emissions increases. Thus, under such a model setup, a CO2 tax levied will increase the cost of emitting greenhouse gases, which then acts as an incentive for producers to alter their production processes so that emissions decline. This technology change is represented as a shift along the value-added production function away from CO2 towards more capital (see Van Heerden et al., 2006 for this type of application).


We adopt a suitable alternative modelling approach that works particularly well here, given that the energy accounts (petroleum and electricity) have been disaggregated into various sub-processes. This method involves calculating the implied taxes on the prices of coal, crude oil and natural gas of a given emissions tax level and using these as the proxy for an economic shock of a CO2 tax. This approach is reasonable since an emissions tax is effectively a tax on those inputs that, when processed, emit greenhouse gases. Coal, in particular, and to a lesser extent crude oil and natural gas, all cause emissions. Given information on the relationship between emissions levels and intermediate input use of coal, crude oil and natural gas in various industries, it is possible to derive the implicit taxes on these commodities. As an example, consider the top row of Table 36 in section 12.4.3, which shows various possible CO2 tax levels, expressed as a Rand value per ton of CO2 emitted. The three rows below show the associated taxes on coal, crude oil and natural gas. Emissions taxes have the largest implicit impact on the price of coal. Even a R25/ton emissions tax equates to a 59.4 per cent tax on the price of coal. A R1000/ton tax is equivalent to the price rising by a factor of 25. The equivalent crude oil and gas prices are much lower.17

To explain the difference (from a technical modelling perspective) between the ‘traditional approach’ to modelling a CO2 tax and the approach taken here, consider the following example. If a CO2 emissions tax is levied on electricity generation processes, it becomes ‘expensive’ to emit greenhouse gases. For electricity producers it then becomes economically sensible to alter their production processes by installing additional capital, since the cost of doing so is lower than maintaining the status quo and paying higher emissions taxes. In the traditional approach this is represented as a shift along the value added function (see top part of Figure 8). Under the alternative modelling approach, the increase in the implicit tax of coal causes electricity generation in coal-fired plants to become more expensive. This now means that nuclear power, for example, becomes relatively less expensive, and hence we observe a shift between production technologies in the commodity aggregation function (bottom part of the figure). This approach is only feasible when your model is set up to so that energy commodities (electricity and petroleum) can in fact be produced by alternative processes. This is in fact the case in our model here given the disaggregation of the electricity and petroleum sectors as explained previously.

Figure 8: Modelling a CO2 Emissions Tax in a Standard CGE Model

Capital

CO2

Value added

‘Traditional approach’(e.g. Van Heerden et al., 2006)

Alternative approach in this study

Lines represent relative price of capital, i.e. PCO2/PK.. As PCO2 increases, producers substitute away from CO2 (A → B) in the value added production function.

Nuclear

Coal

Lines represent relative price of nuclear, i.e. PCOAL/PNUC.. As PCOAL increases due to the cost of coal increasing, we substitute towards nuclear (A → B) energy in the commodity aggregation function.

Electricity (commodity supply)

A

B

A

B

Under our approach emissions associated with a given level of output stay unchanged for each sector. The sectors themselves have no option of reducing emissions through adopting new

17 Gas falls under the ‘other mining’ category in the model and only makes up a small part of this sector. In reality

gas prices are likely to rise by about 200 per cent for a R1000 CO2 tax, but this equates to only a 4 per cent rise in other mining prices.


technology. The only option is to shift production capacity to completely different sectors. Ultimately, Van Heerden et al.’s approach will produce results that show a decline in CO2 emissions, but coal use (for example) will stay proportional to energy supply. Our approach, contrast, will show a decline in coal use, which is certainly a desired outcome. Of course, neither approach is perfect, and certainly both have their advantages and disadvantages.

The CGE model contains various instruments that can be used to simulate the impact of changes in the tax regime. The most appropriate for analysing the effect of an emissions tax is the sales tax. The base-level sales tax rate, which is set as a model parameter during the calibration process, is increased in additive fashion using the percentages in Table 36 in section 12.4.3. Thus, if PQSc is the before-tax supply price (equivalent to the producer price), PQDc, the price faced by consumers (including firms purchasing intermediate inputs), can simply be calculated by levying the sales tax (tsc) and the emissions tax (tec) as follows:

18

(1 )c c c cPQD PQS ts te= × + +

A CO2 tax can potentially become an important source of revenue for government. Government’s choice about how to allocate this extra revenue may have implications for growth, job creation and welfare levels of different households. One option is to use additional revenues to finance production subsidies in cleaner energy production technologies (e.g. nuclear or renewables subsidies). Such subsidies will mitigate the impact of higher emissions taxes on energy prices in general, and also further enhance the substitution away from emissions-intensive processes. Government could also recycle revenue through a variety of other mechanisms, including food subsidies, a reduction in VAT, income tax relief or to increase welfare payments to poor households. Various of these options are explored in the economic impact assessment (section 12.4.3).

12.2.3.4 Modelling Investment Requirements

The MARKAL model produces results on various types of ‘costs’ associated with mitigation scenarios. These include capital outlays (building of new plants, or installation of machinery and equipment required) as well as total energy systems costs. Since production costs are inherently captured in a CGE model, we are only interested in the capital outlay cost under each mitigation scenario. In the CGE model this is captured as investment costs. Since we compare mitigation scenario outcomes against the reference case, we are specifically interested in the marginal investment costs as opposed to the absolute level thereof under each scenario, the rationale being that costs under the reference case are investments that would have been made in any event. In the Start Now scenario the investment costs are actually lower than in the GWC reference case, as energy efficiency measures provide cost savings. Hence the target investment level is reduced. For both the Scale Up and Use the Market scenarios investment costs are higher, hence the target investment level is increased and additional funding has to be raised. One of the macroeconomic balances in a CGE model is the savings-investment balance (see further discussions below in section 12.2.3.5). As investments go up, savings have to increase to meet the targeted investment level. When household increase their savings, their disposable incomes decline, thus leading to welfare losses in the current period.

12.2.3.5 Additional Modelling Information: Model Closures

The model is set up with a range of flexible macro adjustment or closure rules. Model closure rules are typically selected with the objective of providing a realistic representation of the adjustment to shocks in the economy under investigation. Mathematically speaking, closure rules ensure that the number of variables and equations in the model are consistent, a necessary condition for the model to solve. In economic terms closure rules define fundamental differences in perceptions of how economic systems operate under adjustment. In particular, the modeller should select closures for the following markets or accounts:

18 In each instance the subscript refers to the commodity (c), e.g. coal, crude oil or other mining, plus all other

commodities in the model. For commodities not affected by the emissions tax tec = 0 .


• The foreign exchange market is cleared via a flexible exchange rate, which is consistent with South Africa’s exchange rate regime. Therefore, the external balance (or current account balance) is fixed. The alternative closure is a fixed exchange rate and a flexible external balance, which is not considered appropriate for South Africa.

• The capital account (also called the savings-investment account), which records all savings and investment related transactions, can be closed in a variety of ways, ultimately ensuring that investment equals savings in the economy. Under the so-called savings-driven closure the investment level is determined by the level of savings in the economy, with average savings rates of households and enterprises fixed. A further option, often regarded as a more balanced approach, is allowing the share of investment expenditure in total final domestic demand remains constant. Since the analyses here use information on required investment levels, we opt for an investment-driven closure. Under this closure the investment level can be considered as fixed at some target level, which implies that households and enterprises generate enough savings to finance investments. This is achieved by allowing average savings rates of households and enterprises to vary.

• The government account in a CGE model is either closed by variations in the level of government borrowing or savings, i.e. the size of the budget deficit or surplus, whereby all tax rates remain constant, or by allowing tax rates to vary in order to generate a level of government revenue sufficient to maintain the base-level budget deficit or surplus. In this model we opt for the latter. In the Start Now and Scale Up scenarios household income taxes are flexible. Under both these scenarios there is very little variation in government revenue, hence taxes are virtually unchanged. The Use the Market scenario affects government revenue directly and significantly, hence in this scenario we opt for a food subsidy as the optimal way (from a welfare perspective) to recycle additional revenue generated through the CO2 tax.

19

• The factor market closure typically involves different treatments for different factors. If labour categories are subdivided into high-skilled and low-skilled groups – a useful distinction in the South African context – a suitable closure is to assume full employment (flexible wages) for high-skilled workers and unemployment or excess capacity (fixed wages) among low-skilled workers. Workers are usually assumed to be mobile across sectors. Capital stock is often treated as activity-specific and fully employed in the short run, while long run simulations sometimes allow capital mobility between sectors. Land is typically fixed and immobile. In these analyses we treat capital stock as fixed and activity specific, since we want to impose structural shifts in production capacity in the various scenarios.

12.2.4 Final Remarks About the Modelling Approach

12.2.4.1 Combined Scenarios: Economic Effects and Modelling

The four separate sets of input parameters from the MARKAL model, namely structural change, energy efficiency/fuel switching, investment requirements, and tax instruments, each have their own unique impacts on GDP, employment and household income and welfare levels. For example:

• Structural change in electricity involves switching from coal-fired plants to nuclear and renewables. These two electricity generation processes have very different skill compositions and labour intensities. Renewables is assumed to be relatively labour intensive compared to coal-fired and nuclear plants. Nuclear, on the other hand, is highly skill intensive and has a low labour intensity when compared to other electricity generation processes.

19 Arguably a more appropriate way of recycling revenue in this context is to subsidise the cleaner alternative

energy supply processes. In fact, this may well be something that policymakers would consider as it would link the CO2 tax directly to processes associated with lower emissions, thus in a sense ‘ring fencing’ the tax. In the analyses here we opted for the recycling scheme that seems optimal from an economy-wide welfare perspective as the default. In a similar study by Van Heerden et al. (2006) this recycling option was also deemed optimal from a welfare perspective. In section 12.4.3 we elaborate by considering a variety of alternative revenue recycling options.


• Energy efficiency gains generally have positive economic effects due to their associated production price decreases. However, these gains may be offset by increased use of other energy sources due to fuel switching (for example, electricity in transport). Both energy efficiency and fuel switching are considered as part of this study, so the outcome depends on the degree to which these two processes offset each other in terms of economic effects.

• The investment effects observed in a comparative static general equilibrium framework generally only have small compositional effects. When investments increase, additional financing has to be raised. The model closure selected for this study assumes that this is achieved through increasing household and enterprise savings rates. Thus, households’ disposable incomes declines, which reduces final demand, while the increase in investments increase final demand. The compositional effects arise due to the fact that structure of household demand is different from that of investment demand in terms of the types of commodities consumed. Any change in GDP, employment or household welfare depends on the differences in production structures (intermediate input use and value added or employment) of the declining sectors versus those of the growing sectors.

• Increased CO2 taxes have implications for the cost of intermediate inputs associated with high emissions levels, i.e. coal, crude oil and gas. This affects energy prices in the economy, which has adverse effects for all energy users, including productive sectors and households. However, CO2 taxes are also a source of revenue for government, and in the event that increased CO2 tax revenues outweigh income tax losses due to the economic decline associated with the tax, overall government revenue may increase. This allows government to redistribute these funds in a variety of ways, which may mitigate some of the effects of increased CO2 taxes.

The overall outcome of individual wedges are sometimes hard to predict beforehand, and so much more so for combined scenarios. For this reason, the use of an economy-wide model that takes into account complex interactions, is important.

12.2.4.2 The Reference Case, Forecasting and Analysis Period

The decision not to attempt to predict or forecast actual trends in all variables in the reference case may seem surprising. The only structural change that we introduce in the reference case is the change in the energy output mix as determined by the MARKAL model. As such the approach here is a kind of hybrid comparative static-dynamic model, i.e. some form of (re)allocation of capital stock is modelled, which is a key element in dynamic models that explicitly model the link between current period investments and changes in capital stock, but we do not adjust the model for changes in production levels, (un)employment levels, population size and so on. Thus, in terms of the rest of the economy (non-energy sectors) the structure and levels of production remain virtually unchanged, apart from the indirect effects of the structural shifts that take place under the reference case. Given the very long modelling period, we do not want a situation where other dynamic changes over time simply overwhelm the mitigation effects. This approach allows us to focus specifically on mitigation, ceteris paribus (keeping all other things constant).

This approach is however justified bearing in mind that what matters most in comparative static modelling where changes are reported relative to a reference case, are the assumptions about how the economy operates, and not necessarily the level at which it operates. The way in which the economy operates is defined by the behavioural assumptions, which are expressed as mathematical equations and based on micro- and macroeconomic theory. The real concern therefore, when looking at results over a very long period of time, is not necessarily whether levels in the base or reference case are correct, but whether the assumptions about how changes filter through the economy via these behavioural equations are accurate. Put differently, one may ask whether the underlying assumption that the adjustment path from the reference case to the outcome stays unchanged over a 50-year time horizon is relevant? This is difficult to say. While the actual functional forms used in CGE models to define behaviour are based on economic theory, the parameters in these functional forms are calculated during the calibration process, which uses the base data, in this case the SAM for 2000. If there is a strong argument that these parameters will change over time, then model results for distant future periods become less reliable.


Given these concerns it was considered advisable to only present results up to 2015, although the model was in fact set up to cover the same period as the MARKAL model, i.e. up to 2050. We turn to the model results for the combined scenarios, Start Now, Scale Up and Use the Market in the next section.

12.3 Results and Analyses

12.3.1 Start Now and Scale Up

12.3.1.1 Simulation Setup

Results from the MARKAL model are used to obtain estimates of structural shifts in the output mix under various scenarios, including the reference case. The model predicts that electricity supply from coal will drop from around 94 per cent in 2005 to 81 per cent by 2050 in the reference case (see Table 42 in section 12.5). Also in this scenario the nuclear and renewables shares rise from 6 to 14 per cent and zero to 4 per cent over the same period. However, by 2015 the reference case is not much different from the base, with the electricity share from coal actually predicted to rise marginally to 95 per cent, while the share from nuclear, renewables and gas are virtually unchanged.

Under the Start Now scenario the electricity supply from coal declines to 46 per cent by 2050, while nuclear and renewables each contribute around 27 per cent at this point. Most of this relative decline in electricity from coal takes place after 2015, as shown by the MARKAL model results that predict this share to still be quite high in 2015 (87 per cent). The shares from nuclear and renewables are predicted to reach 8 and 5 per cent respectively by 2015. Under the Scale Up scenario a more aggressive decarbonisation strategy sees the coal share drop further to 17 per cent, with nuclear and renewables each contributing about 41 per cent to electricity supply by 2050. However, again much of this change takes place after 2015, as the shares from coal, nuclear and renewables are fairly close to the base at this point, i.e. 86, 8 and 6 per cent.

The Start Now and Scale Up mitigation scenarios in the petroleum sector (also in Table 42) are less aggressive, with little variation in output shares from crude oil refineries and CTL processes in either of these scenarios relative to the reference case. The reference case itself also differs only marginally from the original base in 2000. However, in both the Start Now and Scale Up scenarios the biofuels sector grows from a base of around zero to almost 4 per cent of liquid fuels supply by 2050. While this presents a large growth for the biofuels sector itself, it does not alter the liquid fuels mix significantly and is therefore unlikely to have a large impact on the economy. By 2015 there is very little difference between the petroleum output shares at that point compared to the model base in 2000.

Given that we only report the results up to 2015, it should be clear from the above that not much of the changes observed under the combined scenarios, Start Now and Scale Up, are due to structural shifts. Most of the structural shifts in energy output mix are predicted by the MARKAL model to take place after 2015. Changes observed are therefore largely explained by energy efficiency and fuel switching. We turn to this next.

MARKAL model results are also used to obtain estimates of changes in intermediate input demand associated with energy efficiency or fuel switching. Table 43 in section 12.5 shows percentage changes in fuel use per unit of output relative to the reference case. The reference case assumes no efficiency gains or fuel switching. The first part of the table shows savings in electricity used as intermediate inputs into mining, industrial and commercial production processes for the Start Now scenario. Due to shifts towards electrified transport, overall electricity used in the economy declines by about 2 per cent by 2050.20 This is not significantly higher than the savings by 2015, which suggests that much of the efficiency gains are predicted to take place within the next decade. By 2050 the estimated decline in coal use due to energy efficiency under the Start Now scenario is about

20 Table 43 also shows the percentage decline in electricity demanded as an intermediate input in the economy.

Here we only discuss the decline (or increase) in total electricity demand. The same applies to other energy inputs, i.e. coal, petroleum and gas.


15 per cent total coal supplied. This efficiency gain takes place more gradually over time, with the decline by 2015 estimated at only 7 per cent. Also shown are declines in petroleum use, driven largely by fuel savings in the transport sector. By 2050 petroleum use (liquid fuels such as petrol and diesel) is likely to decline by almost 9 per cent per unit of output, compared to 7 per cent by 2015.

The Scale Up scenario assumes very similar industrial and commercial efficiency increases (electricity and coal). However, a greater shift towards electrified transport actually causes electricity use to increase marginally by 2050 (0.4 per cent). By 2015, however, fuel switching has not yet caused a net increase in electricity, with electricity savings of around 4 per cent expected by this time. Petroleum savings are slightly lower under this scenario compared to the Start Now scenario.

The investment cost estimates (in terms of capital outlay required) under each mitigation scenario is expressed relative to the reference case. In the Start Now scenario these are actually negative, implying that capital outlay under this scenario is less than under the reference case with no loss in terms of production. Investments are initially about 5 per cent below that of the reference case, and thereafter drops further to around 10 per cent below the reference case (see Figure 26 in section 12.5). The Scale Up scenario is almost the opposite, with investments required estimated to be about 5 to 10 per cent higher than under the reference case level over most of the period, reaching a high of 12 per cent by 2050.

12.3.1.2 GDP, Employment and Welfare Effects

Before looking at the results, a brief note on the degree of substitutability assumed in the commodity aggregation function and how this affects prices (see previous discussions in section 12.2.3.2). In the analyses here we are particularly interested in how easily substitution can take place within the electricity and petroleum sectors once capital has been reallocated in the structural shift simulations. We select a moderate elasticity of substitution for all the scenarios. This causes energy prices to rise, especially in the latter periods when substitution away from carbon-based processed is ‘pushed hard’. If we were to assume perfect substitutability, for example, prices would not have risen as much, if at all. Our approach, although more conservative, is considered more appropriate given the general consensus that mitigation actions will probably lead to rising energy prices. A lower substitutability also reflects the fact that commodities produced using different processes are ultimately not homogenous, and that some adjustment costs will have to be borne by the economy, particularly when producers have to alter production processes to accommodate slightly different commodities.

We first consider the impact under the Start Now and Scale Up scenarios on GDP. Figure 9 shows the percentage difference between GDP under each scenario relative to the reference case. Under the Start Now scenario GDP is marginally lower than in the reference case during the initial period, but recovers to a very similar level as the reference case by 2015. Since investment costs under the Start

Now scenario are lower than under the reference case, some interesting household effects are observed, which are discussed in more detail below. This, however, has important implications for GDP levels. Lower investment levels allow consumers (households) to reduce savings, which frees up more funds for consumption. While this is good for consumers from a hedonistic welfare perspective, it is probably short-sighted. If investment levels were maintained, the future production capacity of the economy could be increased more; this would have positive production and employment effects. The dashed line in Figure 9 represents GDP that could be realised if investment levels were maintained at the base level rather than allowing these levels to decline. This ‘potential’ GDP measure excludes the impact that such investment could have on the future production capacity in the economy, and as such only measures the immediate or short term impact of increased investment flows. Clearly, however, maintaining investment levels would imply that positive growth effects are in fact observed over the period, and hence this is something that should be encouraged.


Figure 9: GDP Effects of the Start Now and Scale Up Scenarios

-0.6%

-0.4%

-0.2%

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%

1.2%

2005 2010 2015

% change relative to reference case

Start Now Scale Up Start Now - maintain investment

Source: CGE model results

The Scale Up scenario compares favourably to the reference case and the Start Now scenario as far as GDP is concerned. This is driven largely by the higher level of energy efficiency modelled under this scenario, and also possibly by the higher investment levels under this scenario.

GDP is effectively a measure of value added in the production, which is the sum of producers’ payments for labour (wages), capital and land. Thus, employment effects generally look similar in shape to the GDP effects, at least in the aggregate. As discussed, the factor market closure in the model assumes that excess capacity (unemployment) exists among semi- and unskilled workers (referred to as low-skilled workers), hence their employment levels are flexible and wages are fixed. Skilled and high-skilled workers (high-skilled), on the other hand, are fully employed at flexible wages, reflecting the skill constraints in the South African economy. Figure 10 shows the employment and wage effects for these two groups of workers respectively. At this disaggregated level a better picture is obtained of the relative gains and losses of different types of workers.

Under the Start Now scenario employment levels of semi-skilled workers is below that of the reference case. By 2015 semi-skilled employment is likely to be about 2 per cent lower than the reference case. Unskilled employment, on the other hand, remains above that of the reference case, reaching about 1 per cent by 2015. Under the Scale Up scenario, semi- and unskilled employment levels both remain positive relative to the reference case, with semi-skilled employment peaking at about 3 per cent by 2015. Unskilled employment is marginally lower under this scenario compared to the Start Now scenario.


Figure 10: Employment and Wage Effects of the Start Now and Use the Market Scenarios

-3%

-2%

-1%

0%

1%

2%

3%

2005 2010 2015


Start Now Semi-skilled Start Now Unskilled

Scale Up Semi-skilled Scale Up Unskilled

-0.2%

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%

1.2%

1.4%

1.6%

1.8%

2005 2010 2015% change relative to reference case

Start Now High-skilled Start Now Skilled

Scale Up High-skilled Scale Up Skilled


Wage changes under the Start Now and Scale Up scenarios are quite similar for skilled and high-skilled workers respectively, all showing increasingly higher levels relative to the reference case up to 2015. High-skilled workers are set to gain the most in the Start Now scenario, with wage levels almost 2 per cent higher than the reference case by 2015. In contrast, skilled workers gain marginally more in the Scale Up scenario, with wages estimated to be about 0.7 per cent above that of the reference case.

Welfare is measured using the equivalent variation (EV) measure. This index measures the welfare levels of households taking into account changes in disposable income (i.e. net income, after tax and savings have been deducted) as well as movements in household-specific price indices. Under the investment-driven closure selected households will reduce savings when, as happens in the Start

Now scenario, required investment levels decline. Since high income households contribute the bulk of savings in South Africa, they also benefit the most from a reduction in required savings rates. This results in a large and significant increase in disposable income for this household group. Therefore, as shown in Figure 11, despite the fact that GDP is lower than that of the reference case between 2005 and 2010, high income households experience large, positive welfare effects at significantly higher levels than any of the other household groups. In fact, all the household groups experience positive welfare effects because of their increased disposable incomes, at least during the period under consideration.


Figure 11: Household Welfare Effects of the Start Now and Scale Up Scenarios

0

2

4

6

8

10

12

14

2005 2010 2015

Start Now

Index value

Low income Lower middle income

Upper middle income High income

-6

-5

-4

-3

-2

-1

0

1

2005 2010 2015

Scale Up

Index value




Under the Scale Up scenario the effects are almost the exact opposite. Increased investment requirements cause households, in particular high income households, to increase savings rates in order to generate funding for the investments, which reduces spending power and hence welfare. Even the generally favourable employment effects do little to counter these welfare losses, except for low-income households by 2015.

12.3.1.3 Sensitivity of Results

The preceding discussion already raised the importance of the assumed elasticity of substitution in the commodity aggregation function. An increased substitutability, for example, will cause the negative impacts in the latter periods to be less severe than what the current model results suggest. Also, learning-by-doing or improved technologies are often important in bringing down production costs over time. It may well be that some of these alternative energy supply processes become least-cost optimal production choices in any event if their associated production costs decline in line with technological gains. It may further be that adverse movements in world prices of crude oil and coal could create further incentives for energy supply sectors to switch to alternative processes.

The way in which investment is modelled here also has implications for the outcome. Although investments, as argued, usually only have small compositional effects on the economy, the way in which they are financed may be important in determining the direction of the small compositional effects. The way in which the model is set up assumes that households and enterprises raise additional funding through increased savings. This affects current consumption and welfare levels of households, as shown. Investments may also simply crowd out other investments if the economy is savings-constrained, something that is quite a likely in South Africa given low savings rates, especially among middle- and lower income households. Another alternative that could be considered is to raise funding (through loans) offshore. While, ultimately, such loans still have to paid back with interest, the full impact is not felt within a single year or observation period as we model it here.21 Finally, the import content of new investments matters. If the investments under a scenario lead to an increased demand for imported equipment and machinery, funds will leave South Africa, which ultimately impacts negatively on current GDP. The modelling here basically assumes that the import shares of the base period are preserved, with some degree of flexibility depending on

21 In reality we only model the incremental cost (in the case of rising investment requirements as in the Scale Up

scenario), so the full effect of the energy system cost is not borne by the economy in that year that the investment is made. We further smooth the investment cost vector using moving averages so as to ‘spread’ the burden over longer periods of time.


how relative prices of imports versus domestically produced goods vary. It may well be that the types of investments required under the mitigation scenarios have higher import propensities than general investments in the economy.

12.3.2 Use the Market

12.3.2.1 Simulation Setup

Selected MARKAL model results are also used as simulation parameters in the Use the Market scenario. This scenario sees the use of coal in electricity generation virtually wiped out by 2030, with output share of coal-fired electricity plants declining to 2 per cent, and zero per cent from 2040 onwards (see Table 42 in section 12.5). Again we only analyse and discuss results for the period 2005 to 2015. Even by 2015 coal-fired electricity drops down to 64 per cent, with each of nuclear and renewables contributing a rather substantial 18 per cent each. In terms of the petroleum output mix the Use the Market scenario is also quite aggressive, with CTL output falling to zero by 2030. The share by 2015 is 21 per cent. Most of the CTL output is replaced by crude oil refineries, which implies larger dependence on imported crude oil.

The Use the Market scenario takes a very different angle than the Start Now and Scale Up scenarios as far as energy efficiency is concerned (see Table 43 in section 12.5). The focus in this scenario is much more on fuel switching. Electricity use in mining, manufacturing and commerce does not decline as much as in the other scenarios, while the use of electrified transport is increased even more than in the Scale Up scenario. Consequently electricity use increases quite substantially by about 6 per cent by 2050. However, by 2015 electricity use is down by just over 1 per cent, which implies that fuel switching has not yet started to take place at this point. The Use the Market scenario also considers switching away from coal towards gas as a thermal fuel source. Much of this fuel switching only happens after 2040, and hence does not affect results much in the 2005 to 2015 period reported on here. The MARKAL model results predict that coal use is likely to decline by about 21 per cent relative to the reference case, while natural gas use is likely to increase by over 300 per cent by 2050.22 However, by 2015 none of these changes have yet set in. Petroleum use declines by about 9 per cent by 2050 due to fuel efficiency in transport. This decline is of a similar magnitude to the Start Now scenario. The related decline by 2015 is just over 4 per cent.

As far as investment is concerned the Use the Market scenario initially (by 2015) requires investment levels of up to 20 per cent above the reference case investment levels, but thereafter it drops back to similar levels as the reference case between 2030 and 2050. The CO2 emissions taxes that form a core part of the Use the Market scenario are implemented as an incremental tax in the MARKAL model, ranging from about R250 per ton of emissions in 2008 and increasing to R750 by 2050. The level in 2015 is R353 per ton of CO2. As explained previously (see discussions in section 12.2.3.3) the actual CGE model shock is implemented as an increase in taxes on coal, crude oil and gas. Table 44 in section 12.5 shows the CO2 tax levels in selected years together with the related commodity taxes. Additional revenue from the CO2 tax is recycled in the form of a food subsidy, given that this form of recycling appears to have the most favourable outcome in terms GDP, employment and welfare levels among the poor (see section 12.4.3). This finding is also consistent with that of Van Heerden et al. (2006). Whether this recycling option is politically feasible remains a question for policymakers to consider (see footnote 19).

12.3.2.2 Modelling Issues

Various modelling problems were encountered when attempting to model the Use the Market scenario in a similar way as the Start Now and Scale Up scenarios. The structural change shock under the Use the Market scenario is very severe, especially towards the latter periods. Given the assumption about imperfect substitutability in the commodity aggregation function, energy prices rise significantly in the CGE model, causing the entire economy to take a very big knock as far as real GDP levels are concerned. When run in isolation, the structural change ‘wedge’ of the Use the

22 This is a very crude estimate. Natural gas is not captured as a separate commodity in the model and falls under

other mining commodities, thus making calculations and the modelling of this scenario very difficult. See discussions below.


Market scenarios offers solutions up to 2050, by which time there is virtually no coal used in the domestic economy. However, once the other ‘building blocks’ or wedges of the scenario are added, i.e. energy efficiency/fuel switching, investment requirements and the CO2 tax, the model only offers solutions up to the 2030. It is especially the large increase in gas demanded under the energy efficiency/fuel switching wedge that causes this scenario to create infeasible solutions for the period 2040 to 2050. As expected, the impact of the CO2 tax is small towards the latter period, since most carbon-based processes are removed from the economy by this time. In the absence of modelling some exogenous change in the model that mitigates the negative effects of the Use the Market scenario towards the latter end of the analysis period, results beyond 2030 may seem outrageous

Another concern, as mentioned briefly before, is that natural gas is not a separate commodity in the SAM. Hence it is difficult to define the simulation parameters, as assumptions have to be made about what share of other mining commodities is made up of gas. While in reality additional gas will most likely be imported, it is difficult to ‘force’ the model to choose this option since demand for imports in this model ultimately depends on the relative prices of domestically produced goods vis-à-vis imported goods.

These concerns contributed to the decision to only report results up to 2015, by which time structural shifts, energy efficiency and fuel switching, and CO2 tax levels are not too restrictive to the economy. A more comprehensive analysis of the impact of CO2 taxes in the absence of other mitigation wedges was also done and is reported on in section 12.4.3.

12.3.2.3 GDP, Employment and Welfare Effects

The combined effects of a sharp decline in the coal sector and sharply rising energy prices, driven initially by a CO2 tax, and later by rising energy prices associated with the structural shifts in the energy output mix, causes GDP to decline fairly rapidly in this scenario (see Figure 12). By 2015 GDP is likely to be about 2 per cent below the reference case level, which stands in stark contrast to the Start Now and Scale Up scenarios where gains are in fact expected.

Figure 12: Comparing GDP Effects of Mitigation Scenarios

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2005 2010 2015


Start Now Scale Up Use the Market


The employment effects are interesting, with unskilled and semi-skilled employment initially increasing. This is thanks largely to strong initial growth in the food sector as additional revenue from the CO2 tax is recycled back into the economy as a food subsidy. Output and employment in this sector, as well as the agricultural sector, which is an important supplier to the food processing industry, increases due to strong consumer demand growth. However, by 2015 and beyond (not


reported here) the employment impact for low-skilled workers becomes zero and then negative as the sharp decline in economic activity and employment losses in other economic sectors outweigh the small employment gains in the initial period. Wages of skilled and high-skilled workers decline from the outset. High-skilled workers suffer the biggest losses, with wages falling by about 5 per cent relative to the reference case by 2015.

Figure 13: Employment Effects of the Use the Market Scenario

Change in low-skilled employment

-1%

0%

1%

1%

2%

2%

3%

3%

4%

2005 2010 2015


Semi-skilled Unskilled

Change in high-skilled wages

-6%

-5%

-4%

-3%

-2%

-1%

0%

2005 2010 2015


High-skilled Skilled


The welfare effects are generally negative, with rising prices and reduced wage income impacting negatively on spending power and hence welfare levels. High income households experience the greatest welfare losses, given that they have to raise most of the additional savings required to finance investments, while in relative terms they do not gain as much from food subsidies. As expected, low income households initially benefit from employment growth and the large food subsidy. Low-income households spend a large proportion of their budget on food, hence this result. However, by 2015 much of the initial gains are mitigated by disposable income losses due to the negative labour market effects and rising prices in the economy.


Figure 14: Household Welfare Effects of the Use the Market Scenario

-40

-35

-30

-25

-20

-15

-10

-5

0

5

2005 2010 2015

Index value




12.3.2.4 Final Remarks

In conclusion, we have to reiterate that these results are only indicative of an outcome under a highly restricted economy. In order to try and reproduce structural shifts as predicted by the MARKAL model, the capital stock levels in the energy supply sectors are ‘locked down’ to try and force a certain output mix. It may well be that if the CGE model was allowed to allocate capital in the most efficient way (as is done in the additional CO2 tax analyses reported on in section 12.4.3), that the effect on the economy would not be so large. This highlights one of the difficulties in linking the MARKAL model and the CGE model is sequential manner, as these two models overlap in certain respects.

12.4 Analysing the Effects of Individual Mitigation Components (Wedges)

12.4.1 Energy Efficiency Wedges23

12.4.1.1 Overview

The comparative static computable general equilibrium (CGE) model for South Africa mentioned in the previous section is used to model the effects of increased energy efficiency, one of the proposed mitigation components or ‘wedges’ of the LTMS process. These and other components discussed in this section are combined in various ways to create the Start Now, Scale Up and Use the Market scenarios discussed in the previous section. This section should therefore be seen as informing the reader about the underlying individual impacts.

Energy efficiency in an economic sector is modelled as a reduction in demand for primary or transformed energy sources per unit of output. The analysis considers mining and industrial energy efficiency, commercial energy efficiency and energy efficiency in the freight and passenger transport sectors. Increased residential energy efficiency cannot be analysed in this modelling framework. Below we report on some selected results obtained for industrial and commercial energy efficiency. Results are reported as percentage changes relative to a reference case which assumes zero energy efficiency gains. The simulation parameters, that is, the percentages by which energy demand declines per unit of output are obtained from MARKAL model results for related mitigation wedges.

23 A summary of results is supplied in this section. Please refer to the earlier SBT 5 Technical Document for

further details.


12.4.1.2 Industrial Energy Efficiency

The industrial energy efficiency scenarios consider efficiency gains in the use of electricity and coal in the production processes of manufacturing and mining sectors. Under the industrial energy efficiency scenario electricity demand declines by 22 per cent in 2020, reaching 29 per cent by 2050. Coal use declines by 20 per cent by 2020 and 45 per cent by 2040 (see Table 33 below). These reductions are expressed per unit of output.

The industrial energy efficiency simulation results are summarised in Table 33. For the electricity savings scenarios the CGE model shows a 7.7 and 10 per cent decline in electricity output. Despite this, overall economic activity increases by about 0.2 per cent in both periods. This is partly due to lower producer prices brought about by electricity savings (see PPI in Table 33), which acts as a stimulus for aggregate demand. This results in an increase in demand for labour relative to the base case: wages of skilled workers rise by 0.5 and 0.7 per cent, while employment among low-skilled workers rises by 0.5 per cent in both periods.24 The GDP measure increases only marginally by 0.4 and 0.5 per cent in 2020 and 2050.25 Given the small changes in employment income there are no significant income distribution effects. However, positive welfare effects (as measured in Table 33 by aggregate household expenditure levels) are experienced across all representative household groups in the model.

Under the thermal energy efficiency scenarios overall coal production declines by 12.2 and 29 per cent in 2020 and 2040 respectively. Again, despite these relatively large declines in one sector’s output levels, overall economic activity increases by 0.2 and 0.5 per cent in the two periods, with aggregate demand stimulated by lower producer and consumer prices. The employment effects are slightly higher than under the electricity scenarios; skilled wages increase by 0.5 and 1.1 per cent, and low-skilled employment increases by 0.3 and 0.8 per cent in the two periods. The comparative static GDP impact is also slightly higher, measured at 0.4 and 0.9 per cent respectively. Changes in household expenditure levels are marginally higher among high income households.

12.4.1.3 Commercial Energy Efficiency

The commercial energy efficiency scenarios consider efficiency gains in the use of electricity in the production processes of various services sectors. The weighted average decline in electricity use in the commercial sectors is about 8 per cent in 2015, reaching 15 per cent by 2030 and staying at this level through 2050 (Table 33). Despite the relative size of the commercial sectors in terms of contribution to GDP, electricity use is relatively low compared to industrial sectors. As a result the commercial energy efficiency scenarios have a limited impact on the economy in terms of overall production levels, GDP, employment and household income changes. As shown in Table 33, overall electricity supply declines by 1.5 and 2.7 per cent in 2015 and 2030. A small rise in aggregate demand, however, leads to a 0.02 and 0.04 per cent increase in overall economic activity. Changes in skilled wages, low-skilled employment and household expenditure levels (welfare) are all small but positive (around 0.1 and 0.2 per cent).

24 The model assumes full employment among high skilled workers, and excess capacity (or unemployment)

among lower skilled workers. 25 In simulations of this nature the GDP change should be understood as the comparative static GDP estimate, i.e.

the percentage difference between simulated GDP and the base-level GDP, and not the GDP growth level.


Table 33: Results from Energy Efficiency Simulations: Percentage Changes Relative to Base Case

Industrial Energy Efficiency

Electricity

Efficiency

Thermal

Efficiency (Coal)

Commercial Energy

Efficiency

(Electricity)

2020 2050 2020 2040 2015 2030

Simulation: Weighted average decline in electricity/coal use in industry or commerce 21.9% 28.5% 19.5% 45.2% 8.0% 15.0%

Domestic production effects (activity output)

Electricity supply -7.7% -10.0% -1.5% -2.7%

Coal supply -12.2% -29.0%

Economy-wide production 0.2% 0.2% 0.2% 0.5% 0.02% 0.04%

Production prices (PPI) -0.06% -0.08% -0.03% -0.06% 0.00% -0.01%

Changes in wages/employment and value added

Skilled wages 0.5% 0.7% 0.5% 1.1% 0.2% 0.1%

Low-skilled employment 0.5% 0.6% 0.3% 0.8% 0.1% 0.2%

Gross Domestic Product (value added) 0.4% 0.5% 0.4% 0.9% 0.1% 0.1%

Changes in household expenditure/welfare

Low income 0.3% 0.4% 0.3% 0.7% 0.1% 0.1%

Middle income 0.4% 0.5% 0.4% 0.8% 0.1% 0.1%

High income 0.3% 0.4% 0.4% 0.9% 0.1% 0.1%


In conclusion, energy efficiency gains modelled here generally have small but positive overall production effects in the economy. Output and employment losses in the coal mining and electricity generation sectors are generally offset by gains in other sectors that benefit from lower production costs, resulting in unambiguously positive but small employment effects. Household welfare effects are also small but positive, with the distribution of gains depending on the type of energy efficiency modelled; for example, electricity efficiency appears to benefit low and high income households least, while high income households gain most from thermal efficiency gains. These distributional effects, however, are too small to be overly concerned about the socio-economic implications. Of course, model results from the mining, industrial and commercial energy efficiency scenarios are not directly comparable given that the input parameters for each simulation as well as the production structures of these sectors differ significantly.

12.4.2 Structural Change Wedges26

12.4.2.1 Overview

In these scenarios the economic implications of a relative shift in energy supply away from carbon-based or emissions-intensive production processes towards cleaner, more environmentally friendly production processes are investigated. Three main mitigation scenarios are considered, namely a renewables intensive and a nuclear intensive scenario for electricity generation, and a biofuels scenario for liquid fuel supply. The results under each of these outcomes are compared against a reference case.

The study uses a SAM multiplier model27 to analyse the structural change effects. The model produces results on sectoral output levels, employment and household incomes associated with the structural shifts in the composition of energy supply. The input-output table was adjusted to

26 A summary of results is supplied in this section. Please refer to the earlier SBT 5 Technical Document for

further details. 27 This is similar to an input-output model, except that the SAM allows for so-called multi-product activities and

also links information on factors of production (employment), households, government, savings and investments, as well as trade flows. See technical document from SBT 5 for more information about this type of modelling.


incorporate different production processes within the electricity (coal-fired plants, nuclear power stations, renewable energy and gas turbines) and petroleum (crude oil refineries, CTL plants, GTL plants and biofuels) sectors.28

12.4.2.2 Reference Case

The reference case itself, in this case ‘growth without constraints’, also assumes structural shifts in the output mix of electricity and petroleum over time. The MARKAL model results show the least cost optimisation energy output shares for the electricity and petroleum sectors. The shares by 2015, 2030 and 2050 are extracted and used to generate a counterfactual ‘path’ in the SAM multiplier model. This represents the reference case against which results under mitigation actions can be compared.29

12.4.2.3 Nuclear Intensive Scenario and Renewables Scenarios for the Electricity Sector

Under the nuclear intensive scenario there is a strong drive to increase the electricity output share from nuclear power. As shown in Figure 15 the electricity output share under the nuclear intensive scenario is no different from the reference case in 2015. Consequently no change from the reference case scenario is reported for this year. By 2030 the nuclear share rises rapidly to 27 per cent and remains roughly constant at this level through 2050. The reference case changes somewhat between 2030 and 2050.

Figure 15: Comparison of Output Shares under the Nuclear Intensive Scenario and Reference Case

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Output shares (%)

Coal-fired Nuclear Renewables Gas turbines

Gas turbines 1.0% 0.2% 0.2% 0.1% 0.1% 9.7% 1.2%

Renewables 1.0% 6.0% 6.0% 4.2% 4.0% 8.5% 2.6%

Nuclear 5.0% 4.6% 4.6% 2.9% 26.7% 8.6% 27.1%

Coal-fired 93.0% 89.2% 89.2% 92.7% 69.2% 73.2% 69.0%

2000 Reference Nuclear Reference Nuclear Reference Nuclear

Base 2015 2030 2050

Source: MARKAL model results for initial wedges.

Note: “Base” refers to the SAM multiplier model base for the year 2000. The scenario results are compared

against the reference case.

Under the renewables scenario the renewables output share increases to 11.3, 30.9 and 32.3 per cent in 2015, 2030 and 2050, relative to the reference case shares of 6.0, 4.2 and 8.5 per cent (see Figure 16). As is the case with the nuclear scenario, the output shares under the renewables scenario remains fairly stable between 2030 and 2050, but since the output shares in the reference case do change, some differences between the renewables and reference cases will emerge in the results.

28 This is a similar adjustment to the one explained in section 12.2.2.2 of this report. 29 The structural shifts in output mix reported here are for individual wedges, and therefore differ from those used

in the combined scenarios, Start NowStart Now, Scale UpScale Up and Use the MarketUse the Market.


Figure 16: Comparison of Output Shares under the Renewables Scenario and Reference Case

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%Output shares (%)

Coal-fired Nuclear Renewables Gas turbines

Gas turbines 1.0% 0.2% 1.1% 0.1% 10.3% 9.7% 6.7%

Renewables 1.0% 6.0% 11.3% 4.2% 30.9% 8.5% 32.3%

Nuclear 5.0% 4.6% 4.3% 2.9% 2.4% 8.6% 1.9%

Coal-fired 93.0% 89.2% 83.3% 92.7% 56.3% 73.2% 59.1%

2000 Reference Renew Reference Renew Reference Renew

Base 2015 2030 2050

Source: Earlier MARKAL model results



The percentage change in production under the nuclear intensive scenario compared against the reference case is shown in Table 34. As expected, the demand for coal and lignite products declines significantly (4.4 per cent) under this scenario in 2030, given the drop in electricity generated in coal-fired plants (-25.3 per cent). By 2050, however, under the reference case, output from nuclear energy also rises relative to electricity from coal-fired plants. Hence the change in coal output (relative to the reference case) is lower at this point. The reference case also envisages a sharp increase in electricity output from renewable sources and gas, hence under the nuclear scenario, which does not rely on these energy sources, comparative output levels are much lower.

Under the renewables scenario we notice declines in output from coal mining due to the decline in electricity from coal-fired plants. We also note a decline in nuclear output, which is an important future electricity source in the reference case. Here it is replaced by electricity from renewable sources. The large percentage increase in output from renewables is indicative of the small base from which it grows.


Table 34: Production Levels under the Nuclear/Renewables Scenario Compared Against the Reference Case

Nuclear intensive scenario Renewables intensive scenario

2015 2030 2050 2015 2030 2050

Coal and lignite 0.0% -4.4% -0.8% -1.1% -6.9% -2.9%

Crude oil and other mining 0.0% 0.1% 0.0% 0.0% 0.1% 0.0%

Petroleum 0.0% 0.9% 0.7% 0.0% 0.1% -0.3%

Electricity: Coal-fired 0.0% -25.3% -6.0% -6.6% -38.8% -19.1%

Electricity: Nuclear 0.0% 816.9% 214.9% -5.0% -17.4% -78.4%

Electricity: Renewable sources 0.0% -5.9% -68.9% 86.9% 636.3% 281.6%

Electricity: Gas turbines 0.0% -0.6% -87.4% 524.6% 7047.7% -30.8%

Total activity output 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

Source: SAM multiplier model results

Despite output levels remaining stable, small employment effects can be observed when comparing the nuclear scenario against the reference case. Output-employment ratios and skills intensities in nuclear power plants are different from those of other electricity generation processes.30 Hence we expect to see some relative shifts in employment levels and/or skills distributions. Figure 17 shows the percentage changes in employment under the nuclear scenario compared to the reference case, disaggregated by skill. Given lower output employment ratios in nuclear power plants all skill classes experience negative employment effects relative to the reference case. However, high-skilled workers are likely to be less affected, with employment dropping by only 0.02 per cent compared to the reference case, compared to 0.08 per cent for low-skilled workers by 2050.31

Figure 17 also shows the changes in per capita income levels. Given the small overall employment changes relative to the reference case, income changes are small, yet negative across all household types. Given the skills changes under the nuclear scenario and the fact that low-skilled workers are typically attached to lower income households, poorer households are likely to be disadvantaged more.

30 Information on labour intensity or ‘jobs per megawatt installed capacity’ (operational multipliers only) is

obtained from a report by AGAMA (2003), which draws on Eskom employment figures in various plants in 2003. In particular, the study finds that there are on average 0.93 jobs/MW in coal-fired plants, 0.54 jobs/MW in nuclear plants, 1 job/MW in renewable energy (average of hydro, pumped storage and solar energy) and 0.13 jobs/MW in gas turbines. Data on the skills composition within these sub-industries was not readily available, and hence assumptions had to be made. For coal the assumed skilled to unskilled ratio is 41:59, for nuclear 74:26, for renewables 50:50 and for gas 43:57.

31 In an input-output model prices (and hence wages) are considered fixed, hence in contrast to the CGE model we interpret a change in the value added of labour as a change in employment, irrespective of the skill level. Therefore, the assumption that excess production capacity exists also extends to the labour market.


Figure 17: Employment Effects and Per Capita Incomes under the Nuclear Scenario Compared Against the Reference Case (GWC)

Employment

-0.12%

-0.10%

-0.08%

-0.06%

-0.04%

-0.02%

0.00%

2015 2030 2050

Percentage change from reference case

High-skilled and skilled Semi- and unskilled

Per capita income

-0.14%

-0.12%

-0.10%

-0.08%

-0.06%

-0.04%

-0.02%

0.00%

2015 2030 2050


Low inc Lwr-mid inc Upp-midd inc High inc


Small employment effects can also be observed when comparing the renewables scenario against the reference case. As before, results observed are of course sensitive to the assumed output-employment ratios (relatively high for renewables) and skills intensities (lower skills intensity than, for example, nuclear power, but higher than coal) (see footnote 30). Figure 18 shows the percentage changes in employment under the renewables scenario compared to the reference case, disaggregated by skill. Employment levels rise marginally relative to the reference case, with high-skilled workers likely to gain relatively more. As far as per capita incomes are concerned all household groups gain. However, in terms of the distributional effects we again notice that high-income households gain relatively more, thus leading to increased inequality. These changes are, however, very small, with changes below 0.12 per cent unlikely to have any significant effect on an inequality measure such as the Gini coefficient.


Figure 18: Employment Effects and Per Capita Incomes under the Renewables Scenario Compared Against the Reference Case (GWC)

Employment

0.00%

0.02%

0.04%

0.06%

0.08%

0.10%

0.12%

0.14%

0.16%

0.18%

2015 2030 2050


High-skilled and skilled Semi- and unskilled

Per capita income

0.00%

0.02%

0.04%

0.06%

0.08%

0.10%

0.12%

0.14%

2015 2030 2050Percentage change from reference case



12.4.2.4 Biofuels Scenario for the Petroleum Sector

The biofuels scenario is preliminarily modelled here as an alternative to the reference case rather than a mitigation action. As such it differs very little from the reference case (see Figure 19). The reference case considers, relative to the model base, an initial quadrupling of the production capacity of CTL refineries. This becomes most visible by 2030. In the subsequent period the output share of CTL declines again as an increased dependence on crude oil develops.

Figure 19: Comparison of Output Shares under the Biofuels Scenario and Reference Case

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Crude oil CTL GTL Biofuels

Biofuels 1.0% 1.8% 3.1% 2.0% 2.8% 2.0% 2.7%

GTL 6.0% 5.9% 6.2% 3.2% 3.0% 1.8% 1.8%

CTL 25.0% 27.4% 21.5% 44.7% 41.7% 25.4% 24.3%

Crude oil 68.0% 64.9% 69.1% 50.1% 52.4% 70.7% 71.3%

2000 Reference Biofuels Reference Biofuels Reference Biofuels

Base 2015 2030 2050




In the biofuels scenario a slightly greater reliance on biofuels is modelled, but given the small overall contribution of biofuels, even a large increase in biofuels output will do little to alter production and employment at a national level in any significant way. As shown in Table 35 below, a visible effect under the biofuels scenario is an increase in agricultural output relative to the reference case. This comes at the expense of coal mining output. The scenario also allows for slightly higher output from crude oil refineries. Consequently, CTL is the main liquid fuel source being replaced by biofuels, which also explains the decline in coal production levels.

Table 35: Production Levels under the Biofuels Scenario Compared Against the Reference Case

2015 2030 2050

Agric forestry & fish 0.5% 0.4% 0.3%

Coal and lignite -4.8% -2.2% -0.9%

Crude oil and other mining 0.4% 0.0% 0.0%

Crude oil refineries 5.9% 4.4% 0.6%

Coal to liquids -21.8% -6.9% -4.6%

Gas to liquids 5.5% -6.9% -4.6%

Biofuels 68.9% 41.8% 35.1%

Total activity output -0.1% -0.1% 0.0%


Employment is slightly lower under the biofuels scenario than under the reference case. This may seem surprising, given the high labour intensity of the agricultural sector. However, coal that is displaced is equally labour intensive, while the biofuels scenario in particular assumes a greater crude oil share. Since crude oil is largely imported, an increase in demand for crude oil implies that more funds leave South Africa to pay for imports. Only by 2050 does the biofuels scenario lead to positive employment effects, with low-skilled workers being the main beneficiaries here. However, income levels remain negative compared to those under the reference case, even in 2050, despite positive employment. This suggests that the household income gains associated with employment gains among low-skilled workers in 2050 is more than offset by income losses associated with job losses among high skilled workers. This is true across all household groups, with virtually no distributional effects discernible, as shown in Figure 20.

Figure 20: Employment Effects and Per Capita Incomes under the Biofuels Scenario Compared Against the Reference Case


Employment

-0.12%

-0.10%

-0.08%

-0.06%

-0.04%

-0.02%

0.00%

0.02%

0.04%

2015 2030 2050


High-skilled and skilled Semi- and unskilled Total

Per capita income

-0.16%

-0.14%

-0.12%

-0.10%

-0.08%

-0.06%

-0.04%

-0.02%

0.00%

2015 2030 2050




12.4.3 The Economic Impact of a CO2 Emissions Tax

12.4.3.1 Overview

Taxes are ultimately distortionary since they cause a reallocation of resources away from the so-called Pareto efficient allocation. Consequently, in a CGE model, which is based on neoclassical microeconomic principles, we often expect to see welfare losses arising from taxes. However, depending on how revenue from taxes is used, some of these welfare losses may be mitigated. The aim of this analysis is twofold: firstly, to determine the economic effects of various levels of CO2 taxes in terms of GDP, employment and welfare; and, secondly, to consider which revenue recycling scheme would ultimately cause a CO2 tax to be the least distortionary, given various modelling assumptions. In doing so the analysis may shed some light on changes in the energy output mix that may arise in response to the implementation of a CO2 tax. The analysis may also assist policymakers in deciding on an appropriate level of a CO2 tax.

The analysis here converts a given level of a CO2 tax to a comparative tax on coal, crude oil or natural gas used as intermediate inputs in production processes. Table 36 shows the various CO2 tax levels which we model, expressed as a Rand value per ton of CO2 emitted, while the three rows directly below show the associated taxes on coal, crude oil and natural gas. Given the high emissions associated with coal use, the implied tax rates on coal are extremely high; for example, coal prices are likely to rise 25 times if the CO2 tax were R1000. At present levels of between R250 and R750 are being considered in the Use the Market scenario. The table is useful for putting into perspective the kind of economic shock that the implementation of CO2 taxes at these levels implies.


Table 36: Energy Use Tax Equivalent of Rand per Ton CO2 Taxes

Rand / ton tax R 25 R 50 R 75 R 100 R 200 R 300 R 400 R 600 R 800 R 1,000

Coal 59.4% 118.8% 178.2% 237.6% 475.2% 712.8% 950.5% 1425.7% 1900.9% 2376.1%

Crude 2.8% 5.6% 8.4% 11.2% 22.4% 33.5% 44.7% 67.1% 89.4% 111.8%

Equivalent tax on

energy inputs

Gas 0.1% 0.2% 0.3% 0.4% 0.8% 1.2% 1.6% 2.5% 3.3% 4.1%

Source: Author’s calculations based on South African SAM (2000) and information supplied by Andrew

Marquard, Energy Research Centre.

12.4.3.2 Prices, Output and Employment

The aim of an emissions tax is to reduce emissions through incentivising producers to switch away from processes associated with high levels of emissions. The economic welfare losses of rising energy prices therefore have to be weighed against the social welfare gains of reduced emissions. These social welfare gains are not measured in standard CGE models; what we are concerned about here are only the pure economic effects.

As coal, crude oil and natural gas prices rise we expect to see production costs of producers to increase. Ultimately the impact on the energy prices faced by consumers depends on a number of factors. The extent to which highly taxed commodities (coal, crude oil and gas) are still used in the supply of electricity and petroleum products is an important factor, as are production costs in the alternative energy supply sectors (e.g. non-carbon-based) that expand as a result of the CO2 tax. There are also, of course, costs involved in switching. In the CGE model the cost of switching is influenced by the degree of substitutability allowed for in the commodity aggregation function. To illustrate this effects, we consider low (0.5), average (4) and high (10) elasticities of substitution, as well as a special case with perfect substitution (infinite). Figure 21 shows how these various substitution possibilities impact on energy consumer prices (petroleum and electricity). The easier it is to substitute, the lower the price impact will be. All subsequent simulations were done using the more moderate ‘average subsidy’, also used in the modelling of the various mitigation scenarios.

Figure 21: CO2 Tax Simulations: Impact of Different Substitution Possibilities on Energy Prices

0%

20%

40%

60%

80%

100%

120%

140%

160%

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

% change in commodity price

Petroleum (perf subs) Petroleum (high subs)

Petroleum (avg subs) Petroleum (low subs)

-50%

0%

50%

100%

150%

200%

250%

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

% change in commodity price

Elec (perf subs) Elec (high subs)

Elec (avg subs) Elec (low subs)



Revenue from a CO2 tax can be used in a variety of ways by government. As a base scenario we assume government simply uses the additional revenue to reduce its deficit. This increases savings in the economy, which then under the savings-investment balance implies that investments will go up. We call this the ‘non-neutral’ scenario. Later we compare results under various revenue neutral scenario, whereby additional revenue is recycled in the form of production subsidies for nuclear/renewable energy and biofuels, or in the form of food subsidies, general VAT subsidies or income tax subsidies. We also consider an option whereby additional revenue is passed on to poorer households in the form of increased welfare transfers.

Table 37 shows the percentage changes in production levels in various industries under the ‘non-neutral’ scenario, i.e. additional revenue generated by the tax is added to government revenue, which eventually makes its way to the pool of savings via the budget surplus. The table clearly shows the extent to which substitution takes place, particularly in the electricity sector. Looking at the petroleum sector, we note that output from CTL plants is virtually wiped out once the tax reaches R600 and beyond. At this level output from coal-fired electricity plants is down by two-thirds. While a reduction in coal use would, of course, ultimately be the aim of a mitigation action such as this, we also note that overall production levels in other sectors decline as a result of the CO2 tax. This is due to indirect increases in transformed energy prices (electricity and petroleum) and direct increases in primary energy prices (coal, crude oil and gas).

Table 37: CO2 Tax Simulations: Percentage Change in Output Levels (Selected Sectors)

R25 C02 tax R75 C02 tax R100 C02 tax R200 C02 tax R300 C02 tax R600 C02 tax R1000 C02 tax

Agriculture 0.0% -0.2% -0.3% -0.7% -1.3% -2.8% -4.6%

Coal and lignite -11.7% -23.9% -27.5% -35.9% -40.6% -48.5% -53.2%

Petroleum: Crude oil refineries 6.0% 11.3% 12.1% 10.8% 7.0% -6.0% -20.9%

Petroleum: CTL -36.9% -70.2% -78.2% -91.9% -96.0% -98.9% -99.6%

Petroleum: GTL 17.3% 45.8% 58.3% 103.0% 144.6% 267.2% 431.4%

Petroleum: Biofuels -0.4% -6.2% -9.9% -24.6% -36.1% -56.4% -69.0%

Electricity: Coal-fired -4.1% -11.6% -15.2% -28.3% -39.8% -64.7% -81.7%

Electricity: Nuclear 26.9% 91.4% 128.7% 305.6% 510.9% 1126.1% 1717.9%

Electricity: Renewables 28.8% 97.2% 136.5% 322.6% 538.6% 1193.5% 1842.2%

Electricity: Gas turbines 27.2% 90.6% 126.3% 289.9% 471.1% 975.6% 1410.1%

Other sectors 0.0% -0.2% -0.3% -1.0% -1.6% -3.6% -5.7%

Total -0.3% -0.8% -1.0% -1.8% -2.4% -4.1% -5.7%


The shift in output mix in electricity and petroleum looks very different from the output mix predicted by the MARKAL model for the Use the Market scenario (see Figure 22). Of course, the CGE model does not take into account emissions constraints. However, this explains why the CGE model results from the Use the Market scenario showed such large negative effects, as the output mix imposed was not the optimal one (from an economic point of view) under the conditions.

Figure 22: CO2 Tax Simulations: CGE Model Predicted Shifts Output Mix


Petroleum

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

base

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

Output mix

Petro oil Petro CTL Petro GTL Petro Bio

Electricity

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

base

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

Output mix

Elec coal Elec nuc Elec ren Elec gas


Declining production levels already give an indication that GDP is likely to decline as a result of the CO2 tax. As noted, in addition to the non-neutral scenario, we also consider various revenue recycling options. In Figure 23 we compare the GDP effects under a variety of these closures, namely a renewables and nuclear subsidy, a biofuels subsidy, a food subsidy, a general VAT subsidy, an income tax subsidy and a general increase in welfare transfers. Under the non-neutral scenario GDP declines from about 0.5 per cent for a R25 CO2 tax, increasing to 13.9 per cent for a R1000 tax. At the proposed tax of R250 per ton of CO2 GDP is likely to decline by about 5 per cent. Of all the alternative revenue recycling options the food subsidy appears to be the best option, while the two production subsidies yield the worst results. In fact, at low levels of taxation the food subsidy may actually cause GDP to increase marginally. This result is consistent with Van Heerden et al.’s (2006) results for a R35/ton CO2 tax.

32

The actual amount recycled under each scenario is not exactly the same, hence the scenarios are not directly comparable. It may well be that in the absence of a CO2 tax that an income tax subsidy, for example, may have a more favourable outcome than a food subsidy.

32 Van Heerden et al. (2006) only consider a R35 tax, hence it is not possible to say whether our results are

consistent at all levels of taxation.


Figure 23: CO2 Tax Simulations: GDP Effects Under Alternative Revenue Recycling Options

-25.0%

-20.0%

-15.0%

-10.0%

-5.0%

0.0%

5.0%

R25

C02 tax

R75

C02 tax

R100

C02 tax

R200

C02 tax

R300

C02 tax

R600

C02 tax

R1000

C02 tax

Change in GDP

Non-neutral

Renewable and nuclear subsidy

Biofuels subsidy

Food subsidy

VAT subsidy

Income tax

Household transfers


Production subsidies should not be summarily dismissed because they fail to reduce the negative impact of a CO2 tax on GDP. If the aim is to mitigate the rise in energy prices they can be very successful. Figure 11 shows the rise in energy prices (electricity and petroleum) with and without nuclear/renewables or biofuels production subsidies. However, ultimately, because GDP declines more when a production subsidy is introduced suggests that the subsidisation of a less efficient production process is not an economically sensible option.

Figure 24: CO2 Tax Simulations: Impact of Renewables/Nuclear and Biofuels Production Subsidies on Energy Prices

Biofuels subsidy

0%

20%

40%

60%

80%

100%

120%

140%

160%

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

Petroleum price

Electricity price

Petroleum price with no subsidy

Renewables and neuclar subsidy

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

120%

140%

R25 C02 tax

R50 C02 tax

R75 C02 tax

R100 C02 tax

R200 C02 tax

R300 C02 tax

R400 C02 tax

R600 C02 tax

R800 C02 tax

R1000 C02 tax

Petroleum price

Electricity price

Electricity price with no subsidy


Next, we turn to employment effects. The CO2 tax generally has a negative employment impact, especially at high levels of taxation. At the lower levels some of the revenue recycling schemes, in particular the biofuels subsidy, the food subsidy and the general VAT subsidy have a positive effect


on employment . In the case of the biofuels and food subsidies, for example, demand for food and agricultural output increases. Both these sectors are characterised by fairly high labour intensities, hence the positive employment effect. However, at higher levels of taxation employment effects become negative, as the effect of high energy prices overwhelm any employment gains associated with the targeted revenue recycling schemes.


Table 38: CO2 Tax Simulations: Employment and Wage Changes

R25

C02 tax

R75

C02 tax

R100

C02 tax

R200

C02 tax

R300

C02 tax

R600

C02 tax

R1000

C02 tax

Non-neutral closure

Employment changes Semi-skilled -0.4% -1.4% -2.0% -4.4% -6.6% -12.2% -17.4%

Unskilled -1.1% -3.0% -3.8% -6.8% -9.4% -15.8% -22.0%

Wage changes High-skilled -1.6% -4.4% -5.6% -9.8% -13.4% -21.7% -29.5%

Skilled -1.5% -4.1% -5.2% -9.1% -12.5% -20.6% -28.2% Neutral - subsidise renewables and

nuclear


Unskilled -1.0% -2.8% -3.6% -6.2% -8.2% -12.4% -16.1%


Skilled -1.3% -3.7% -4.7% -8.1% -10.7% -16.0% -20.7%

Neutral - subsidise biofuels

Employment changes Semi-skilled 0.0% -0.7% -1.2% -3.2% -5.2% -10.1% -14.7%

Unskilled 0.4% 0.5% 0.4% -0.1% -0.9% -4.0% -7.8%


Skilled -1.0% -3.1% -4.1% -7.8% -11.2% -19.3% -27.2%

Neutral - food subsidy

Employment changes Semi-skilled 0.5% 0.6% 0.4% -0.9% -2.5% -7.4% -12.5%

Unskilled 0.5% 0.8% 0.8% 0.4% -0.5% -4.1% -8.7%

Wage changes High-skilled 0.3% 0.2% -0.1% -1.5% -3.3% -9.1% -15.9%

Skilled 0.4% 0.4% 0.3% -0.8% -2.4% -7.9% -14.3%

Neutral - general VAT subsidy

Employment changes Semi-skilled 0.1% -0.3% -0.6% -2.0% -3.5% -7.5% -11.4%

Unskilled 0.1% -0.2% -0.4% -1.2% -2.2% -5.4% -9.2%

Wage changes High-skilled 0.0% -0.5% -0.8% -2.2% -3.6% -7.9% -12.7%

Skilled 0.0% -0.4% -0.6% -1.8% -3.1% -7.3% -12.2%

Neutral - income tax relief


Unskilled -1.0% -2.6% -3.4% -6.2% -8.6% -14.5% -20.3%


Skilled -1.4% -3.6% -4.7% -8.3% -11.4% -18.9% -26.1%

Neutral - welfare transfers


Unskilled -1.0% -2.6% -3.4% -6.2% -8.6% -14.5% -20.3%


Skilled -1.4% -3.7% -4.7% -8.3% -11.5% -19.0% -26.3%

Finally, we consider some of the welfare effects. Figure 25 shows the results under all the different revenue recycling options. None of the outcomes are necessarily surprising. The food subsidy benefits low-income households most, given the share of poor households’ budget spent on food. Similarly, the welfare transfer scenario also benefits the poor, given that welfare transfers are means tested and targeted at poor people. In contrast, an income tax relief programme benefits mostly high income households, given that they contribute the bulk of income taxes in South Africa.

Appendices to the Technical Report – LTMS

83

Figure 25: CO2 Tax Simulations: Welfare Effects Under Alternative Revenue Recycling Options

Renewable and nuclear subsidy

-35.00

-30.00

-25.00

-20.00

-15.00

-10.00

-5.00

0.00

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C02

tax

R20

0 C02

tax

R30

0 C02

tax

R40

0 C02

tax

R60

0 C02

tax

R80

0 C02

tax

R10

00 C

02 ta

xIndex value

Low income

Lower middle income

Upper middle income

High income

Non-neutral closure

-60.00

-50.00

-40.00

-30.00

-20.00

-10.00

0.00

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C02

tax

R20

0 C02

tax

R30

0 C02

tax

R40

0 C02

tax

R60

0 C02

tax

R80

0 C02

tax

R10

00 C

02 ta

x

Index value

Low income

Lower middle income

Upper middle income

High income

Food subsidy

-35.00

-30.00

-25.00

-20.00

-15.00

-10.00

-5.00

0.00

5.00

10.00

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C02

tax

R20

0 C0

2 ta

xR30

0 C02

tax

R40

0 C02

tax

R60

0 C02

tax

R80

0 C02

tax

R10

00 C

02 ta

x

Index value

Low income

Lower middle income

Upper middle income

High income

General VAT subsidy

-25.0

-20.0

-15.0

-10.0

-5.0

0.0

5.0

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C02

tax

R20

0 C02

tax

R30

0 C02

tax

R40

0 C02

tax

R60

0 C02

tax

R80

0 C02

tax

R10

00 C

02 ta

x

Index value

Low income

Lower middle income

Upper middle income

High income

Income tax relief

-30.0

-25.0

-20.0

-15.0

-10.0

-5.0

0.0

5.0

10.0

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C02

tax

R20

0 C02

tax

R30

0 C02

tax

R40

0 C02

tax

R60

0 C0

2 ta

xR80

0 C02

tax

R10

00 C

02 ta

xIndex value

Low income

Lower middle income

Upper middle income

High income

Welfare transfers

-50.0

-40.0

-30.0

-20.0

-10.0

0.0

10.0

20.0

30.0

R25

C02

tax

R50

C02

tax

R75

C02

tax

R10

0 C0

2 ta

xR20

0 C02

tax

R30

0 C02

tax

R40

0 C02

tax

R60

0 C02

tax

R80

0 C02

tax

R10

00 C

02 ta

x

Index value

Low income

Lower middle income

Upper middle income

High income


LONG-TERM MITIGATION SCENARIOS ENERGY RESEARCH CENTRE

12.4.3.3 Concluding Remarks

This section briefly reviewed some of the price, production, employment, GDP, savings, investments and welfare effects of a proposed tax on CO2 emissions. The analysis can be used to determine a suitable level of taxation that would bring about a positive social outcome as far as emissions reductions are concerned without causing too much harm to the economy at large.

It was shown that any level of taxation induces switching away from CTL and coal-fired electricity plants. Although switching comes with a cost, increasing tax levels act as incentives to switch further away from coal-based processes, which is a desirable outcome from a mitigation point of view. It is clear, however, given the modelling assumptions, that at levels beyond R75 per ton of CO2, and despite using the most efficient of the revenue recycling options available, the economic impact will be negative. At high levels of taxation, therefore, overall economic activity (production) and employment levels are likely to decline. GDP may fall by anything between 2 and 7 per cent for a R250 tax, and by between 9 and 17 per cent as the tax reaches R750 per ton of CO2. It is for policymakers to decide what level of GDP decline is deemed acceptable given the associated mitigation reductions of the tax instrument.



12.5 Additional Tables and Figures

Table 39: Accounts in the SAM

SAM Code Description SAM Code Description

Commodities Activities

cagfield Agric field crops & forestry aagric Agric forestry & fish

caghort Agric horticulture acoal Coal and lignite

caglive Agric livestock fishing agold Gold and uranium ore

ccoal Coal and lignite products aomin Crude oil and other mining

cgold Gold and uranium ore product afood Food products

ccroil Crude oil products abev Beverages and tobacco

comin Other mining products atext Textiles

cfood Food products alwpap Leather Wood and Paper

cbevs Beverages and tobacco apetro Petroleum

ctext Textile products afert Fertilisers

clwpap Leather wood and paper products apest Pesticides

cpetro Petroleum products apharm Pharmaceuticals

cfert Fertilisers aochem Other Chemicals

cpcides Pesticides anonmet Non metallics

cpharm Pharmaceutical products ametals Metals

ochem All other chemical products amach Machinery

cnonmet Non metalic products avehic Vehicles

cmetprod Metal products aomanu Other manufacturing

cmach Machinery aelec Electricity

cvehic Vehicles awater Water

comanu Other manufacturing aconst Construction and Building

celec Electricity atrad Trade and transposrt services

cwater Water aoserv Other services

cconst Construction and building

ctraserv Trade and transport services

coserv Other services

Disaggregation of Petroleum and Electricity Activities

apet_oil Petroleum Crude oil based aelec_coal Electricity Coal based

apet_ctl Petroleum Coal to liquids aelec_nuclear Electricity Nuclear

apet_gtl Petroleum Gas to liquids aelec_renew Electricity Hydro & Renewables

apet_bio Petroleum Biofuels aelec_gas Electricity Gas turbines

Factors of production Households

fgos Gross operating surplus hhlow Low income households

fland Land hhlowmid Lower middle income households

fhskil High-skilled labour hhuppmid Upper middle income households

fskil Skilled labour hhhigh High income households

fsskil Semi-skilled labour

fuskil Unskilled labour



Table 40: Disaggregated Petroleum and Electricity SAM Accounts

Petroleum Electricity

Original

Petroleum

Activity in

SAM Crude oil CTL GTL Biofuels

Original

Electricity

Activity in

SAM Coal-fired Nuclear

Renew-

ables

Gas

turbines

Agric field crops & forestry 350.72 0.04 0.12 0.03 350.54 15.11 13.86 0.84 0.19 0.23

Agric horticulture 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Agric livestock fishing 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Coal and lignite products 5,198.57 0.35 5,155.23 0.27 42.72 4,467.72 4,463.80 1.12 1.26 1.53

Gold and uranium ore product 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Crude oil products 24,306.48 24,296.70 2.86 0.77 6.16 0.00 0.00 0.00 0.00 0.00

Other mining products 2,706.07 1.29 3.68 2,669.43 31.67 4.33 0.43 0.48 0.55 2.86

Food products 1.53 0.88 0.63 0.01 0.00 49.80 45.68 2.76 0.62 0.75

Beverages and tobacco 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Textile products 0.00 0.00 0.00 0.00 0.00 19.56 17.94 1.08 0.24 0.30 Leather wood and paper products 92.35 53.37 38.17 0.61 0.20 105.18 96.46 5.82 1.31 1.59

Petroleum products 4,100.44 1,593.86 2,453.00 39.49 14.09 199.56 0.85 191.09 1.07 6.54

Fertilisers and pesticides 0.00 0.00 0.00 0.00 0.00 12.64 11.59 0.70 0.16 0.19 Pharmaceuticals and other chemicals 1,793.20 1,036.28 741.16 11.93 3.83 102.39 93.91 5.67 1.27 1.55

Non metalic products 299.74 173.22 123.89 1.99 0.64 54.50 49.98 3.02 0.68 0.82

Metal products 246.94 142.70 102.06 1.64 0.53 411.80 377.66 22.79 5.13 6.22

Machinery 892.77 515.92 369.00 5.94 1.91 427.23 391.81 23.64 5.32 6.46

Vehicles 90.64 52.38 37.46 0.60 0.19 158.53 145.39 8.77 1.97 2.40

Other manufacturing 44.28 25.59 18.30 0.29 0.09 1,981.02 1,816.78 109.63 24.67 29.94

Electricity 825.03 476.78 341.00 5.49 1.76 1,911.61 1,753.13 105.79 23.80 28.89

Water 208.08 120.25 86.00 1.38 0.44 149.12 136.76 8.25 1.86 2.25

Construction and building 0.00 0.00 0.00 0.00 0.00 2,462.63 2,258.47 136.29 30.66 37.21

Trade and transport services 2,961.21 1,711.26 1,223.92 19.70 6.33 633.32 580.81 35.05 7.89 9.57

Other services 2,111.35 1,220.13 872.66 14.05 4.51 1,722.59 1,579.78 95.33 21.45 26.03

Gross operating surplus 8,895.22 6,037.16 2,228.99 537.22 91.85 14,197.85 13,187.90 727.86 124.11 157.97

Land 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

High-skilled labour 1,812.37 1,248.95 445.33 107.33 10.77 1,849.04 1,677.36 123.29 44.41 3.99

Skilled labour 452.40 344.91 84.23 20.30 2.96 598.37 566.58 20.73 10.15 0.91

Semi-skilled labour 846.76 560.01 223.11 53.77 9.87 1,312.60 1,274.79 18.02 18.32 1.47

Unskilled labour 47.66 20.48 20.38 4.91 1.89 173.99 167.63 2.67 3.44 0.25

Production rebates -491.39 -333.51 -123.13 -29.68 -5.07 -33.22 -30.85 -1.70 -0.29 -0.37

Production taxes 106.36 72.18 26.65 6.42 1.10 293.43 272.56 15.04 2.57 3.26

Production subsidies -3.80 -2.58 -0.95 -0.23 -0.04 -5.39 -5.00 -0.28 -0.05 -0.06

TOTAL 57,894.98 39,368.59 14,473.75 3,473.70 578.95 33,275.30 30,946.03 1,663.77 332.75 332.75

Note: Figures and millions of Rands, year 2000 prices.

Source: Control totals from SAM (2000); disaggregation based on author’s calculations.



Table 41: Industry-level Employment Levels by Skill

High-skilled

labour

Skilled

labour

Semi-skilled

labour

Unskilled

labour Total

Agric forestry & fish 26,286 20,963 399,864 545,856 992,969

Coal and lignite 6,091 2,580 41,269 5,762 55,702

Gold and uranium ore 23,178 31,269 230,744 28,181 313,372

Crude oil and other mining 9,238 16,541 93,869 28,398 148,046

Food products 25,274 25,363 107,176 61,288 219,101

Beverages and tobacco 10,152 15,124 33,322 10,895 69,493

Textiles 21,720 32,823 245,570 28,896 329,009

Leather Wood and Paper 26,087 24,546 134,899 36,034 221,566

Petroleum: Crude oil based 3,634 4,868 5,922 339 14,763

Petroleum: Coal to liquids 1,357 1,245 2,472 353 5,428

Petroleum: Gas to liquids 326 299 593 85 1,303

Petroleum: Biofuels 33 43 109 33 217

Fertilisers and pesticides 1,187 630 1,479 910 4,206

Pharmaceuticals and other chemicals 15,743 7,836 21,352 9,341 54,272

Non-metals 16,194 9,464 90,147 28,047 143,852

Metals 27,405 12,301 154,328 21,768 215,802

Machinery 13,095 7,443 26,863 8,873 56,274

Vehicles 13,674 8,673 44,806 9,517 76,670

Other manufacturing 18,703 12,373 60,513 27,108 118,697

Electricity: Coal based 11,498 8,499 21,997 7,999 49,993

Electricity: Nuclear 836 307 307 126 1,577

Electricity: Hydroelectricity & Renewables 349 175 363 188 1,075

Electricity: Gas turbines 25 12 23 11 71

Water 5,057 4,092 13,655 2,945 25,749

Construction and Building 21,972 18,631 473,904 117,962 632,469

Trade and transport services 371,009 1,062,879 749,030 609,437 2,792,355

Other services 1,208,900 1,092,047 442,353 1,769,222 4,512,522

Total 1,879,022 2,421,027 3,396,929 3,359,573 11,056,551



Table 42: Simulation Parameters for Combined Scenarios: Structural Shifts in Energy Output Mix

Reference Case - GWC

base simref05 simref10 simref15 simref20 simref30 simref40 simref50

aelec_coal 92.9% 93.9% 94.6% 94.7% 92.2% 83.4% 81.6% 81.3%

aelec_nuclear 5.1% 5.9% 5.2% 4.9% 6.4% 7.8% 12.1% 14.0%

aelec_renew 0.9% 0.01% 0.01% 0.2% 1.1% 8.6% 6.2% 3.7%

aelec_gas 1.1% 0.2% 0.2% 0.3% 0.3% 0.2% 0.1% 1.0%

base simref05 simref10 simref15 simref20 simref30 simref40 simref50

apet_oil 67.9% 69.5% 70.9% 69.4% 66.8% 68.2% 71.3% 76.6%

apet_ctl 25.1% 23.7% 22.0% 23.2% 26.4% 26.6% 24.4% 19.5%

apet_gtl 6.0% 6.8% 6.3% 5.7% 4.9% 3.2% 2.3% 1.8%

apet_bio 1.0% 0.04% 0.8% 1.7% 1.9% 1.9% 2.0% 2.0%

Mitigation - Start Now (initial wedges)

base simshd05 simshd10 simshd15 simshd20 simshd30 simshd40 simshd50

aelec_coal 92.9% 93.9% 94.0% 87.3% 73.5% 51.3% 47.6% 45.7%

aelec_nuclear 5.1% 5.9% 5.4% 7.5% 13.5% 24.2% 26.2% 26.6%

aelec_renew 0.9% 0.2% 0.6% 5.2% 12.8% 23.7% 26.0% 27.7%

aelec_gas 1.1% 0.01% 0.01% 0.01% 0.27% 0.8% 0.2% 0.1%

base simshd05 simshd10 simshd15 simshd20 simshd30 simshd40 simshd50

apet_oil 67.9% 69.5% 69.9% 66.5% 61.3% 57.6% 60.1% 65.8%

apet_ctl 25.1% 23.6% 22.4% 24.6% 29.9% 34.7% 33.2% 28.0%

apet_gtl 6.0% 6.8% 6.4% 6.0% 5.5% 4.2% 3.2% 2.6%

apet_bio 1.0% 0.1% 1.3% 2.8% 3.3% 3.6% 3.5% 3.6%

Mitigation - Scale Up (extended wedges)

base simcan05 simcan10 simcan15 simcan20 simcan30 simcan40 simcan50

aelec_coal 92.9% 93.9% 93.0% 86.3% 73.8% 51.4% 31.8% 17.2%

aelec_nuclear 5.1% 5.9% 5.4% 7.5% 13.4% 24.5% 34.3% 41.5%

aelec_renew 0.9% 0.2% 1.6% 6.2% 12.8% 24.1% 33.9% 41.3%

aelec_gas 1.1% 0.01% 0.01% 0.01% 0.01% 0.02% 0.01% 0.01%

base simcan05 simcan10 simcan15 simcan20 simcan30 simcan40 simcan50

apet_oil 67.9% 69.4% 69.8% 67.5% 64.4% 64.3% 63.8% 70.5%

apet_ctl 25.1% 23.6% 22.3% 23.4% 26.9% 28.3% 29.4% 23.7%

apet_gtl 6.0% 6.8% 6.4% 5.8% 5.0% 3.4% 2.8% 2.2%

apet_bio 1.0% 0.1% 1.5% 3.3% 3.8% 4.0% 3.9% 3.6%

Mitigation - Use the Market (economic instruments)

base simcld05 simcld10 simcld15 simcld20 simcld30 simcld40 simcld50

aelec_coal 92.9% 94.0% 90.8% 63.7% 22.0% 2.0% 0.0% 0.0%

aelec_nuclear 5.1% 5.7% 5.4% 18.1% 43.9% 44.9% 27.4% 25.6%

aelec_renew 0.9% 0.3% 3.9% 18.2% 34.1% 53.1% 72.6% 74.4%

aelec_gas 1.1% 0.01% 0.01% 0.01% 0.01% 0.00% 0.00% 0.02%

base simcld05 simcld10 simcld15 simcld20 simcld30 simcld40 simcld50

apet_oil 67.9% 72.3% 72.5% 72.1% 81.1% 94.5% 95.6% 96.1%

apet_ctl 25.1% 24.0% 21.7% 20.5% 12.2% 0.2% 0.0% 0.0%

apet_gtl 6.0% 3.7% 5.1% 5.8% 5.0% 3.7% 2.8% 2.3%

apet_bio 1.0% 0.0% 0.7% 1.6% 1.7% 1.6% 1.6% 1.6%

Note: In the results section we only report on results up to 2015, i.e. simshd15, simcan15 and simcld15.

Source: MARKAL model results


89

LONG-TERM MITIGATION SCENARIOS

ENERGY RESEARCH CENTRE

Table 43: Simulation Parameters for Combined Scenarios: Energy Efficiency

Sta

rt N

ow

Electricity

Coal

Petroleum

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

Mining

0.0%

-5.7%-23.7%

-23.8%

-23.7%

-26.3%

-29.8%

0.0%

0.0% 0.0%

0.0%

0.0%

-2.4%

-4.8%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Industry

0.0%

-1.4%

-8.0%

-8.2%

-7.2%

-7.9%

-7.5%

0.0%

0.0% -16.2%

-19.7%

-26.0%

-41.9%

-33.7%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Commerce -0.4%

-3.5%

-8.5%-12.0%

-15.6%

-15.9%

-16.6%

-0.3%

-6.1% -16.8%

-24.7%

-36.0%

-37.7%

-39.7%

0.0%

0.1%

0.3%

0.5%

0.7%

0.7%

0.9%

Transport

9.5%

30.1%

36.4%

38.1%

43.3%

43.3%

42.5%

0.0%

0.0% 0.0%

0.0%

0.0%

0.0%

0.0%

-3.7%-12.5%

-16.6%

-17.7%

-19.9%

-22.6%

-25.1%

Int input

1.8%

2.2%

-2.6%

-2.7%

-1.6%

-2.4%

-3.0%

0.0%

-0.1% -11.1%

-13.5%

-17.8%

-28.6%

-23.1%

-8.0%

-11.4%

-13.0%

-13.4%

-14.3%

-15.3%

-16.3%

Tot supply

1.1%

1.4%

-1.6%

-1.7%

-1.0%

-1.4%

-1.8%

0.0%

0.0% -7.1%

-8.7%

-11.5%

-18.4%

-14.8%

-4.3%

-6.2%

-7.1%

-7.3%

-7.7%

-8.3%

-8.8%

Sca

le U

p

Electricity

Coal

Petroleum

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

Mining

0.0%

-5.7%-23.7%

-23.8%

-23.7%

-26.3%

-29.8%

0.0%

0.0% 0.0%

0.0%

0.0%

-2.4%

-4.8%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Industry

0.0%

-1.3%

-8.0%

-8.2%

-7.2%

-7.9%

-7.5%

0.0%

0.0% -16.2%

-19.7%

-26.0%

-41.9%

-33.7%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Commerce -0.4%

-3.5%

-8.5%-12.0%

-15.6%

-15.9%

-16.6%

-0.3%

-6.1% -16.8%

-24.7%

-36.0%

-37.7%

-39.7%

0.0%

0.1%

0.3%

0.5%

0.7%

0.8%

0.9%

Transport

0.2%

3.5%

7.6%

7.2%

7.2%

35.1%

51.9%

0.0%

0.0% 0.0%

0.0%

0.0%

0.0%

0.0%

-0.1%

-1.1%

-2.2%

-1.5%

-1.2%-10.1%

-15.4%

Int input

0.0%

-1.1%

-6.4%

-6.9%

-6.8%

-2.1%

0.6%

0.0%

-0.1% -11.1%

-13.5%

-17.8%

-28.6%

-23.1%

-6.5%

-6.9%

-7.4%

-7.0%

-6.9%

-10.4%

-12.5%

Tot supply

0.0%

-0.7%

-3.9%

-4.2%

-4.2%

-1.3%

0.4%

0.0%

0.0% -7.1%

-8.7%

-11.5%

-18.4%

-14.8%

-3.5%

-3.8%

-4.0%

-3.8%

-3.7%

-5.6%

-6.8%


90

LONG-TERM MITIGATION SCENARIOS


Use

the M

ark

et

Electricity

Coal

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

Mining

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

-66.5%

Industry

0.0%

-1.3%

-8.0%

-8.2%

-7.2%

-7.9%

-7.5%

0.0%

0.0%

0.0%

0.0%

0.0%

-0.2%

-45.2%

Commerce

-0.4%

-3.6%

-8.7%

-12.3%

-15.5%

-14.8%

-15.0%

-0.1%

-1.2%

-2.7%

-8.8%

-35.7%

-44.2%

-45.5%

Transport

0.1%

3.1%

11.3%

25.9%

60.5%

70.3%

77.6%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Int input

0.0%

-0.3%

-1.9%

0.5%

7.3%

9.0%

10.5%

0.0%

0.0%

0.0%

-0.1%

-0.3%

-0.5%

-31.9%

Tot supply

0.0%

-0.2%

-1.2%

0.3%

4.5%

5.5%

6.4%

0.0%

0.0%

0.0%

-0.1%

-0.2%

-0.3%

-20.5%

Petroleum

Gas switching

2005

2010

2015

2020

2030

2040

2050

2005

2010

2015

2020

2030

2040

2050

Mining

179.5%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

-0.1%

-0.1%

1760.5%

Industry

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

4.2%

1075.2%

Commerce

0.0%

-0.1%

-0.3%

-0.4%

-0.4%

-0.5%

-0.5%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Transport

-0.1%

-1.0%

-3.7%

-8.2%

-17.9%

-20.9%

-23.4%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Int input

-0.1%

-6.9%

-8.0%

-9.8%

-13.6%

-14.8%

-15.8%

0.0%

0.0%

0.0%

0.0%

0.0%

3.7%

966.6%

Tot supply

0.0%

-3.8%

-4.3%

-5.3%

-7.4%

-8.0%

-8.6%

0.0%

0.0%

0.0%

0.0%

0.0%

1.3%

342.8%

Note:

In the results section we only report on results up to 2015, i.e. simshd15, simcan15 and simcld15.


Appendices to the Technical Report - LTMS 91


Figure 26: Simulation Parameters for Combined Scenarios: Investment Costs

-15%

-10%

-5%

0%

5%

10%

15%

20%

25%

2005 2010 2015 2020 2030 2040 2050


Should do Can do Could do


Table 44: Simulation Parameters for Combined Scenarios: CO2 Emissions Taxes

Year

Tax: Rands per ton

of CO2

Effective tax on

crude oil

Effective tax on

other mining

commodities (gas) Effective tax on coal

2005 R 0 0.0% 0.0% 0.0%

2010 R 279 31.2% 1.2% 663.9%

2015 R 353 39.5% 1.5% 838.5%

2020 R 427 47.7% 1.8% 1013.4%

2030 R 574 64.1% 2.4% 1362.7%

2040 R 721 80.6% 3.0% 1712.0%

2050 R 750 83.8% 3.1% 1781.9%




References

References for energy emissions

From Technical report An, F & Sauer, A 2004. Comparison of Passenger Vehicle Fuel Economy and GHG Emission Standards Around the World. Pew Center on Global Climate Change. December 2004. www.pewclimate.org.

Azar, C & Schneider, S H 2002. Are the economic costs of stabilising the atmosphere prohibitive? Ecological

Economics 42: 73-80.

Clark, A & Spalding-Fecher, R 1999. Financial protocol for South Africa's climate change mitigation assessment. Cape Town, Energy & Development Research Centre, University of Cape Town.

DWAF 2001. Thukela water project feasibility study: Main feasibility report. PB V000-00-9700. Prepared by BKS Incorporated Water Division. November 2001.

DWAF 2006. Vaal River system: large bulk water supply reconciliation strategy: First stage reconciliation strategy. P RSA C000/00/4405/07. December 2006.

EPRI & NRDC (Electric Power Research Institute & Natural Resources Defense Council) 2007. Environmental assessment of plug-in hybrid electric vehicles. Volume 1: Nationwide greenhouse gas emissions. www.epri.com (accessed 24 July 2007).

ESD (Energy for Sustainable Development) 2004. Special issue on renewable energy policies in Europe. Bangalore.

Fraser, R 2007. Personal communication. Response to long term mitigation scenario project. 27 June 2007.

Gan, L, Eskelanda, G S & Kolshusa, H H 2007. Green electricity market development: Lessons from Europe and the US. Energy Policy 35 (1): 144-155.

Halsnaes, K, Callaway, J M & Meyer, H J 1998a. Economics of greenhouse gas limitations: Methodological guidelines. Roskilde, Denmark, UNEP Collaborating Centre on Energy and Environment.

Halsnaes, K, Callaway, J M & Meyer, H J 1998b. Economics of greenhouse gas limitations: Technical guidelines. Draft February '98. Roskilde, Denmark, UNEP Collaborating Centre on Energy and Environment.

Howells, M I & Solomon, M 2002. An optimal energy and greenhouse gas mitigation path for South Africa in the short to medium term. Journal of Energy in Southern Africa 13 (4): 123-129.

IPCC (Intergovernmental Panel on Climate Change) 1996. Revised 1996 guidelines for national greenhouse gas inventories. Paris, Organisation for Economic Co-operation and Development.

IPCC (Intergovernmental Panel on Climate Change) 2001. Climate Change 2001: Mitigation. Contribution of WG III to the Third Assessment Report of the IPCC. Cambridge, Cambridge University Press for Intergovernmental Panel on Climate Change.

IPCC (Intergovernmental Panel on Climate Change) 2005. Special report on carbon dioxide capture and storage. Geneva.

IPCC 2007. Climate Change 2007: Mitigation of Climate Change. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva.

Johansson, B & Ahman, M 2002. A comparison of technologies for carbon-neutral passenger transport. Transportation Research Part D 7: 175–196.

Kahane, A 2000. How to change the world: Lessons for entrepreneurs from activists. Draft. Hamilton, Centre for Generative Leadership.

Mayet, M A G 1993. Domestic waste generation in the urban core of the Durban Functional Region. MSc thesis. University of Natal.

Midttun, A & Koefoed, A L 2003. Greening of electricity in Europe: challenges and developments. Energy Policy 31: 677-687.

Morris, G 2002. Criteria for certification of electricity generation plant for participation in a green electricity market in South Africa. Cape Town, AGAMA Energy.

Nordhaus, W D 1993. Rolling the 'DICE': An optimal transition path for controlling greenhouse gases. Resource and

Energy Economics 15: 27-50.

OECD 2000. Ancillary benefits and costs of greenhouse gas mitigation. IPCC co-sponsored workshop, Washington, Organisation for Economic Co-operation and Development.

Rabe, B G 2006. Race to the top: The expanding role of US state renewable portfolio standards. Arlington, Pew Center on Global Climate Change. www.pewclimate.org.

Rademeyer, J 2007. Personal communication. Vaal River system planning. 3 August 2007.

Scholes, R J, Van der Merwe, M R, Kruger, A J & Crookes, D 2000. Mitigation of climate change through land use practices. South African Country Studies on Climate Change. Pretoria, Council for Scientific and Industrial Research.

Shell (Shell International) 2001. Energy needs, choices and possibilities: Scenarios to 2050. London.



Sims, R E H, Rogner, H-H & Gregory, K 2003. Carbon emission and mitigation cost comparisons between fossil fuel, nuclear and renewable energy resources for electricity generation. Energy Policy 31: 1315-1326.

Stern Review (Stern review team) 2006. The economics of climate change. London, Treasury. http://www.hm-treasury.gov.uk/independent_reviews/stern_review_economics_climate_change/sternreview_index.cfm (accessed October 2006).

UK (United Kingdom) 2001. The renewables obligation / NFFO. London, Department of Industry and Trade. www.dti.gov.uk/renewable/nffo.html.

Van der Heijden, K 1996. Scenarios: the art of strategic conversation. Chichester, Wiley.

Van der Merwe, M R & Scholes, R J 1998. South African Greenhouse Gas Emissions Inventory for the years 1990 and 1994. Pretoria, National Committee on Climate Change.

van Rooyen, J 2007. Personal communication. Vaal River system planning. 13 July 2007.

Weitzman, M 1974. Prices vs Quantity. The Review of Economic Studies 41 (4): 477-491.

Wellington, F, Bradley, R, Childs, B, Rigdon, C & Pershing, J 2007. Scaling up: Global technology deployment to stabilise emissions. Washington, World Resources Institute.

Winkler, H 2005. Renewable energy policy in South Africa: Policy options for renewable electricity. Energy Policy 33 (1): 27-38.

Wiser, R, Porter, K & Bolinger, M 2002. Comparing state portfolio standards and system-benefits charges under restructuring. Unpublished table. Berkeley, Lawrence Berkeley National Laboratory.

References in appendices 25 Degrees 2006. Africa’s first bio-ethanol plant. 25 Degrees in Africa 1 (1): pp. 28.

AEJ 2006a. Botswana's king coal. African Energy Journal 8 (5): pp. 6-9.

AEJ 2006b. Energy imports on the agenda. African Energy Journal 8 (5): pp. 10-12.

AGAMA 2003. Employment Potential of Renewable Energy In South Africa. Constantia, AGAMA Energy (Pty) Ltd. 14 November 2003.

Armington, P S 1969. A theory of demand for products distinguished by place of production. IMF Staff Papers Vol. 16.

AsgiSA (Republic of South Africa) 2006. Accelerated and Shared Growth Initiative for South Africa (AsgiSA). Pretoria. http://www.info.gov.za/asgisa/asgisa.htm, accessed December 2006.

Banks, D & Schäffler, J L 2005. The potential contribution of renewable energy in South Africa. Prepared for the Sustainable Energy & Climate Change Partnership, Earthlife Africa. Pretoria, RAPS Consulting & Nano Energy.

Chikkatur, A P & Sagar, A D 2006. Towards a better technology policy for coal-based power in India. Proceedings of the 1st National Conference on Advances in Energy Research.

CIC (CIC Energy Corporation) 2006. CIC Mmamabula Energy Project Update. December 7. Road Town, Tortola, British Virgin Islands. http://www.cicenergycorp.com/investor_relations/in_the_news/index.php?&content_id=102.

De Vries, B, Van Vuuren, D & Hoogwijk, M M 2007. Renewable energy sources: Their global potential for the first-half of the 21st century at a global level: An integrated approach. Energy Policy 35 (4): 2590–2610.

Dervis, K, de Melo, J & Robinson, S 1982. General Equilibrium Models for Development Policy. New York, Cambridge University Press.

Devarajan, S 2007. Notes on Dynamics in CGE Models. World Bank. 25 January 2007,

Devarajan, S, Lewis, J D & Robinson, S 1994. Getting the Model Right: The General Equilibrium Approach to Adjustment Policy. Mimeo.

DME (Department of Minerals and Energy) 2003. Integrated energy plan for the Republic of South Africa. Pretoria. www.dme.gov.za.

DME (Department of Minerals and Energy) 2004. Economic and financial calculations modelling for the renewable energy white paper and strategy formulation. Report No. 2.3.4 - 19, prepared by Conningarth Economists. Pretoria, DME - Capacity Building in Energy Efficiency and Renewable Energy project.

DSW & PCF (Durban Solid Waste and Prototype Carbon Fund) 2006. Durban Landfill-gas-to-electricity project – Mariannhill and La Mercy Landfills, CDM Project 0545. http://cdm.unfccc.int/Projects/DB/TUEV-SUED1154520464.04/view.html, accessed 17 November 2006.

EDRC (Energy & Development Research Centre) 2003. Policies and measures for renewable energy and energy efficiency in South Africa. Prepared for the Sustainable Energy & Climate Change Partnership. Cape Town, EDRC, University of Cape Town.

EIA (Energy Information Administration) 2006. Annual Energy Outlook 2006 with Projections to 2030. Washington DC, US Department of Energy. http://www.eia.doe.gov/oiaf/aeo/gas.html.

Energy Innovations 1997. Energy innovations: a prosperous path to a clean environment. Washington, D.C., Alliance to Save Energy, American Council for an Energy Efficient Economy, Natural Resources Defense Council, Tellus Institute, Union of Concerned Scientists.



Engelbrecht, A, Golding, A, Hietkamp, S & Scholes, R J 2004. The potential for sequestration of carbon dioxide in South Africa. Report for the Department of Minerals & Energy. Pretoria, Council for Scientific and Industrial Research.

ENS (Environmental News Service) 2004. Durban landfill gas project to generate power, Cool the Climate. http://www.ens-newswire.com/ens/jun2004/2004-06-16-02.asp, accessed 17 November 2006.

Fisher, J & Grubb, M 1997. The use of economic models in climate change policy analysis. 5. Royal Institute of International Affairs. October 1997.

Hoogeveen, J G & Özler, B 2004. Not Separate, Not Equal. Poverty and Inequality in Post-Apartheid South Africa. World Bank.

IEA (International Energy Agency) 2003. Renewables for power generation: Status & prospects. Paris, OECD/IEA. ww.iea.org/textbase/nppdf/free/2000/renewpower_2003.pdf.

IEA (International Energy Agency) 2006. World Energy Outlook. Paris, IEA.

IEA & OECD (International Energy Agency, Organisation for Economic Cooperation and Development) 2000. Experience curves for energy technology policy. Paris.

IEA & OECD 2006. Energy technology perspectives: Scenarios and strategies to 2050. Paris, International Energy Agency & Organisation for Economic Co-operation and Development.

IPCC (Intergovernmental Panel on Climate Change) 1996. Revised 1996 guidelines for national greenhouse gas inventories. Paris, Organisation for Economic Co-operation and Development.

IPCC (Intergovernmental Panel on Climate Change) 2000. Special Report on Emissions Scenarios. A special report of Working Group III of the IPCC. Cambridge University Press.

IPCC (Intergovernmental Panel on Climate Change) 2001. Climate Change 2001: Mitigation. Contribution of WG III to the Third Assessment Report of the IPCC. Cambridge, Cambridge University Press for Intergovernmental Panel on Climate Change.

IPCC (Intergovernmental Panel on Climate Change) 2005. Special report on carbon dioxide capture and storage. Technical summary. Geneva.

Junginger, M, Faaij, A & Turkenburg, W 2004. Global experience curves for wind farms. Energy Policy 32: 133-150.

Kilkenny, M 1991. Computable General Equilibrium Modeling of Agricultural Policies: Documentation of the 30-Sector FPGE GAMS Model of the United States. USDA ERS Staff Report AGES 9125.

King, B J 1985. What is a SAM? in G Pyatt and J Round (Eds). Social Accounting Matrices: A Basis for Planning. Washington D.C., World Bank.

Kwon, T-H 2006. The determinants of the changes in car fuel efficiency in Great Britain (1978-2000). Energy Policy 34: 2405-2412.

Laitner, J A 2002. LBD technology data. Spreadsheet. Washington DC, US Environmental Protection Agency.

Laitner, J A, DeCanio, S J & Peters, I 2000. Incorporating Behavioral, Social and Organizational Phenomena in the Assessment of Climate Change Mitigation Options. Karlsruhe, Germany. March 2000.

Le Roux, H 2006. Independent Power: South Africa’s first gas-fired cogen plant under construction. Engineering

News 3-9 November.

Lloyd, P J D & Trikam, A 2004. The determination of emission factors for South African power stations. Eskom Contract 1RE-000046. Cape Town, Energy Research Centre, University of Cape Town.

Löfgren, H, Harris, R L & Robinson, S 2001. A Standard Computable General Equilibrium (CGE) Model in GAMS. International Food Policy Research Institute (IFPRI): Trade and Macroeconomics Division Discussion Paper 75.

Mwakasonda, S & Winkler, H 2005. Carbon capture and storage in South Africa. Chapter 6 in. in R Bradley, K Baumert and J Pershing (Eds). Growing in the greenhouse: Protecting the climate by putting development first. Washington DC, World Resources Institute: 94-109.

National Treasury 2005. Medium term budget policy statement: Chapter 1 Accelerated and shared growth. Pretoria, National Treasury. http://www.polity.org.za/pdf/Chapter%201.pdf.

NEA, IEA & OECD (Nuclear Energy Agency, International Energy Agency and Organisation for Economic Co-operation and Development) 2005. Projected costs of generating electricity. Paris.

Nemet, G F 2006. Beyond the learning curve: factors influencing cost reductions in photovoltaics. Energy Policy 34: 3218–3232.

NER (National Electricity Regulator) 2004. National Integrated Resource Plan 2 (NIRP2) 2003/4. Pretoria, National Electricity Regulator. www.ner.org.za.

NREL (National Renewable Energy Laboratory) 1999. The potential for low-cost Concentrating Solar Power systems. Report NREL/CP-550-26649. Golden, Colorado, NREL.

Pacala, S & Socolow, R H 2004. Stabilization wedges: Solving the climate problem for the next 50 years with current technologies. Science 305 (13 August): 968-972.

Papineau, M 2006. An economic perspective on experience curves and dynamic economies in renewable energy technologies. Energy Policy 34 (4): 422-432.



Pauw, K 2005. Forming Representative Household and Factor Groups for a South African SAM. 2005:2. Elsenburg, PROVIDE Project. www.elsenburg.com/provide.

Pauw, K 2006. Personal communication. Economist, University of Cape Town.

Pauw, K, McDonald, S & Punt, C 2007. Agricultural Efficiency and Welfare in South Africa. Development Southern

Africa Forthcoming.

Philibert, C 2005. The present and future use of solar thermal energy as a primary source of energy. The InterAcademy Council. www.iea.org/textbase/papers/2005/solarthermal.pdf.

PROVIDE 2005a. Compiling a Social Accounting Matrix for South Africa: 2000. Forthcoming. Elsenburg, PROVIDE Project.

PROVIDE 2005b. The PROVIDE Project Standard Computable General Equilibrium Model. Version 2. 2005:3. Elsenburg, PROVIDE Project. www.elsenburg.com/provide.

PROVIDE 2006. Compiling National, Multiregional and Regional Social Accounting Matrices for South Africa. 2006:1. Elsenburg, PROVIDE Project. www.elsenburg.com/provide.

Pyatt, G 1998. A SAM Approach to Modelling. Journal of Policy Modelling 10: 327-352.

Repetto, R & Austin, D 1997. The costs of climate protection: A guide for the perplexed. Washington, World Resources Institute.

Robinson, S, Kilkenny, M & Hanson, K 1990. USDA/ERS Computable General Equilibrium Model of the United States. USDA ERS Staff Report AGES 9049.

RSA (Republic of South Africa) 2004. South Africa: Initial National Communication under the United Nations Framework Convention on Climate Change. Submitted at COP-9. Pretoria. unfccc.int/resource/docs/natc/zafnc01.pdf.

Sargent & Lundy (Sargent & Lundy LLC Consulting Group) 2003. Assessment of parabolic trough and power tower solar technology cost and performance forecasts. NREL/SR-550-34440. Chicago. www.nrel.gov/solar/parabolic_trough.html.


SSA 2002. Labour Force Survey September 2000. Pretoria, Statistics South Africa.

SSA 2003. Final supply and use tables, 2000: an input-output framework. Pretoria, Statistics South Africa.

Thurlow, J 2004. A Dynamic Computable General Equilibrium (CGE) Model for South Africa: Extending the Static IFPRI Model. No. 1-2004. Trade and Industrial Policy Strategy. February 2004.

Turrentine, T & Kurani, K in press 2006. Car buyers and fuel economy? Energy Policy.

UNEP (United Nations Environment Programme) 2006. Changing climate: The role of renewable energy in a carbon-constrained world. Roskilde.

Vessia, Ø 2006. Long-term CO2 mitigation scenarios for South Africa 2001 - 2050. Masters Thesis. Department of Electrical Engineering, Norwegian University of Science and Technology & Energy Research Centre, University of Cape Town.

Von Blottnitz, H & Curran, M A in press. A review of assessments conducted on bio-ethanol as a transportation fuel from a net energy, greenhouse gas, and environmental life cycle perspective. Journal of Cleaner Production.

Winkler, H (Ed) 2006. Energy policies for sustainable development in South Africa: Options for the future. ISBN: 0-

620-36294-4. Contributors: O Davidson, H Winkler, A Kenny, G Prasad, D Sparks, M Howells, T Alfstad, S Mwakasonda, B Cowan and E Visagie. Cape Town, Energy Research Centre. www.erc.uct.ac.za/publications/Energy%20policies%20for%20SD.pdf.

World Bank 1999. Cost reduction study for solar thermal power plants. Prepared by Enermodal Engineering Ltd. Washington DC, World Bank.

World Bank 2006. Assessment of the World Bank Group / GEF strategy for market development of concentrating solar power. Washington DC, GEF/ WorldBank. http://siteresources.worldbank.org/GLOBALENVIRONMENTFACILITYGEFOPERATIONS/Resources/Publications-Presentations/SolarThermal.pdf.

References for industrial process emissions Australia (2004): Australian Greenhouse Gas Office. Australian methodology for the Estimation of Greenhouse Gas Emissions and Sinks: Industrial Processes. Canberra.

Borland J., Matsabu M., Gontshi N. and Wiechers H. (2000): SA Country Studies. Mitigation Options Project: Industrial Processes. Pretoria

Cluett (2006): Personal communication A Cluett via the Association of Cementitious Materials Producers (ACMP)

De Wit, M.P. (2005): Coal Mining and Carbon Constraints. Coaltech 2020 Final report on task 6.1. De Wit Sustainable Options (Pty) Ltd, Cape Town



DME (2003): A review of the Dolomite and Limestone Industry in SA. Directorate Mineral Economics Report R43/2003.

Dong Woon Noh (2006): GHG Emission Reduction Potential in Industrial Sector Cement industry in Korea Economic and Environmental Modeling workshop, Beijing

Engelbrecht A., Golding A., Hietkamp S., Scholes R.J.: The Potential for Sequestration of Carbon Dioxide in South Africa. CSIR Report 86DD / HT339. Pretoria.

Freund, P. and Davison, J. (n.d.): General overview of costs. IPPC workshop on carbon dioxide capture and storage.

Goede (2007): Personal communication F.Goede and H. van der Walt, Sasol Ltd.

Human (2007): Personal communication C Human, Anglogold Ashanti.

IAI (2006): International Aluminium Institute. The Aluminium Sector Greenhouse Gas protocol October 2006

IEA (International Energy Agency) (2004): Low Emission Fuels - the impact of CO2 Capture and Storage on Selected Pathways. Report produced for the IEA Greenhouse Gas R&D programme. Vienna.

IPPC (2007): Fourth Assessment Report, Working Group III. New York

Liebenberg (2007): Personal communication J Liebenberg, Sasol Ltd.

McCollum,and Ogden, J. (2006) : Techno-Economic Models for Carbon Dioxide Compression, Transport and Storage & Correlations for Estimating Carbon Dioxide Density and Viscosity. Institute of Transportation Studies, University of California, Davis, Report UCD-ITS-RR-065-14

Marais (2007) Personal communication E. Marais, Sasol Synfuels (Pty) Ltd

SA DNA (2007) South African Designated National Authority www.dme.gov.za Accessed April 2007

Sasol (2003) Sasol Sustainability Report 2000-2002. Johannesburg

South Africa (2000): Initial National Communication under the United Nations Framework. Department of Environmental Affairs and Tourism, Pretoria.

Scholes, R.J. and van der Merwe, M (1993):. South African GHG inventory 1990. CSIR, Pretoria.

US EPA (2006): Global mitigation of non-CO2 greenhouse gases. Report EPA 430-R-06-005. Research Triangle Park

US EPA (2002): Air Pollution Control Technology Fact Sheet EPA-452/F-03-022

Van der Linde (2007): Personal communication Dr G van der Linde, Fertiliser Society of SA

Van der Walt (2007): Personal communication H J van der Walt, Sasol Ltd SHE Centre

References for non-energy emissions Ball JM, 2006 Bridging the gap from landfill to ‘zero waste’ In: Proceedings, WasteCon 2006. International Waste Management Biennial Congress and Exhibition. Somerset West, 5-8 September 2006.

Bond, W.J., Woodward, F.I. and Midgley, G.F. 2003. Does elevated CO2 play in bush encroachment? Proceedings of Natural Forests and Savanna Woodlands Symposium III, May 2002. Dept Water Affairs and Forestry.

Botha J, 2006 ‘The South African Experience of Conservation and Social Forestry Outreach Nurseries’ Ed T. F. Witkowski Æ Jacklyn Cock, Environ Management vol. 38, pp. 733–749

City of Johannesburg. (2003). The Wastes Management Plan for the City of Johannesburg. Final Report. A Framework for Sustainable Waste Management in the City of Johannesburg. June 2003.

CSIR, 2001: South Africa Greenhouse Gas Inventory, 1994, prepared by the CSIR for the Department of Environmental Affairs and Tourism, Pretoria.

Department of Agriculture (DoA), 2006 Abstract of Agricultural Statistics

Department of Environmental Affairs and Tourism (DEAT). (1999). National waste management strategies and action plans, South Africa. Strategy formulation phase. National Waste Management Strategy. Version D, 15 October 1999.

Department of Environmental Affairs and Tourism (DEAT), 2005 Polokwane Declaration of Waste Management, 2001, Department of Environment and Tourism, www.environment.gov.za/ProjProg/WasteMgmt/Polokwane_ declare.htm

Department of Environmental Affairs and Tourism (DEAT), 2006 ‘Draft National Environmental Management: Waste Management Bill’ published for comments, November 2006

Department of Mineral and Energy (DME), 2004 ‘Landfill gas resources for power generation in South Africa, CBEERE report No. 2.3.4-37, prepared by R Lombard

Department of Water Affairs and Forestry (DWAF), 2001 ‘Waste generation in South Africa (Baseline study in preparation for the National Waste Management Strategy for South Africa)’ Water quality management Series report

Department of Water Affairs and Forestry (DWAF), 2004 ‘Roundwood supply and demand to 2030’ Report prepared by LHA management consultants

DST, 2006 Energy recovery from municipal solid waste in South Africa. A prefeasibility study for the Department of Science and Technology by AGAMA Energy (Pty) Ltd



EPA, 2005 ‘Global Mitigation of Non-CO2 Greenhouse Gases’, EPA Report 430-R-06-005, http://www.epa.gov/ nonco2/econ-inv/pdfs/SectionIIIWaste.pdf

Fairbanks DHK and Scholes RJ, 1999 ‘South African Country study on Climate Change: vulnerability and adaptation assessment for plantation forestry’ prepared for National Research Facility, Pretoria, May 1999

Food and Agriculture Organisation (FAO), 2004 ‘Global Forest Resources Assessment 2005: South Africa’ Country report prepared by Department of Water Affairs and Forestry

Food and Agriculture Organisation (FAO), 2006 South Africa: Country pasture profile- http://www.fao.org/ag/Agp/ agpc/doc/ Counprof/southafrica/southafrica.htm

Forsyth, GG, van Wilgen, BW, Scholes, RJ, Levendal, MR, Bosch, JM, Jayiya, TP and Le Roux, R (2006). Integrated veldfire management in South Africa: An assessment of the current conditions and future approaches. Report CSIR/NRE/ECO/ER/2006/0064/C, CSIR, Stellenbosch.

IPCC (Intergovernmental Panel on Climate Change), 2006. 2006 IPCC guidelines for national greenhouse gas inventories. Prepared by the National Greenhouse Gas Inventories Programme. Eggleston H.S., Buendia L., Miwa K., Ngara T. and Tanabe K. (eds). Kanagawa, Japan, Institute for Global Environmental Strategies. http://www.ipcc-nggip.iges.or.jp/public/2006gl/index.htm.

IPCC, 2007 4th Assessment report by Working Group3 (in preparation)

IPCC (Intergovernmental Panel on Climate Change) 2007. Summary for policymakers: Climate Change 2007: The Physical Science Basis. IPCC WG1 Fourth Assessment Report. Geneva, Intergovernmental Panel on Climate Change. 1 Feb 2007, http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_SPM.pdf (Accessed in Feb 2007).

Hendricks LB, 2006 speech by Minister of Water Affairs and Forestry at Eastern Cape Summit on Forestry and Timber Processing, Mthatha Health Resource Centre, Mthatha, Eastern Cape 26 February 2007

Global Resource Action Center for the Environment (GRACE), 2004 ‘Methane digesters’ [email protected]

Godfrey, L. and Dambuza, T. (2006). Integrated waste management plans – A useful management tool for local government or a bureaucratic burden? WasteCon 2006 Biennial International Waste Congress, Somerset West, South Africa, 4 - 8 September 2006.

Mayet M A G 1993. Domestic Waste Generation in the Urban Core of the Durban Functional Region. MSc Thesis, University of Natal.

Meyer W and Rusk G, 2003 ‘Financial analysis and cost of forestry operations for South Africa and Regions’ by Forestry Economic Services, Pietermaritzburg

Paustian K., Antle JM, Sheenan J., 2006 ‘Agriculture’s role in Greenhouse gas mitigation’, Pew Centre on Global Climate Change publication, September 2006 www.pewclimate.org

Palmer Development Group (PDG), 2004 ‘Methane emissions reduction opportunities in twelve South African Cities: turning liability into a resources’ report to USAID Contact num 674-c-00-0110051-00 December 2004

Phiri A., 2007a ‘Modelling the generation of domestic waste to support the planning of Municipal Waste Services’ CSIR report ??

Phiri A., 2007b ‘Composting as a tool for wealth creation in developing countries’ CSIR report ??

Shackleton, C.M., Buiten, E., Annecke, W., Banks, D., Bester J., Everson, T., Fabricius, C., Ham, C., Kees, M., Modise, M., Phago, M., Prasad, G., Smit, W., Twine, W., Underwood, M., von Maltitz, G. & Wenzel, P. 2004. Fuelwood and poverty alleviation in South Africa: opportunities, constraints and intervention options. Report to Dept of Water Affairs & Forestry, Pretoria. 38 pp


S Sirohi, A Michaelowa and AK Sirohi, 2007 ‘Mitigation options for enteric methane emissions from dairy animals: an evaluation for potential CDM projects in India’ Mitigation and Adaptation Strategies for Global Change (2007) 12:259-274STATS SA, 2006 Mid-year population estimates, South Africa

L.J. Strachan , 2006 ‘Landfill gas to electricity project’ http://www.durban.gov.za/ethekwini/services/dsw/ recycling/har

Van der Merwe, M R & Scholes, R J 1998. South African Greenhouse Gas Emissions Inventory for the years 1990 and 1994. Pretoria, National Committee on Climate Change.

Von Blottnitz, H., Austin, G, Nissing, C., Schmalbein, N, liphoto, L, Ncwadi, N., Gets, A and Fedorsky. C. (2006). Burn, gasify, Pyrolyse or Ferment. Making sense of the many possibilities for energy from waste in South Africa. In: Proceedings, WasteCon 2006. International Waste Management Biennial Congress and Exhibition. Somerset West, 5-8 September 2006.

Kleyn R, 2004 The impact of technical efficiency in the poultry industry on the animal feed industry. Animal Feed Manufacturer Association (AFMA) (http://www.afma.co.za/AFMA_Template/mrt_04_3.html)

B.W. van Wilgen, N. Govender, H.C. Biggs, D. Ntsala, and X.N. Funda, 2004 ‘Response of savannah fire regimes to changing fire-regime policies in a large African national Park’ Conservation Biology, Vol. 18., No.6 pp. 1533-1540



References for macro-economic modeling AGAMA (2003). ‘Employment Potential of Renewable Energy In South Africa,’ Prepared for The Sustainable

Energy and Climate Change Partnership, Johannesburg. AGAMA Energy (Pty) Ltd, Constantia. 14 November 2003.

Armington, P.S. (1969). ‘A theory of demand for products distinguished by place of production,’ IMF Staff Papers, Vol. 16.

Dervis, K., de Melo, J. and Robinson, S. (1982). General Equilibrium Models for Development Policy. Cambridge University Press: New York.

Devarajan, S. (2007). ‘Notes on Dynamics in CGE Models,’ published online at www.cepii.fr. World Bank.

Devarajan, S., Lewis, J.D. and Robinson, S. (1994). ‘Getting the Model Right: The General Equilibrium Approach to Adjustment Policy,’ Mimeo.

DOE/EPA (2006). ‘National Action Plan for Energy Efficiency (NAPEE).’ United States Department of Energy and Environmental Protection Agency. August 2006.

Hoogeveen, J.G. and Özler, B. (2004). ‘Not Separate, Not Equal. Poverty and Inequality in Post-Apartheid South Africa.’ World Bank.

Islam, N. (1999). ‘Capital Mobility in CGE Models: A Survey.’ Department of Economics, Emory University, Atlanta. March 1999.

Kilkenny, M. (1991). ‘Computable General Equilibrium Modeling of Agricultural Policies: Documentation of the 30-Sector FPGE GAMS Model of the United States.’ USDA ERS Staff Report AGES 9125.

King, B.J. (1985). ‘What is a SAM?’ In Social Accounting Matrices: A Basis for Planning, edited by Pyatt, G. and Round, J. Washington D.C.: World Bank.

Laitner, J.A., DeCanio, S.J. and Peters, I. (2000). ‘Incorporating Behavioral, Social and Organizational Phenomena in the Assessment of Climate Change Mitigation Options,’ Prepared for the IPCC Meeting Conceptual Frameworks

for Mitigation Assessment from the Perspective of Social Science. Karlsruhe, Germany. March 2000.

Löfgren, H., Harris, R.L. and Robinson, S. (2001). ‘A Standard Computable General Equilibrium (CGE) Model in GAMS,’ International Food Policy Research Institute (IFPRI): Trade and Macroeconomics Division Discussion

Paper, 75.

Pauw, K. (2005). ‘Forming Representative Household and Factor Groups for a South African SAM,’ PROVIDE

Technical Paper Series, 2005:2. PROVIDE Project, Elsenburg. Available online at www.elsenburg.com/provide.

Pauw, K., Leibbrandt, M., Edwards, L. and Dieden, S. (2006). ‘Trade and Poverty in South Africa: Traded sector employment and vulnerability,’ Report Prepared for the Trade and Poverty Project. Available online at www.saldru.uct.ac.za.

Pauw, K., McDonald, S. and Punt, C. (2007). ‘Agricultural Efficiency and Welfare in South Africa,’ Development

Southern Africa, Forthcoming.

PROVIDE (2005a). ‘Compiling a Social Accounting Matrix for South Africa: 2000,’ PROVIDE Technical Paper

Series, Forthcoming. PROVIDE Project, Elsenburg.

PROVIDE (2005b). ‘The PROVIDE Project Standard Computable General Equilibrium Model. Version 2.,’ PROVIDE Technical Paper Series, 2005:3. PROVIDE Project, Elsenburg. Available online at www.elsenburg.com/provide.

PROVIDE (2006). ‘Compiling National, Multiregional and Regional Social Accounting Matrices for South Africa,’ PROVIDE Technical Paper Series, 2006:1. PROVIDE Project, Elsenburg. Available online at www.elsenburg.com/provide.

Pyatt, G. (1998). ‘A SAM Approach to Modelling,’ Journal of Policy Modelling, 10: 327-352.

Robinson, S., Kilkenny, M. and Hanson, K. (1990). ‘USDA/ERS Computable General Equilibrium Model of the United States.’ USDA ERS Staff Report AGES 9049.

Sadoulet, E. and De Janvry, A. (1995). Quantitative Development Policy Analysis. The John Hopkins University Press: London.

SSA (2002). Labour Force Survey September 2000, Pretoria: Statistics South Africa.

SSA (2003). Final supply and use tables, 2000: an input-output framework, Pretoria: Statistics South Africa.

Thurlow, J. (2004). ‘A Dynamic Computable General Equilibrium (CGE) Model for South Africa: Extending the Static IFPRI Model,’ TIPS Working Paper, No. 1-2004. Trade and Industrial Policy Strategy. February 2004.

long term mitigation scenarios - energy research … · a long term mitigation scenarios for south...

Documents