The Role of Land Use in Determining Greenhouse Gases

Mitigation Costs

Thomas Hertel (Purdue Univ.), Huey-Lin Lee (NIES),Steven Rose (U.S. EPA), and

Brent Sohngen (Ohio State Univ.)

GTAP Working Paper No. 36(Output from the research project sponsored by the US-EPA)

2

Motivation• Land is a significant source of GHG emissions

– Deforestation: 1/3 of total emissions since 1850– Land management: 75% of N2O, 50% of CH4

• Previous studies suggest land-based mitigation is cost-effective – e.g., Sohngen and Mendelsohn (2003), Rao and Riahi (in

press), van Vuuren et al. (in press)• Analytical challenges for land modeling

– Competition for land between land-based sectors– Land-based mitigation competition and net emissions effects– Land heterogeneity and dynamics– Lack of key consistent global data—land, emissions,

mitigation costs• New global datasets—land, emissions, mitigation costs

Provide opportunities for improving our understanding of the role of land in determining GHG mitigation costs.

3

Objective: To analyze the impact of GHG mitigation on land use change in general equilibrium framework

Outline of this presentation: • Land, GHG emissions/sequestration data• Land supply and demand and land-based

emissions modeling in GTAP• Analysis set-up• Results• Conclusions

4

GTAP AEZ Land Use Data

• Our work builds on path-breaking work by Darwin et al. at ERS/USDA, by adding:– More refined definition of AEZs– Climate dimension—tropical, temperate,

boreal– Implementation at the 226-country level– Documented in Lee et al. (2005) and

available on the GTAP website

5

Definition of AEZs in GTAP• 18 AEZs = 6 LGPs x 3 climatic zones

– 6 LGPs = 0–59, 60–119, 120–179,…., 300–365 days– 3 climatic zones = boreal, temperate, tropical

• Follows pioneering work by FAO and IIASA in definition of an AEZ as– land with given “length of growing period” (LGP), as

determined by: temperature, precipitation, soil condition and topography, combined with information from a water balance model and knowledge of physical requirements for growing certain crop.

• Lands classified in same AEZ have homogeneous units within the country—i.e., with similar climate and soil conditions for crop growing.

6

Global Distribution of AEZs

7

Global Land Cover: distribution

000000E+0

0200E+6

0400E+6

0600E+6

0800E+6

1000E+6

1200E+6

1400E+6

1600E+6

1800E+6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

LGP1

LGP2

LGP3

LGP4

LGP5

LGP6

Forest Savan./grassland Shrubland Cropland Pasture Built-up land Other land

Are

a (ha)

Boreal

TemperateTropical

8

-

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

180,000

AEZ

1

AEZ

2

AEZ

3

AEZ

4

AEZ

5

AEZ

6

AEZ

7

AEZ

8

AEZ

9

AEZ

10

AEZ

11

AEZ

12

AEZ

13

AEZ

14

AEZ

15

AEZ

16

AEZ

17

AEZ

18

1000

hec

tare

s

Paddy riceWheatCereals grain necCrops necVegetables, fruit, nutsPlant-based fibresOil seedsSugar cane, sugar beet

Tropical Temperate Boreal

Cropland Hectares by AEZ and crop

Paddy riceWheat

Cereal grain

9

Distribution of Crop Land Rents, within AEZs

0%

20%

40%

60%

80%

100%

1 A

EZ

1

2 A

EZ

2

3 A

EZ

3

4 A

EZ

4

5 A

EZ

5

6 A

EZ

6

7 A

EZ

7

8 A

EZ

8

9 A

EZ

9

10 A

EZ

10

11 A

EZ

11

12 A

EZ

12

13 A

EZ

13

14 A

EZ

14

15 A

EZ

15

16 A

EZ

16

17 A

EZ

17

18 A

EZ

18

8 ocr

7 p fb

6 c_b

5 os d

4 v_f

3 g ro

2 w h t

1 pd r

Vegies/fruits: quick growing, higher value.

10

Non-CO2 emissions & forest sequest’n data• New 2001 non-CO2 emissions data

– Corresponds to GTAP v6 data 2001 base year and complements GTAP 2001 CO2 emissions data

– Highly disaggregated – explicitly for more precise mapping to economic activity (output and input)

• 226 countries• 21 non-CO2 GHG emissions categories (N2O,

CH4, F-gases)– ~145 types of emissions with subcategory

disaggregation• Regional 2000 forest carbon stock data by AEZ,

management type, and tree age cohort• Soil carbon stock data and Other CO2 (non-fossil fuel

combustion) emissions data also available (but yet implemented in the model)

11

GTAP-AEZ sectoral Non-CO2 emissions distribution by region

0

200

400

600

800

1000

1200

1400

1600

1800

2000

1 USA 2 CHN 3 ROW

MtC

eq

24 RWTrade

23 NTrdServices

22 OthManufact

21 OthExtractn

20 Transport

19 Services

18 EnrgIntnsMnf

17 WoodProcessn

16 OthFoodPrcsn

15 ProcessdRice

14 OtherMeatPrd

13 Rumint_Prods

12 GasDistribut

11 Electricity

10 RefinedFuels

9 Gas

8 Oil

7 Coal

6 Forest

5 NonRuminLivs

4 Ruminants

3 OtherCrops

2 OtherGrain

1 PaddyRice

Sectoral distribution of non-CO2 emissions, by region

China Paddy Rice85% CH415% N2O

(in Ceq)

12

The GTAP-AEZ model• Static global CGE• 3 Regions (for now): USA, China, ROW• 24 Sectors – 5 land-based sectors (3 crops, ruminant

livestock, forestry)• Key features:

– Land in 6 AEZs: aggregated from the 18 AEZs– 3-tier CET structure of AEZ-specific land supply

• GHG emissions and sequestration modelling– Incorporate new detailed non-CO2 GHG emissions

data (N2O, CH4, F-gases) and forest carbon sequestration data

– 3 classifications of emissions – output, intermediate inputs, primary factor related emissions

– Introduce emissions pricing– Calibrate mitigation responses

13

Land supply in GTAP

• Standard, yet counterfactual– 1 type of ag. land, imperfect mobility across

uses• Now: Agro-ecologically zoned land

– Heterogeneous in terms of rainfall, temperature, topography, soil type and moisture, etc.

– length of growing period (LGP) varies– Suitability for growing of certain crops– Restricting land mobility across uses

14

3-tier CET structure for AEZ-specific land supply

Land of AEZ i

Forestry Agriculture

GrazingCrops

Crop 1 Crop 2 ..... Crop N

ETREAL1 = -0.25yi

ETRAEL3 = -yi

ETRAEL2 = -0.5yi

Index i allows for AEZ heterogeneity.

15

Sector-specific CES structure for AEZ land demand

• Big enough ESUBAEZ ensures returns to AEZ lands to move closely together• A good approximate of an alternative specification where:

– one prod. function for each AEZ in each activity– AEZ-specific comm. are perfect substitutes– Similar production function for AEZ-specific activity– Each AEZ-specific sector faces same input/factor prices.

Total land demanded by sector j

AEZ NAEZ 2

.....

AEZ 1

ESUBAEZ = 20

16

The GTAP-AEZ model• Static global CGE• 3 Regions (for now): USA, China, ROW• 24 Sectors – 5 land-based sectors (3 crops, ruminant

livestock, forestry)• Production with intra- and inter-regional land

heterogeneity– Land in 6 AEZs: aggregated from the 18 AEZs– CET – 6 different AEZ land endowments

• GHG emissions and sequestration modelling– Incorporate new detailed non-CO2 GHG emissions

data (N2O, CH4, F-gases) and forest carbon sequestration data

– 3 classifications of emissions – output, intermediate inputs, primary factor related emissions

– Introduce emissions pricing– Calibrate mitigation responses

17

Modeling emissions- 3 categories -

• Output – emissions treated as an input to production, represents alternative technologies, introduce new CES elasticity – Follows Hyman et al. (2003) – e.g., coal, oil, energy intensive

manufacturing• Input – emissions proportional

to input use– Endowment – e.g, ruminant :

capital stock (animal herd)– Intermediate input – e.g., grain

crop : fertilizer use

USA China ROW1 PaddyRice 2.971 70.160 137.364

Total non-CO2 GHG emissions (MtCeq)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

USA China ROW

Intermediate inputEndowmentOutput

Paddy Rice

18

Prod. structure: output related emissions incl.

Output

Intermediate InputsValue Added

LandSkilledLabor

UnskilledLabor

NaturalResource

Capital

Output-Emissions composite

Output-related emissions

19

Emissions pricing• The economic impact of an emissions tax associated

with input usage depends on the size of the tax AND the emissions intensity (tC/$) of the input.

• The larger the emissions intensity, the greater the impact of a given carbon tax on the sector’s input use and production.

Emission intensities (tC/$ of input) Input USA China Fertilizer in crops production 0.0061 0.0043 Ruminant livestock capital 0.0099 0.9562 Land in paddy rice 0.0040 0.0125

20

Calibrating mitigation responses• Non-CO2 mitigation

– Engineering mitigation cost estimates for detailed technologies (Delhotal and Kruger, in press; USEPA, 2006)

– Calibrate substitution elasticities with partial equilibrium closure

• Output emissions – ESUBMAC • Endowment emissions – ESUBT • Intermediate input emissions – ESUBVA

• Forest sequestration supply– Calibrated to regional forest carbon supply curves

Sohngen (2005) – afforestation (extensification) and forest management (intensification)

– Calibrated to forest carbon intensities due to presence of unmanaged land in the base year data

21

Calibrated ROW forest carbon sequestration curve via extensification (20-year annual equivalent abatement)

ROW forest sector sequestration MAC: extensification, 20-year annual equivalent abatement

0

25

50

75

100

125

150

175

200

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200Abatement - MMTCE

200

1 U

SD

$ p

er t

on

ne

of

C

GTAP-AEZ_extensification

DGTM_Land Storage

22

Calibrated USA forest carbon sequestration curve via intensification (20-year annual equivalent abatement)

USA forest sector sequestration MAC: intensification, 20-year annual equivalent abatement

0

25

50

75

100

125

150

175

200

-5 5 15 25 35 45 55 65 75 85 95 105 115 125 135 145 155Abatement - MMTCE

2001

US

D $

per

ton

ne o

f C

GTAP-AEZ_intensification

DGTM_age/management storage

23

Analysis of mitigation responses

1. GE global competition: USA-only vs. global carbon tax

2. GE inter-sector competition: USA-only v.s. global carbon tax

24

Mitigation affects regional land competition

GE % change in land rents and land use by sector due to a $50/tonne carbon tax: USA only

1. For a given use, similar land rent responses across AEZs2. Changes in land rents reflect the net effect of mitigation costs and land competition (i.e.,

changes in land prices and changes in acreage) – mitigation cost/subsidy dominates in rice and forestry, land competition dominates in other ag sectors

Percentage change in land rents Forest PaddyRice OtherGrain Other Crops Ruminants

AEZ1 253.5 -15.9 2 3.3 5.1 AEZ2 254.3 -15.3 1.9 3.2 5.1 AEZ3 236.9 -15.5 1.9 3.2 5.1 AEZ4 267.5 -15.5 2 3.2 5.1 AEZ5 295.2 -16.5 2 3.3 5.2 AEZ6 320.2 -20.2 2.3 3.6 5.4

Percentage change in land use, weighted by AEZ land rent share Forest PaddyRice OtherGrain Other Crops Ruminants

AEZ1 0.286 -0.023 -0.211 -0.05 0 AEZ2 0.013 -0.017 -0.226 0.068 0.162 AEZ3 0.009 -0.029 -0.131 0.098 0.053 AEZ4 0.427 -0.026 -0.431 -0.003 0.034 AEZ5 2.134 -0.412 -1.215 -0.391 -0.084 AEZ6 5.604 -0.626 -1.366 -3.242 -0.153

25

Mitigation affects global competition - Regional

GE MAC: 3 re gions

0

10

20

30

40

50

60

70

80

90

100

-20 0 20 40 60 80 100 120 140 160a ba te me nt - mmtce

20

01

US

D p

er

ton

ne

of

C.

US A

CHN

ROW

World

GE MAC: 3 re gions

0

10

20

30

40

50

60

70

80

90

100

0 200 400 600 800 1000 1200 1400 1600 1800a ba te me nt - mmtce

20

01

US

D p

er

ton

ne

of

C.

US A

CHN

ROW

World

US only tax Global tax

WorldUSAROW

China

WorldUSAROWChina

1. USA only carbon tax – USA less competitive, international emissions leakage (primarily deforestation)

2. Vs. global carbon tax – all regions with net reductions, global emissions reductions large, ROW mitigation least expensive (primarily forest carbon increases), USA mitigation most expensive.

26

Mitigation affects global competition – sectors

GE MAC: U.S .A, s e ctora l, to ta l

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140a ba te me nt - mmtce

20

01

US

D p

er

ton

ne

of

C.

p_REGEMIT_AGR

p_REGEMIT_FRS

p_REGEMIT_IND

p_REGEMIT_S VC

p_REGEMIT

GE MAC: U.S .A, s e ctora l, tota l

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140 160 180a ba te me nt - mmtce

20

01

US

D p

er

ton

ne

of

C.

p_REGEMIT_AGR

p_REGEMIT_FRS

p_REGEMIT_IND

p_REGEMIT_S VC

p_REGEMIT

US only taxUSA

IndForest

Svc Ag

Global vs. US tax – US mitigation responses diminished in all sectors.

Global taxUSA

IndForest

Svc Ag

27

Conclusions

• Biophysical and economic land characteristics create comparative abatement advantages for land endowments

intra- and inter-regional reallocation of production, and thus land use change.

• International market structure influences regional mitigation responses

• International leakage is an important component of total GHG emissions

28

Access to LU/GHG data and WP• Land use data:

– https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=1900

• Greenhouse gas emissions data:– CO2:

• https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=1143

– CH4, N2O, F-gases:• https://www.gtap.agecon.purdue.edu/resources/res_display.asp

?RecordID=1186• GTAP Working Paper No. 36:

– https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2230