1. Harnessing the carbon market to sustain ecosystems and alleviate poverty Draft REDD Methodology BioCarbon Fund Workshop, February 5 – 8, 2008

2. History and Plan 2007: 1 st Draft Reviewers: Ben de Jong, Bernhard Schlamadinger, Tim Pearson + Comments from Andrea Garcia and Marc Steiniger 2008: January:2 nd Draft February:Comments from WB colleagues Finalized 2 nd Draft March: 2 nd round of comments by reviewers June:3 rd draft submitted to VCS?

3. Existing guidance IPCC (www.ipcc.ch): Revised 1996 GL for National GHG Inventories. 2003 GPG for Land Use, Land Use-Change, and Forestry. 2006 GL for National GHG Inventories, Vol. 4, Agriculture, Forestry and Other land Uses (AFOLU). Winrock International (www.winrock.org): Reducing GHG Emissions from Deforestation and Degradation in Developing Countries: a Sourcebook of Methods and Procedures for Monitoring, Measuring and Reporting. Land Use, Land Use Change and Forestry Projects. Voluntary Carbon Standard (www.v-c-s.org): Guidance for Agriculture, Forestry and Other Land Use Projects

4. Issues Insufficient policy guidance on critical methodological issues: –Definitions (forest, deforestation, degradation) –Project and Baseline Emissions –Emissions from land-use after deforestation –Leakage (types, attribution, assessment) –Etc. Components of a baseline –Quantity –Location –Carbon stock changes

5. Applicability conditions (a)Only “gross deforestation”. (b)Deforestation agents, drivers and underlying causes can be identified. (c)Deforestation and forest degradation agents are bound to a certain region and cannot migrate beyond that region. (d)Information on forest cover and forest status before the start of the project activity is available for at least two points in time. (e)Information on land use-use and land-cover on deforested land before the start of the project activity is available.

6. Step 1. Define the boundaries of the proposed REDD project activity: spatial, temporal, carbon pools and sources of greenhouse gas emissions. Step 4. Project the quantity of future deforestation and forest degradation Step 3. Analyze agents, drivers and underlying causes of deforestation and forest degradation Step 5. Project the location of future deforestation and forest degradation by analyzing the spatial correlation between historical land-use and land-cover change and biogeophysical and socioeconomic factors (vicinity to roads, slope, population density, etc.) Step 7. Estimate the expected baseline carbon stock changes and non-CO 2 emissions. Step 2. Analyze historical land-use and land-cover change in the reference region during the past 10-15 years and project potential forest regeneration. Approach a: Linear projection. Deforestation and forest degradation are linearly projected from observed historical trends. Approach c: Corridor. Deforestation and forest degradation are projected as a likelihood corridor to reflect uncertainty. Step 6. Project future baseline activity data (the land-use and land-cover change component of the baseline) by combining the results of steps 2, 4 and 5. Step 10. Calculate the expected ex ante net anthropogenic GHG emission reductions. Step 8. Estimate the expected actual carbon stock changes and non-CO 2 emissions. Step 9. Estimate the expected leakage carbon stock changes and non-CO 2 emissions. Approach b: Multiple regression. Deforestation and forest degradation are projected using multiple regression. Select the applicable baseline approach and justify the choice. Step 1. Define the boundaries of the proposed REDD project activity: spatial, temporal, carbon pools and sources of greenhouse gas emissions. Step 4. Project the quantity of future deforestation and forest degradation Step 3. Analyze agents, drivers and underlying causes of deforestation and forest degradation Step 5. Project the location of future deforestation and forest degradation by analyzing the spatial correlation between historical land-use and land-cover change and biogeophysical and socioeconomic factors (vicinity to roads, slope, population density, etc.) Step 7. Estimate the expected baseline carbon stock changes and non-CO 2 emissions. Step 2. Analyze historical land-use and land-cover change in the reference region during the past 10-15 years and project potential forest regeneration. Approach a: Linear projection. Deforestation and forest degradation are linearly projected from observed historical trends. Approach c: Corridor. Deforestation and forest degradation are projected as a likelihood corridor to reflect uncertainty. Step 6. Project future baseline activity data (the land-use and land-cover change component of the baseline) by combining the results of steps 2, 4 and 5. Step 10. Calculate the expected ex ante net anthropogenic GHG emission reductions. Step 8. Estimate the expected actual carbon stock changes and non-CO 2 emissions. Step 9. Estimate the expected leakage carbon stock changes and non-CO 2 emissions. Approach b: Multiple regression. Deforestation and forest degradation are projected using multiple regression. Select the applicable baseline approach and justify the choice. Ex ante methodology steps Ex ante = before validation

7. Ex post methodology steps Step 1. Define the boundaries of the proposed REDD project activity: spatial, temporal, carbon pools and sources of greenhouse gas emissions. Step 3. Analyze agents, drivers and underlying causes of deforestation and forest degradation Step 2. Analyze historical land-use and land-cover change in the reference region during the past 10-15 years and project potential forest regeneration. Step 11. Project monitoring. Step 13. Adjustment of the baseline for future crediting periods. Step 12. Ex post calculation of net anthropogenic GHG emssion reduction Ex post = during project implementation

8. Ex ante methodology steps

9. Step 1. Define the boundaries of the proposed REDD project activity. 1.1Spatial boundaries 1.2Temporal boundaries 1.3Carbon pools 1.4Sources of emissions and GHG

10. Reference region = Forest to be protected / managed = Area where pre-project activities could be displaced = Domain from which information on DD agents, drivers and rates is extracted and projected. Step 1. Define the boundaries of the proposed REDD project activity. 1.1Spatial boundaries Leakage belt Project area

11. Step 1. Define the boundaries of the proposed REDD project activity. 1. 2Temporal boundaries time Project start Project end End of 1 st crediting period and start of 2 nd crediting period End of 2 st crediting period and start of 3 rd crediting period min 20 max 100 yrs max 10 yrs Start of the historical reference period End of the historical reference period 10-15 yrs

12. Step 1. Define the boundaries of the proposed REDD project activity. 1. 3Carbon pools 1.Above-ground biomass Trees Non-Trees 2.Below-ground biomass 3.Dead wood Standing Lying Wood products 4.Litter 5.Soil organic carbon Selection criteria: Principle of conservadurism. Expected magnitude of carbon stock change. Cost Could be different carbon pools depending on the land-use/land- cover change category.

13. TreesDead Wood Soil Carbon Non-tree Vegetation Wood Products Before Deforestation After Deforestation Carbon Stock (Brown et al., 2007)

14. Carbon Stock time Trees Non-Tree Vegetation Dead Wood Soil Carbon Harvested wood

15. Step 1. Define the boundaries of the proposed REDD project activity. 1.4 Sources of emissions and GHG Under the REDD activity scenario: GHG emissions that would occur in the baseline are avoided. GHG emissions due to project activities are likely to occur. GHG emissions due to leakage are likely to increase. Conservatively ignore them. Can count non-CO 2 emissions from forest fire. Reasonably assume that project emissions are less than baseline emissions  ignore. Leakage prevention measures and activity displacement may lead to significant GHG emissions  consider.

16. Step 1. Define the boundaries of the proposed REDD project activity. 1.4 Sources of emissions and GHG SourceGHGComment Forest firesCH 4 and N 2 OAvoided baseline emission FertilizationN2ON2OLeakage prevention measures Enteric fermentation CH 4 Leakage prevention measures Manure management CH 4 and N 2 O Leakage prevention measures Non-significant sources are omitted Significance is tested using the EB-CDM approved tool

17. Step 2. Analysis of historical Land-Use and Land-Cover Change. 2.1Select data sources 2.2Define land-use/land-cover classes 2.3Define LULC-change categories 2.4Prepare LULC and LULC-change maps and LULC-change matrices 2.5Assess map accuracy 2.6Prepare a methodological annex

18. Step 2. Analysis of historical Land-Use and Land-Cover Change. 2.1Select data sources Prefer existing LU/LC and LU/LC-change maps if they are already approved by the national authority and/or independently validated. Unclassified remotely sensed data (medium resolution – Landsat, Spot). Ground-truth data (high resolution remotely sensed data and/or field data).

19. Step 2. Analysis of historical Land-Use and Land-Cover Change. 2.2Define land-use / land-cover classes Consistent with IPCC classes: -Forest Land- Wetland -Cropland- Settlement -Grassland- Other land Consult National GHG inventory for the definitions. Subdivide, as necessary, to build homogeneous carbon stock density classes.

20. Time tCO 2 e ha -1 Undisturbed Forest Initial degradation Intermediate degradation Advanced degradation Forest Degradation Forest Non- Forest Carbon Density Classes Degradation

21. tCO 2 e ha -1 Time Recovered Forest Initial succession Intermediate succession Advanced succession Forest Plantation or Succession Forest Non-Forest Carbon stock enhancement (“Regeneration”) Carbon Density Classes

22. Step 2. Analysis of historical Land-Use and Land-Cover Change. 2.3Define LU/LC-change categories Describe class transitions in case of degradation and carbon stock enhancement (= “ Forest Regeneration”). Intact forest Cropland Grassland Wetland Settlement Other Land Forest LandNon-Forest Land Forest Regeneration Forest Degradation Deforestation Degraded forest Managed forest

23. Non-Forest Degraded Forest Class X ForestNon-Forest Forest Class X Forest Regeneration Class A Non-Forest Degraded Forest Class X Forest Regeneration Class B Forest Regeneration Class A Degraded Forest Class X Forest Regeneration Class B Regeneration (C-stock enhancement) DegradationDeforestation Forest Regeneration Class A Forest Regeneration Class B Forest Regeneration Class C Forest Regeneration Class D Regeneration (C-stock enhancement) DegradationDeforestation Forest Class X Degraded Forest Class Y Possible class transitions during one crediting period

24. Step 2. Analysis of historical Land-Use and Land- Cover Change. 2.3Prepare LU/LC and LU/LC-change maps and LU/LC-change matrices Forest cover benchmark map Past and current LU/LC Past LU/LC change Potential forest regeneration map 2.5Assess map accuracy 2.6Prepare a methodological annex

25. Step 3. Analysis of agents, drivers, underlying causes and chain of events 3.1Agents Who is deforesting / degrading? 3.2Drivers: What drives the agents to cut the trees? (a)Spatial variables (predisposing factors) (b)Economic and social variables. 3.3Underlying causes: Ultimate reasons explaining the drivers. 3.4Chain of events: Relationships between agents. Typical sequence of events leading to deforestation or degradation.

26. Step 4. Project the quantity of future deforestation and forest degradation 4.1Analysis of remaining forest area that is suitable for conversion to non-forest use and logging activities 4.2Selection of the baseline approach 4.3Quantitative projection of future deforestation and forest degradation

27. Step 4. Project the quantity of future deforestation and forest degradation 4.1Analysis of remaining forest area that is suitable for conversion to non-forest use and logging activities Time DD rates are likely to decrease once only “sub-optimal” areas remain available. DD rates should decrease once only “marginal” areas remain available. DD rates should be zero once no suitable area remains available. DD rates are likely to continue at the historical level as long as “optimal” areas are available. Ha yr -1

28. Step 4. Project the quantity of future deforestation and forest degradation 4.2Selection of the baseline approach (a)Linear projection: Future deforestation and degradation is linearly projected from past trends. (b)Modeling approach: Future deforestation and degradation is projected using multiple regression techniques (requires sufficient data points). (c)Corridor approach: Future deforestation and degradation is projected as an average of the historical rate. The 90% confidence interval of the mean is calculated to determine a “corridor” of likely baseline deforestation and forest degradation.

29. Variable DD Observed Projected Approach b: Modeling approach Monitoring and ex post adjustment (dynamic baseline) time ha Constant DDDecreasing DD Increasing DD Observed Projected Approach a: Linear projection time ha time ha time ha Variable DD Observed Projected Approach c: Corridor approach 90% CI time ha 50% credit 0 credit Full credit

30. Step 4. Project the quantity of future deforestation and forest degradation 4.3Quantitative projection of future deforestation and forest degradation YearDeforestation (ha)Degradation (ha) 1 2 3 4 5 6 7 8 … Project end

31. Step 5. Project the location of future deforestation and forest degradation Distance to roads Deforestation in year 1 Deforestation in year 2 Deforestation in year 3 Deforestation in year … Suitability Road

32. Step 5. Project the location of future deforestation and forest degradation 5.1Create driver maps from spatial variables. 5.2Create suitability maps for deforestation and for degradation. 5.3Select the most accurate suitability map for deforestation and for degradation. 5.4Locate future deforestation and forest degradation.

33. Slope Vicinity to roads Logging areas Land allocation projects Suitability Map Spatial variables  Driver Maps CATIE Study in Costa Rica (1996-2006) Ex post correlation with actual deforestation: r = 0.91 (p < 0.001)

34. Step 4 Step 5 Deforested Potential Degradation Map year 1 Category 1 of Categorized Potential Degradation Map Area that would be degraded in year 1 Deforested Potential Degradation Map year 2 Area that would be degraded in year 2 Category 2 of Categorized Potential Degradation Map Deforested Potential Degradation Map year 3 Area that would be degraded in year 3 Category 3 of Categorized Potential Degradation Map Continue Categorized Potential Degradation. Map In case that degradation is taken into account…

35. Step 5. Project the location of future deforestation and forest degradation At the end of Step 5 we have: Initial LU/LC Cover Map Potential Forest Regeneration Map Potential Deforestation Map Potential Forest Degradation Map From step 2 From step 5

36. Step 6. Project future land-use and land-cover change 6.1Identification of LU/LC-change categories in “forest land remaining forest land” Forest degradation Forest regeneration (= carbon stock enhancement) 6.2Identification of LU/LC-change categories in “forest land converted to non-forest land”. Deforestation

37. LU/LC-Map year 0 Potential Deforestation year 1 Deforested LU/LC-Map year 1 Potential Degradation year 1 Deforested & Degraded LU/LC-Map year 1 Potential Regeneration year 1 D&D & Regenerated LU/LC-Map year 1 Potential Degradation year 2 Potential Deforestation year 2 Deforested LU/LC-Map year 2 Deforested & Degraded LU/LC-Map year 2 Potential Regeneration year 2 D&D & Regenerated LU/LC-Map year 2 LU/LC-Map year 1 Continue with all future years Step 2 Step 5 Step 6 6.1Identification of LU/LC-change categories in “forest land remaining forest land”

38. Degradation and carbon stock enhancement (“regeneration”).

39. 6.1Identification of LU/LC-change categories in “forest land remaining forest land” Deforestation.

40. Step 6. Project future land-use and land-cover change 6.2Identification of LU/LC-change categories in “forest land converted to non-forest land”. Three methods are available to do this projection: (1)Historical LULC-change; (2)Suitability modeling; and (3) Ex post observations (dynamic baseline).

41. 6.2Identification of LU/LC-change categories in “forest land converted to non-forest land”.

42. Step 7. Estimate the expected baseline carbon stock changes and non-CO 2 emissions. 7.1Estimation of the average carbon stock density of each LU/LC class. 7.2Estimation of non-CO 2 emissions from forest fires (if applicable). 7.3Calculation of Emission Factors. 7.4Calculations of carbon stock changes due to forest degradation and regeneration. 7.5Calculation of carbon stock changes (and non- CO 2 emissions) due to deforestation. 7.6Estimation of total baseline carbon stock changes and non-CO 2 emissions (C BASELINE )

43. LU/LC- change category Deg Reg Def 7.3 Calculation of Emission Factors.

44. 7.4Calculations of carbon stock changes due to forest degradation and regeneration.

45. 7.5Calculation of carbon stock changes due to deforestation.

46. 7.6Estimation of total baseline carbon stock changes and non-CO 2 emissions (C BASELINE )

47. Step 8. Estimate the expected actual carbon stock changes. Estimations are based on planned project activities. The expected level of “activity data” is reported in tables similar to the previous ones. The numbers of reduced activity data for “degradation” and “deforestation” and the underlying assumptions must be explained and justified. If specific measures are undertaken to enhance carbon stocks in “regeneration” forest classes, the Potential Forest Regeneration Map must be adjusted accordingly.

48. Step 9. Estimation of expected leakage: carbon stock changes and non-CO 2 emissions (C LEAKAGE ) E Displacement = (C BASELINE - C ACTUAL ) * X%

49. Assumptions about attributability of different types of leakage Attributable to a REDD project activity are: –Activity displacement. –Emissions from measures implemented to prevent activity displacement. Not attributable are: –Positive leakage (consistently with CDM). –Market effects: Project can not control markets and market actors, but government can Market effects are attributable to governments If market effects are not controlled by governments and carbon prices, DD will continue.

50. Step 10. Calculate the expected ex ante net anthropogenic GHG emission reductions.

51. Ex post methodology steps Step 1. Define the boundaries of the proposed REDD project activity: spatial, temporal, carbon pools and sources of greenhouse gas emissions. Step 3. Analyze agents, drivers and underlying causes of deforestation and forest degradation Step 2. Analyze historical land-use and land-cover change in the reference region during the past 10-15 years and project potential forest regeneration. Step 11. Project monitoring. Step 13. Adjustment of the baseline for future crediting periods. Step 12. Ex post calculation of net anthropogenic GHG emssion reduction

52. Step 11. Project monitoring. 11.1Project implementation: Measures to reduce deforestation and forest degradation; Measures to enhance carbon stocks; and Measures to reduce the risk of leakage. 11.2Land-use and land-cover change in the reference region, project area and leakage belt 11.3Driver variables used to estimate the quantity and location of future deforestation and forest degradation 11.4Carbon stocks

53. Step 12.Calculation of ex post net anthropogenic GHG emission reductions Ex ante projection Ex post measured

54. Step 13.Adjustment of the baseline projections for future crediting periods 13.1Adjustment of the land-use / land-cover change component of the baseline 13.2Adjustment of the carbon stock-change component of the baseline

55. ? B B B time T1T1 T2T2 T3T3 T4T4 t -2 t -1 t1t1 t2t2 t0t0 T3: 4 x A A A A A T1: 1 x B T1: 1 x A A B T2: 3 x A T2: 2 x B A A A B B 13.1Adjustment of the land-use / land-cover change component of the baseline T2: (PA+LKB)/RR = 2/3 T3: (PA+LKB)/RR= ?/4 T3: 2/3 = ?/4 ? = 2/3*4 = 2.6 Reference Region Project Area and Leakage Belt

56. 13.2Adjustment of the carbon stock-change component of the baseline Measure and adjust, as necessary, the estimated average carbon stock density of LU/LC classes Adjust emission factors

57. End of ex post methodology steps

58. Thank you!