USAID LEAF Regional Climate Change Curriculum Development Module: Carbon Measurement and Monitoring (CMM) Section 2. Forest carbon stocks and change 2.6. Establishing Lam Dong’s Reference Level for a Provincial REDD+ Action Plan: A Case Study

NameAffiliationNameAffiliation Deborah Lawrence, Co-leadUniversity of VirginiaMegan McGroddy, Co-leadUniversity of Virginia Bui The Doi, Co-leadVietnam Forestry UniversityAhmad Ainuddin NuruddinUniversiti Putra Malaysia Prasit Wang, Co-leadChiang Mai University, Thailand Mohd Nizam SaidUniversiti Kebangsaan Malaysia Sapit DiloksumpunKasetsart University, ThailandPimonrat TiansawatChiang Mai University, Thailand Pasuta SunthornhaoKasetsart University, ThailandPanitnard TunjaiChiang Mai University, Thailand Wathinee SuanpagaKasetsart University, ThailandLawong BalunUniversity of Papua New Guinea Jessada PhattralerphongKasetsart University, ThailandMex Memisang PekiPNG University of Technology Pham Minh ToaiVietnam Forestry UniversityKim SobenRoyal University of Agriculture, Cambodia Nguyen The DzungVietnam Forestry UniversityPheng SoklineRoyal University of Phnom Penh, Cambodia Nguyen Hai HoaVietnam Forestry UniversitySeak SophatRoyal University of Phnom Penh, Cambodia Le Xuan TruongVietnam Forestry UniversityChoeun KimsengRoyal University of Phnom Penh, Cambodia Phan Thi Quynh NgaVinh University, VietnamRajendra ShresthaAsian Institute of Technology, Thailand Erin SwailsWinrock InternationalIsmail ParlanFRIM Malaysia Sarah WalkerWinrock InternationalNur Hajar Zamah ShariFRIM Malaysia Sandra BrownWinrock InternationalSamsudin MusaFRIM Malaysia Karen VandecarUS Forest ServiceLy Thi Minh HaiUSAID LEAF Vietnam Geoffrey BlateUS Forest ServiceDavid GanzUSAID LEAF Bangkok Chi PhamUSAID LEAF Bangkok

IOVERVIEW: CLIMATE CHANGE AND FOREST CARBON 1.1Overview: Tropical Forests and Climate Change 1.2Tropical forests, the global carbon cycle and climate change 1.3Role of forest carbon and forests in global climate negotiations 1.4Theoretical and practical challenges for forest-based climate mitigation IIFOREST CARBON STOCKS AND CHANGE 2.1Overview of forest carbon pools (stocks) 2.2Land use, land use change, and forestry (LULUCF) and CO 2 emissions and sequestration 2.3Overview of Forest Carbon Measurement and Monitoring 2.4IPCC approach for carbon measurement and monitoring 2.5 Reference levels – Monitoring against a baseline (forest area, forest emissions) 2.6 Establishing Lam Dong’s Reference Level for Provincial REDD+ Action Plan : A Case Study IIICARBON MEASUREMENT AND MONITORING DESIGN 3.1Considerations in developing a monitoring system IVCARBON STOCK MEASUREMENT METHODS 4.1Forest Carbon Measurement and Monitoring 4.2Design of field sampling framework for carbon stock inventory 4.3Plot Design for Carbon Stock Inventory 4.4Forest Carbon Field Measurement Methods 4.5Carbon Stock Calculations and Available Tools 4.6Creating Activity Data and Emission Factors 4.7Carbon Emission from Selective Logging 4.8Monitoring non-CO 2 GHGs VNATIONAL SCALE MONITORING SYSTEMS

By the end of this session, learners will be able to:  Identify deciding components of reference level for country-specifics  Establish a workable process for reference level development based on the local context  Develop reference level at the national and sub- national levels taking into account national circumstances

 Lam Dong Working Group – RL establishment  FREC, AFC– Forest Cover Change Map  FIPI – Carbon Stock Analysis  USAID LEAF – Technical Assistance

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

 Reference Level Training – November 2012  Biomass Training – March 2013  Carbon Stock Analysis – 2013  Forest cover change maps – 2013; NFI analysis  Applied RL and Stratification – November 2013  RL Report Final Draft – February 2014

Key Decision for REDD+ RL/REL Lam Dong Decisions 1.Determine Scope of Activities  Include deforestation  Include forest degradation  Include forest enhancement 2.Finalize Forest Definition  Minimum tree cover: 10%  Minimum height: 5 m  Minimum area: 0.5 ha 3.Scale of REDD+ RLSubnational 4.Pools/Gases  Measured Pools: Live tree aboveground biomass  Pools using default values: Dead wood, litter, soil carbon, and live tree belowground biomass  Gases: CO 2 5.Link REDD+ to National Forest Inventory? Yes 6.Adjust for National Circumstances? To be determined. This is expected to be based on specific detailed identification of events, and conservative assessment of the likely impacts. 7.Location Analysis?Possibly in the future, to be determined.

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

Defined and analyzed by FREC’s Forest Cover Change Assessment (FCCA) work Drivers of deforestation  Agriculture expansion  Plantation establishment  Infrastructure and hydropower projects  Logging Drivers of forest degradation  Timber extraction  Non-timber forest extraction

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

 According to Circular No. 34/2009/TT-BNNPTNT of June 10, 2009  Given the mosaic pattern of forests  Consistency with national context Forest/ land cover class Evergreen - Broadleaf - Rich Evergreen - Broadleaf - Medium Evergreen - Broadleaf - Poor Evergreen - Broadleaf - Regrowth Deciduous Bamboo forest Mixed Wood and Bamboo Coniferous - Rich Coniferous - Medium Coniferous - Poor Coniferous - Regrowth Mixed Broadleaf and Coniferous Plantation Bare land Agricultural and other land Water area Residential area

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

 Activity Data for deforestation, forest degradation and afforestation/reforestation  Historical periods: , , ,  Derived from pairwise comparison of the land cover maps for 1990, 1995, 2000, 2005 and 2010  Landsat and SPOT satellite images  Forest knowledge from local partners  Thirteen forest/land cover classes

Image classification 1990 Corrected JICA/IC map 1990 Image classification 1995 Corrected IC map 1995 Image classification 2000 Corrected JICA/IC map 2000 Image classification 2005 Corrected IC map 2005 Image classification 2010 Corrected JICA/IC map 2010 Ancillary data Land cover change map Final map 1990 Final map 1995 Land cover change map Final map 2000 Land cover change map Final map 2005 Land cover change map Final map 2010 Input data: Landsat images for 1990, 1995 SPOT images for 2000, 2005, 2010 Aerial photographs and 2000 Previously developed maps for Lam Dong province Methods: Image segmentation with eCognition will be used for image classification for all years Overlay between two consequent maps and manual correction of wrong classified land cover transitions Accuracy Assessment analysis Result of maps Full standard maps. Which have been combined from commune levels Formal accuracy assessment will be done for all maps Ancillary data JICA 2000 Ancillary data JICA 2010 Ancillary data JICA 1990 Manual correction of wrong classified LC transitions. And visualization interpret again for every forest boundary to adjust the forest boundaries for more detail Image segmentation with eCognition JICA 1995 JICA Collected interpretation keys - Aerial photos; -Experiential interpreters. -Local experts and forest owners Ground truth points should be set up in the protected areas, national parks in which the forest may not be changed over all time intervals Task 1: Computer based method Task2: empirical method Step3: Ground trust accuracy assessment method

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

The carbon stocks for each forest stratum undergoing change is determined to define emission factors

Forest typeSum of Average Carbon Stocks (tC/ha) Average tree Carbon Stocks (tC/ha) Default IPCC Dead Wood Carbon Stocks (tC/ha) Default IPCC Litter Carbon Stocks (tC/ha) Evergreen - Broadleaf forest - Rich Evergreen - Broadleaf forest - Medium Evergreen - Broadleaf forest - Poor Evergreen - Broadleaf forest - Regrowth Deciduous forest Bamboo forest220.0 Mixed Broadleaf and Coniferous forest Coniferous forest - Rich Coniferous forest - Medium Coniferous forest - Poor Coniferous forest – Regrowth(*) Mixed Wood and Bamboo forest Plantation forest23 0.2

Forest Type Average emission factor from deforestation (t C/ha) AgricultureBare land Residential/ infrastructure Evergreen - Broadleaf forest - Rich Evergreen - Broadleaf forest - Medium Evergreen - Broadleaf forest - Poor Evergreen - Broadleaf forest - Regrowth Deciduous forest Bamboo forest Mixed Broadleaf and Coniferous forest Coniferous forest - Rich10691 Coniferous forest - Medium9378 Coniferous forest - Poor7358 Coniferous forest – Regrowth Mixed Wood and Bamboo forest65 50 Plantation forest48 33

 For deforestation, degradation, and afforestation/reforestation  Live tree carbon stock estimates from NFIMAP Cycle IV raw field data ( ), collected by FIPI  Litter, dead wood and soil carbon pools based on IPCC defaults  Post deforestation land uses: Agriculture, Settlements, Bare Land (assumed 0 t C)

Forest Carbon Stratum/ Forest type Live Tree Carbon Stock (t C.ha -1 ) Uncertainty (%) Evergreen - Broadleaf forest - Rich Evergreen - Broadleaf forest - Medium Evergreen - Broadleaf forest - Poor Evergreen - Broadleaf forest - Regrowth Deciduous forest Bamboo forest Mixed Broadleaf and Coniferous forest Coniferous forest - Rich Coniferous forest - Medium Coniferous forest - Poor Coniferous forest – Regrowth* Mixed Wood and Bamboo forest Plantation forest

 Uncertainty reflects the degree of both precision and accuracy in a reported measurement  Increasing the number of measurements (in this case plots) can help reduce uncertainty associated with precision  Refining the strata to reduce variation with a forest type can also reduce uncertainty

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

ActivityAnnual average (million tons CO 2 e) Sum total (million tons CO 2 e) Deforestation Forest Degradation Afforestation/ Reforestation Net Total

 Uncertainty was estimated for all Activity Data and for each Emission Factor, then for emissions from each historical period using simple propagation of errors.  Uncertainty for deforestation ranged from 14.4% to 207.4%, with a median value of 45.7%.  Uncertainty for degradation ranged from 11.2% to 52.4%, with a median value of 30.1%.  Uncertainty for A/R ranged from 22.7% to 216.7%, with a median value of 82.5%.

Sum Emissions Reference Level MRV System Emissions from Deforestation Emissions from Forest Degradation Emission Reduction from Forest Enhancement Future Projection of Emissions Stratification Driver Analysis Deforestation Activity Data Deforestation Emission Factors Forest Degradation Activity Data Forest Degradation Emission Factors Forest Enhancement Activity Data Forest Enhancement Emission Factors KEY Final Outcome Policy decision Intermediate outcome Technical input

 The options are presented for a RL projected 10 years from the base year (2010) to  Average  Continuation of historical trend  Adjusted for national (subnational) circumstances RL  Regardless of the RL that Lam Dong decides to follow, the emissions after 2010 (the start year of REDD+) must fall below the trend line for Lam Dong to demonstrate successful performance.

The average RL is set as continuation of historical average.

Estimation of HISTORICAL emissions and removal is a key element for constructing RL  AD – measure of extent of activity  EF – GHG emissions/ removal per unit activity

 Improved inventory data, specific for carbon stock estimates, including all relevant pools, strata, and post-deforestation land uses  Develop EFs appropriate for older historical periods (or reduce reference period to )  Ground-truth land use classes and activity data  Improve uncertainty values in certain strata  Develop driver-specific activity data and emission factors for degradation and A/R

 Address data gaps by establishing a system that takes both the sampling design and geospatial aspects of this analysis into consideration from the early stages.  A forest monitoring system that is designed with REDD+, (or RELU), as one of its goals will enable those implementing it to better understand what data must be collected and how they can be applied to GHG emission estimates.  Ideally, those responsible for the design of any data collection plan will also be involved in the analysis and reporting of that data.