1. Field Monitoring for LULUCF Projects Training Seminar for BioCarbon Fund Projects February 6 th 2008 Timothy Pearson and Sarah M. Walker Winrock International

2. Project Monitoring – Components:  Carbon stocks  Leakage  Project Emissions  Project implementation  Forest Establishment  Forest Management  Project Boundary

3. Project Monitoring – Goals:  Estimate GHG sequestration or avoided emissions by project  Estimate changes in carbon stocks  Estimate emissions caused by project  Overall Goal:  Conservatively estimate changes with low uncertainty, minimizing errors

4. Project Monitoring  Detailed monitoring plan will need to be created for each component monitored  Standard operating procedures for collection  Excel data sheets and calculation tools  Frequency of collection  Tools to monitor and reduce error  Protocol for data storage  Project member responsible for data collection, analysis and storage

5. Monitoring Project Implementation  Forest Establishment  Site preparation conforms to PDD  Species and planting density inline with PDD  Forest Management  Cleaning and thinning  Fertilization  Harvesting

6. Monitoring Carbon Stocks  Should you monitor the baseline?  When project starts baseline no longer exists  Proxy sites?  No proxy in avoided deforestation  Possible in FM and AR but doubles monitoring costs  Monitor project  Monitor project area  Monitor project carbon stocks

7. IPCC GPG Chapter 4.3  Provides good practice guidance for JI and CDM projects and includes guidance on:  defining project boundaries  measuring, monitoring, and estimating changes in carbon stocks and non-CO 2 greenhouse gases  implementing plans to measure and monitor  developing quality assurance and quality control plans

9. Measurement Plan  Baseline  Step 1: MEASURE Carbon at beginning of project 14 t C/ha

10. Measurement Plan  Baseline  Step 1: Carbon at beginning of project  Step 2: ESTIMATE Carbon over time E.g.: Change in Carbon Stocks Baseline: 14 t C/ha Baseline: Year 40 32 t C/ha

11. Measurement Plan  Project: Plant Trees  Step 1: Carbon at beginning of project  Step 2: MEASURE Carbon over time Project: Year 40 110 t C/ha Baseline: Year 40 32 t C/ha

12. Measurement Plan  Project: Plant Trees  Step 1: Carbon at beginning of project  Step 2: Carbon over time 40 years Net Sequestered: 110 - 32 = 78 t C/ha Project: Year 40 110 t C/ha Baseline: Year 40 32 t C/ha

13. Accuracy and precision  Accuracy:  agreement between the true value and repeated measured observations or estimations  Precision  illustrates the level of agreement among repeated measurements of the same quantity Accurate but not precise Precise but not accurate Accurate and Precise

14. Accurate but imprecise 145 95 170 110 80 Average120 95 % confidence interval 45.8 True mean of forest carbon stock = 120 t C/ha

15. Accurate but imprecise Inaccurate but precise 145180 95183 170177 110178 80182 Average120180 95 % confidence interval 45.83.2 True mean of forest carbon stock = 120 t C/ha

16. Accurate and Precise Accurate but imprecise Inaccurate but precise 118145180 12395183 121170177 118110178 12080182 Average120 180 95 % confidence interval 2.645.83.2 True mean of forest carbon stock = 120 t C/ha

17. Developing a measurement plan Define project boundary Stratify project area Decide which carbon pools to measure Develop sampling design--plot type, shape, size, number, and layout Determine measurement frequency

18.  May been different sets of ‘boundaries’ in the project  Project area = where participants are doing activities  A/R CDM project boundary = this ONLY contains the areas where trees will be planted for carbon credits  Forest Enrichment/Protection boundary = forestry area where voluntary carbon credits will be accounted Define project boundary Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

19.  For accurate measuring and monitoring, boundaries must be clearly defined from start of project  Also a requirement for project registration  Goal is to monitor that boundaries do not change through project due to encroachment, disturbance  Define boundaries using features on map or coordinates attained using a global positioning system  Maintain areas in GIS Define project boundary Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

20.  Maximize the area for carbon credits  Easy and efficient to monitor and verify using GPS  Excludes areas that have little to no carbon benefit  Excludes areas where baseline carbon stocks (and leakage) are more difficult to estimate than the potential carbon benefit warrants (e.g., villages) Define project boundary Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

21.  Project can vary in size: 10’s ha  1000’s ha  Project can be one contiguous block OR many small blocks of land spread over a wide area  One OR many landowners  Only includes lands that meet eligibility conditions Define project boundary Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

22. Principles of monitoring carbon  Methods for measuring carbon credits are based on measuring changes in carbon stocks  Not practical to measure everything - so we sample  Sample subset of land by taking relevant measurements of selected pool components in plots  Number of plots measured predetermined to ensure both accuracy and precision

23. Principles of monitoring carbon  There is a trade-off between the desired precision level of carbon-stock estimates and cost  In general, the costs will increase with:  Greater spatial variability of the carbon stocks  The number of pools that need to be monitored;  Precision level that is targeted;  Frequency of monitoring;  Complexity of monitoring methods.  Stratification of the project lands into a number of relatively homogeneous units can reduce the number of plots needed.

24. Stratify Project Area  If different areas will contain different amounts of carbon, ‘stratify’ project = divide area into different ‘strata’  Stratify based on factors that will affect CARBON stock  One stratum can be made up of one large block of land or several small blocks of land, as long all of the blocks have similar carbon stocks Baseline strata: Grazing land – 10,000 ha Crop land – 8,000 ha Not part of carbon project Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

25. Stratify Project Area  Will most likely have separate strata for baseline and monitoring  Must monitor area of each strata over project  Area and number of strata may change over life of project Baseline strata: Grazing land – 10,000 ha Crop land – 8,000 has Not part of carbon project Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency Project strata: Agroforestry Natural forest restoration

26. Stratify Project Area  Land use  Slope  Drainage e.g. flooded, dry  Elevation  Proximity to villages, towns  Age of vegetation e.g. ‘cohort’  Species composition, stand model Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

27. Decide which carbon pools to measure Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency Soil or Peat Carbon AG non-tree non-woody vegetation Above Ground Live Trees Litter Dead wood Belowground Live Trees (roots) Above Ground Non-tree Woody Wood products

28.  Selection of pools depends on:  Expected rate of change  Expected magnitude and direction of change  Availability of methods, accuracy and cost of methods to measure and monitor  For afforestation and reforestation over < 60 years it is always most economic and efficient to measure live tree biomass (above and belowground) Decide which carbon pools to measure Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

29. Decide which Carbon pools  MUST measure ALL pools that predicted to be smaller in project than in baseline Under project Baseline Under project Baseline Required:Optional: Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

30. Decide which carbon pools to measure Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency Project Type Project Type Carbon Pools Live Biomass Dead Biomass Soil Wood TreesUnderstoryRootsFineCoarse Products -Restore native forests YMYMMM N -Plantations for timber YNYMMM Y -AgroforestryYYYNNM M -Soil carbon management NMMMNY N -Short-rotation plantations YNMNNM * -Forest management YMNMYN Y -Reduce deforestation YMYMMM Y Y=recommended, M=maybe, N=not recommended, * Stores carbon in unburned fossil fuels  Selection of pools varies by project type  Different measuring and monitoring designs are needed for different types of projects

31. Sampling Design: Distribution of plots  Sample units must be located without bias  Randomly distribute sample units using GIS  The entirety of the project site should be sampled  To ensure above, location of plots determined prior to field work Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

32. Sampling Design: Type of plots  Permanent plots  Statistically more efficient for measurements through time  Permit verification  Must mark trees to track ingrowth and mortality  Temporary plots  Measurements made only one time  Preliminary data  Non-tree pools Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

33. Sampling Design: Number of plots  Number of plots:  Identify the desired precision level  ±10 % of the mean is most common but as low as ±20 % of the mean could be used  Collect preliminary data to estimate variability of carbon stocks Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

34. Sampling Design: Number of plots  Preliminary data collection:  Collect data in areas similar to what vegetation will be in your project and baseline  For example: collect data in:  5-10 year old forest  20-30 year old forest  Randomly locate ~10 plots within each strata  Take field measurements using same methods will use for measurement plan Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

35. Sampling Design: Number of plots  Estimate mean carbon stock and variance from preliminary data  Calculate the required number of plots using equation or Excel plot calculator file Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

36. Sampling Design: Number of plots  More variable C stocks  more plots needed for precision level  If a stratified project area requires more plots than an unstratified area  remove 1+ strata  If strata analyzed together  C stocks in each strata cannot be reported separately but fewer plots needed to attain precision level Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

37. Sampling Design: Number of plots  Non-tree biomass pools:  Above method can be used  OR: # non-tree pools in proportion to # tree plots For example:  For every tree plot, sample:  Single 100 m line transect for dead wood  4 sub-plots for herbaceous, forest floor, soil  May result in large variance, but overall amount small in comparison to tree carbon stock Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

38. Achieving Precision – Noel Kempff 452

39. Cost of Precision – Noel Kempff 0 50 100 150 200 250 300 350 Monitoring Cost ($1000s) 5%10%20%30% Precision Level Variable Fixed

40. Sampling Design: Shape of plots  Trees  Large trees:few, very spread out  Small trees:many, close together Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

41. Sampling Design: Shape of plots  Nested plots  efficient for regenerating forests with trees growing into new size classes  Plots can be either circular or square  ~10 stems per strata ‘rule of thumb’ to determine plot size  Single size plots  Requires lower expertise  can be efficient where trees will be planted and will be single-aged Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

42. Sampling Design: Shape of plots

43. Sampling Design: Carbon pool measurement methods  Develop standard methods to collect carbon pool measurements  Develop standard methods to calculate carbon stocks, according to approved methodology Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

44. Sampling Design: Carbon pool measurement methods  Project must verify existing allometric equations or volume and BEF equations  Can use allometric equations for one stratum, volume+BEF for another stratum  Alternatively, may need to create local equations Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

45. Frequency of measurement  For CDM verification and certification must occur every five years  It is therefore logical to re-measure at this time  However, for slowly changing pools such as soil it will be necessary to measure less frequently Define Project Boundary Stratify Which carbon pools? Sampling design Measurement frequency

46. Field Measurement Techniques

47. Carbon vs Biomass Generally: Carbon = 50% of Biomass

48. Create Plot  Install permanent measuring and monitoring plots in a standard design  Permanently mark plot center and locate with a GPS Plots marked with rebar and PVC, trees marked with aluminum nails and tags

49. Estimate carbon pools - tree biomass  In each strata, measure DBH of appropriate size trees  DBH measured at 1.3 m above the ground with a DBH tape

50. Estimate carbon pools - tree biomass – Allometric equations  Local regression equations in literature  Verify applicability of equation(s) through limited destructive sampling or volume and wood density estimations  If no equations exist, need to harvest and measure a representative sample of trees to develop equations

51. Estimate carbon pools - tree biomass– Allometric equations r 2 = 0.99 N = 454 (Schroeder et al. 1997)

53. Estimate carbon pools – Dead wood  Dead wood can be a significant component of biomass pools  Particularly in mature forests – not eligible in first reporting period  For standing dead trees estimate biomass using regression equations or volume from detailed measurements

54. Estimate carbon pools– Understory/herbaceous vegetation  Use small frames  Frame placed on ground  Cut all herbaceous vegetation, remove leaf litter, within the frame  Place in new location on repeat measurements Aluminum or PVC frame of ~60 cm 2 is placed on the ground

55. Estimate carbon pools - Mineral soil carbon  Expose mineral soil surface  Collect 4 samples, mix and sieve for C analysis  Collect samples for bulk density in each plot

56. Estimate carbon pools - Mineral soil carbon  How deep to sample?  trade-off among magnitude of change expected, detectability of change, precision, and cost  How many samples?  relates to desired precision and cost  unlike trees where change can be measured directly (measure same trees through time), a new soil sample is collected each time  results in decreased precision of change in carbon stocks for soil

57. Monitoring Project Emissions  Tracking:  Vehicle and machinery use  Use of fertilizers  Use of fire  Incidental fires  Livestock emissions

58. Monitoring Leakage  All projects  Monitoring vehicle use  Some projects  Activity shifting  Tracking individuals  Tracking livestock  Tracking wood fuel use  All require good organization and good database skills

59. Minimizing Project Errors  Errors will arise through:  Poorly chosen methods  Poorly applied methods  Insufficient sampling  Errors can be predicted and minimized to decrease project costs and maximize claimable credits

60. Sources of error in estimating carbon pools  For estimating carbon stocks, three main sources of error are:  Sampling error—number and selection of plots to represent the population of interest  Measurement error —e.g. errors in field measurements of tree diameters, laboratory analysis of soil samples  Regression error — e.g. based on use of regression equations to convert diameters to biomass  All these sources can be quantified and “added”

61. Uncertainty  Two methods for determining uncertainty in estimates  Error Propagation  Monte Carlo Analysis (commercial software available)  Complex but should be used if there are correlations between datasets or if error (>100 %)  Correlations will exist between various measured carbon pools and between estimates at different times e.g.

62. Quality Assurance/Quality Control plans  Monitoring requires provisions for quality assurance (QA) and quality control (QC) to be implemented via a QA/QC plan  = How to guarantee that methods are applied correctly  The plan should become part of project documentation and cover the following procedures:  collecting reliable field measurements;  verifying methods used to collect field data;  verifying data entry and analysis techniques;  data maintenance and archiving.

63. QA/QC field measurements  Develop a set of Standard Operating Procedures (SOPs)  Thorough training of all field crews in procedures, followed by:  Hot Checks - supervisor visits crew in field and verifies measurements  Cold Checks - supervisor revisits plots after the departure of crew and reviews recorded measurements  Blind Checks - supervisor re-measures a proportion of plots with no knowledge of data recorded by crew

64. QA/QC for laboratory measurements, data entry, and archiving  Laboratory measurements  check equipment and measurement with known quantity samples added blindly  Data entry  test of out of range values  recheck proportion for errors  Archiving  off-site storage of data

65. Acknowledgements  Team at Winrock International  Including Sandra Brown, Ken MacDicken, David Shoch, Matt Delaney and John Kadyszewski  Funding agencies  Including TNC, USAID, USFS, UNDP and World Bank

66. Thank You!  For more information see:  http://www.winrock.org/Ecosystems/tools.asp http://www.winrock.org/Ecosystems/tools.asp  Or contact us:  tpearson@winrock.org tpearson@winrock.org  swalker@winrock.org swalker@winrock.org