REDD Methodological Module. Estimation of the baseline rate of unplanned deforestation

Transcription

1 REDD Methodological Module Estimation of the baseline rate of unplanned deforestation Version 1.0 April 2009 I. SCOPE, APPLICABILITY, DATA REQUIREMENT AND OUTPUT PARAMETERS Scope This module provides methods for estimating the annual rate (hectares per year) of unplanned deforestation in the baseline case. Applicability conditions This module is applicable for estimating the rate of unplanned conversion of forest land to non-forest land in the baseline case. The forest landscape configuration can be either mosaic or frontier. Data requirements Spatial data on historical deforestation and an analysis of agents and drivers of deforestation must be available to apply this module. This module calls upon the following other VCS-approved Modules and Tools: BL-UL LK-ASU Location and quantification of the threat of unplanned baseline deforestation - Version 1.0 Estimation of emissions from activity shifting for avoided unplanned deforestation Version 1.0 Output parameters This module provides methods to determine the following parameters: Parameter A BSL,RR,unplanned,t ha yr -1 A BSL,PA,unplanned,t ha yr -1 A BSL,LK,unplanned,t ha yr -1 SI Unit Description Annual area of unplanned baseline deforestation in the Reference Region at year t Annual area of unplanned baseline deforestation in the Project Area at year t Annual area of unplanned baseline deforestation in the Leakage Belt Area at year t 1

2 II. PROCEDURE The procedure is implemented by applying the following 8 steps: STEP 0. Selection of the procedure STEP 1. Definition of the project boundaries STEP 2. Analysis of historical deforestation STEP 3. Analysis of agents and proxy drivers of deforestation STEP 4. Selection of the most appropriate baseline approach STEP 5. Analysis of deforestation constraints STEP 6. Estimation of the annual areas of unplanned baseline deforestation in the reference region STEP 7. Analysis of the location of the risk of unplanned baseline deforestation STEP 8. Estimation of the annual areas of unplanned baseline deforestation in the project area and leakage belt area STEP 0. Selection of the procedure The REDD project activity may be located: 1. in a region for which no regional deforestation baseline has been projected yet (Scenario 1); or 2. in a region for which a regional deforestation baseline has already been projected by a third party (Scenario 2). If Scenario 1 applies: Steps 1 to 8 must be applied. If Scenario 2 applies: If the third party that determined the regional deforestation baseline is approved or sanctioned by the national or regional government then, the existing regional rate must be used, unless it is not applicable according to the criteria listed below, in which case Steps 1 to 8 of this module will apply. If the third party is not the national or regional government, project participants can decide not to use the existing regional baseline rate if they consider that it does not reflect the baseline circumstances expected to occur in the project area during the crediting period. In this case Steps 1 to 8 of this module will apply. An existing regional deforestation baseline rate is applicable under the following conditions: a) The baseline deforestation rate has been established for a reference region that includes the entire project area of the proposed REDD project activity. b) If the area for which the existing baseline rate has been projected is larger than the project area, the projected baseline must include the location of the expected baseline deforestation, so that the deforestation rate in the 2

3 project area can transparently be determined as per Step 8 of this module. If no location analysis exists, Steps 7 and 8 of this module must be applied. c) The existing baseline rate has been projected for at least 10 years into the future. If the regional rate has been determined for a fewer number of years than the crediting period, Steps 1 to 8 of this module must be used for the years of the crediting period for which the existing regional rate is not applicable. d) Methods used to project the baseline deforestation rate are transparently documented so that assumptions and data used to do the projections can be verified. This provision does not apply in case of baseline rates established by a national government having adopted a REDD scheme recognized by the UNFCCC or VCS. e) To use the existing regional deforestation baseline rate, it must be either: independently validated by a VCS accredited verifier, or is registered under a VCS acknowledged system, or has been established by the national or regional government having adopted a REDD scheme recognized by the UNFCCC or VCS; or it has been determined by an independent team and has been peerreviewed. If the previous two requirements are not satisfied, VCS verifiers must perform an independent validation of the existing regional deforestation baseline rate. STEP 1. Definition of the project boundaries The analytical domain from which information on the historical deforestation rate is extracted and projected into the future must be delimited by spatial and temporal boundaries. 1.1 Definition of the spatial boundaries of the analytical domain The boundaries of the following spatial features must be defined: Reference region Project area Leakage belt Forest For each spatial feature the criteria used to define their boundaries must be described and justified. Appropriate sources of spatial data should be used for each of these criteria, such as remotely sensed data, field information, and other verifiable sources of information. 3

4 Shape files, maps, GPS coordinates or any other location information that allows the identification of the boundaries unambiguously (< the minimum mapping unit of forest land) must be available for review by verifiers Reference region The boundary of the reference region is the spatial delimitation of the analytic domain from which information about regional rates and spatial patterns of deforestation are obtained, projected into the future and monitored. The reference region should be representative of the general patterns of unplanned deforestation that are influencing the project area and its leakage belt. The boundary of the reference region must be defined using criteria such as the following: a) Agents and drivers of deforestation that exist or are expected to exist within the project area during the project term must be of the same type of those found historically within the reference region 1. Areas where deforestation agents and drivers are dissimilar to those that will be found in the project area should not be included in the reference region. b) Landscape factors, such as vegetation type, soil fertility, slope, elevation, distance to navigable rivers and water bodies, etc. that increase the likelihood of deforestation must be similar in the reference region and project area. c) Human infrastructure, such as roads, sawmills, and settlements, that increase the likelihood of deforestation and that exist or are expected to exist within the project area during the project term must be of the same type of those found historically in the reference region. The following distinction must be made: No further infrastructure development is expected in or near the project area. In this case, the boundary of the reference region must be determined so that access conditions to forest land within the reference region and the project area are about similar. Further infrastructure development is expected in or near the project area. In this case, the reference region must include strata that are representative of the changes in access conditions that are likely to occur within or near the project area during the project term. For example, if there is a plan to build a new road or to improve an existing 1 For instance, if deforestation within the project area is linked to population growth of small farmers practicing subsistence agriculture and fuel-wood collection on land that is considered marginal for commercial agriculture, areas outside the project boundary that are subject to deforestation by large cattle ranchers and cash-crop growers should not be included in the reference region. However, if the forest land within the project boundary is suitable for deforestation agents that have not encroached into the project area historically (e.g. large cattle ranchers and cash-crop growers) but that may do so during the project term, then the reference region must include areas where such agents have been deforesting during the historical reference period. 4

5 road in or near the project area, the reference region should include strata where a similar road was constructed or improved during the historical reference period. If the latter condition is not feasible, then a proxy area, in addition to the reference area, that contains similar infrastructure as planned must be identified and included in the analysis. d) Policies and regulations having an impact on land-use change patterns within the reference region and the project area must be of the same type taking into account the current level of enforcement. For instance, in a country where subnational administrative units are governed by different forest institutions it would make sense to consider the boundary of the administrative unit as the reference region s boundary Project area The project area is the area or areas of land on which the project participants will undertake the project activities and that are forest land at the start date of the REDD project activity. Lands on which the REDD project activities will not be undertaken cannot be included in the project area Leakage belt Depending on the methods chosen to address activity displacement leakage, a leakage belt area may have to be defined in the surroundings or immediate vicinity of the project area. See the Module Estimation of emissions from activity shifting for avoided unplanned deforestation (LK-ASU) to decide whether a leakage belt is necessary and how it should be defined. If a leakage belt is defined, a baseline deforestation rate must be estimated for it using the procedures described in this module Forest To quantify deforestation it is necessary to define the boundary of forest land. This requires a clear definition of: Forest land; and of The Minimum Mapping Unit (MMU). A consistent definition of forest land 2 and MMU must be used during the entire project term. The MMU should be meet the minimum criteria defined in Module Methods for the monitoring forest cover changes in REDD project activities (M-FCC). An initial Forest Cover Benchmark Map, representing forest land at the earliest date of the historical reference period, is required to report only gross deforestation going 2 If the definition of forest will change in future periods, the baseline must be reassessed with the new definition. 5

6 forward. This map has to be updated at the starting date of each period analyzed and before each verification event. It should depict the locations where forest land exists. Areas covered by clouds or shadows and for which no spatially explicit and verifiable information on forest cover can be found or collected (using ground-based or other methods) shall be excluded permanently, unless it can reasonably be assumed that these areas are covered by forests (e.g. due to their location). 1.2 Temporal boundaries The following temporal boundaries must be defined (see also the REDD Methodology Framework REDD-MF): Start date and end date of the historical reference period The historical reference period is the temporal domain from which information on historical deforestation is extracted, analyzed and projected into the future. The starting date of this period should be about years in the past and the end date must be as close as possible to project start. Start date and end date of the REDD crediting period The crediting period is the period of time for which the net GHG emissions reductions or removals will be verified, which under the VCS is equivalent to the project lifetime. The project crediting period for REDD projects shall be between 20 and 100 years. Date at which the project baseline will be revisited The REDD project baseline must be revisited periodically. The date of the next revision must be specified and cannot be more than 10 years after the project start date. Duration of the monitoring periods The minimum duration of a monitoring period is one year and the maximum duration is 10 years. Project proponents are free to decide the periodicity of verifications. STEP 2. Analysis of historical deforestation This step is to quantify the historical deforestation rate during the historical reference period within the reference region and project area. This is performed by implementing the following sub-steps: 2.1 Collection of appropriate data sources 2.2 Mapping of historical deforestation 2.3 Calculation of the historical deforestation rate 2.4 Map accuracy assessment 2.1 Collection of appropriate data sources Collect the data that will be used to analyze deforestation during the historical reference period within the reference region. It is good practice to do this for at least 6

7 three time points, about 3-5 years apart to obtain sufficient data for calibrating and validating a deforestation model 3. For still intact forest areas, it is sufficient to collect data for one single date, which must be as close as possible to the present. As a minimum requirement: Collect medium resolution remotely sensed spatial data 4 (30m x 30m resolution or less, such as Landsat or Spot sensor data) covering the past 10 years or so. Collect high-resolution data from remote sensors (< 5 x 5 m pixels) and/or from direct field observations for ground-truth validation. Describe the type of data, coordinates and the sampling design used to collect them. In a Table, provide a summary of the data sources used. Where already interpreted data of adequate spatial and temporal resolution are available, with some caution 5 these can also be considered for posterior analysis. 2.2 Mapping of historical deforestation 6 Using the data collected in Step 2.1 divide the reference region into polygons 7 representing forest land and non-forest land at different dates in the past (Forest Cover Maps) as well as deforested land (Deforestation Maps) at different time periods in the past. Given the heterogeneity of methods, data sources and software, no specific methodology is prescribed for forest land and deforestation mapping. Refer to Module Methods for the monitoring forest cover changes in REDD project activities (M-FCC) for specific guidance on deforestation mapping. 2.3 Calculation of the historical deforestation rate This is required for the module Location and quantification of the threat of unplanned baseline deforestation (BL-UL). Guidance on the selection of data sources (such as remotely sensed data) can be found in Chapter 3A.2.4 of the IPCC 2006 GL AFOLU and in GOFC-GOLD. (2008), Section Appendix 2 gives an overview of present availability of optical mid-resolution (10-60m) sensors. Existing maps should be used with caution because they often do not provide documentation, error estimates, whether they were of the site or region in question or extracted from a national map, or whether they were obtained by change detection techniques rather than by static map comparison, etc. If data about historical LU/LC and/or LU/LC-change is already available, information about the minimum mapping unit, the methods used to produce these data, and descriptions of the LU/LC classes and/or LU/LC-change categories must be compiled, including on how these classes may match with IPCC classes and categories. Note: For the purpose of this module, mapping forest and non-forest land is sufficient. However, project participants may consider to divide these two classes in sub-classes representing different carbon densities, as long as such classes can be accurately mapped using the data collected in Step 2.1 and such mapping is useful for other methodology steps. Raster or grid data formats are allowed. 7

8 The outcome of the calculations must be the number of hectare deforested per each year of the historical reference period. Calculating the rate of deforestation when maps have gaps due to cloud cover is a challenge. If there are clouds in either date in question in the area for which the rate is being calculated, then the rate should come from areas that were cloud free in both dates in question. This should be estimated in % per year. Then, a maximum possible forest cover map should be made for the most recent date. The historical rate in % should be multiplied by the maximum forest cover area at the start of the period for estimating the total area of deforestation during the period. 2.4 Map accuracy assessment A verifiable accuracy assessment of the maps produced in the previous sub-step is necessary to produce a credible estimate of the historical deforestation rate 8. The minimum map accuracy should be 90% for both, the forest class and for the deforestation category. If the classification accuracy is less than 90%: Exclude from the definition of forest the forest sub-classes that are causing the greatest confusion with non-forest classes (e.g. initial secondary succession and heavily degraded forest may be difficult to distinguish from certain types of grassland or cropland, such as agro-forestry and silvo-pastoral systems not meeting the definition of forest ) and redo the classification. Repeat mapping and accuracy assessment until the required minimum map accuracy is achieved. If ground-truth data are not available for time periods in the past, the accuracy shall be assessed only at the most recent date, for which ground-truth data can be collected. In order to allow a consistent time-series of land use-change data to be produced, the methodology used to map historical forest cover and deforestation must be carefully documented as specified in Module Methods for the monitoring forest cover changes in REDD project activities (M-FCC). STEP 3. Analysis of agents and drivers of deforestation To assess whether the future rates of deforestation in the reference region and project area are likely to change compared to the rates measured in the previous step it is necessary to analyze the main groups of deforestation agents (farmers, ranchers, loggers, etc.), the drivers that motivate their land-use decisions, and their likely future evolution. Existing studies, expert-consultations, field-surveys and other verifiable sources of information can be used, as needed 9, to perform this analysis. 8 9 See Chapter 5 of IPCC 2003 GPG, Chapter 3A.2.4 of IPPC 2006 Guidelines for AFOLU, and Section of Sourcebook on REDD (GOFC-GOLD, 2008) for guidance on map accuracy assessment. See Angelsen and Kaimowitz (1999) and Chomitz et al. (2006) for comprehensive analysis of deforestation agents and drivers. 8

9 For each identified main agent group, provide the following information: a) Brief description of the social, economic, cultural and other attributes of the agent group that are relevant to understand the agent s motivations to deforest. b) Current and most likely future spatial distribution of the agent group within the reference region and the project area. c) Current and likely future development of the population size of the agent group in the reference region and project area. d) The key driver variables that motivate the agent group s decision to deforest, including explanations and references to verifiable and credible sources of information. e) Current and most likely future development of the identified driver variables, including references to verifiable and credible sources of information. The analysis of agents and drivers must conclude with a statement about whether the available evidence about the most likely future deforestation trend within the reference region and project area is: Inconclusive or Conclusive In case the evidence is conclusive state whether the weight of the available evidence suggests that the future baseline deforestation rates will be: decreasing about constant increasing STEP 4. Selection of the baseline approach Three baseline approaches are available: a) Historical average approach: Under this approach, the regional rate of unplanned baseline deforestation is assumed to be a continuation of the average annual rate measured during the historical reference period. In case of inconclusive evidence in Step 3, a discount factor will be used to allow conservative estimates. b) Linear extrapolation approach: With this approach, the regional rate of unplanned baseline deforestation will be estimated by extrapolating the historical trend using linear regression. c) Modeling approach: With this approach, the rate of unplanned baseline deforestation will be estimated using a model that expresses deforestation as a function of driver variables selected by the project proponents. Select and justify the most appropriate baseline approach following the decision criteria described below: 9

10 1. The regional deforestation rates measured in different historical sub-periods do not reveal any trend (decreasing, constant or increasing deforestation) and: 1.1. No conclusive evidence emerges from the analysis of agents and drivers explaining the different historical deforestation rates: use approach a and the lower boundary of the 90% confidence interval of the average historical rate Conclusive evidence emerges from the analysis of agents and drivers explaining the different historical deforestation rates: use approach c. 2. The regional deforestation rates measured in different historical sub-periods reveal a clear trend and this trend is: 2.1. A decrease of the regional deforestation rate and: Conclusive evidence emerges from the analysis of agents and drivers explaining the decreasing trend and making it plausible that this trend will continue in the future: use approach b. Conclusive evidence emerges from the analysis of agents and drivers explaining the decreasing trend but this evidence also suggests that the decreasing trend will change in the future due to predictable changes at the level of agents and drivers: use approach c. No conclusive evidence emerges from the analysis of agents and drivers explaining the decreasing trend: use approach a and the lower boundary of the 90% confidence interval of the average historical rate A constant regional deforestation rate and: Conclusive evidence emerges from the analysis of agents and drivers explaining the historical trend and making it plausible that this trend will continue in the future: use approach a. Conclusive evidence emerges from the analysis of agents and drivers explaining the historical trend and this evidence also suggest that the historical trend will change in the future due to predictable changes at the level of agents and drivers: use approach c. No conclusive evidence emerges from the analysis of agents and drivers explaining the historical trend: use approach a and the lower boundary of the 90% confidence interval of the average historical rate 2.3. An increase of the regional deforestation rate and: Conclusive evidence emerges from the analysis of agents and drivers explaining the increased trend and making it plausible that this trend will continue in the future: use approach b. If the future deforestation trend is likely to be higher than predicted with approach b, use approach c. 10

11 Conclusive evidence emerges from the analysis of agents and drivers explaining the increased trend but this evidence also suggests that the future trend will change: use approach a or develop a model (approach c ). No conclusive evidence emerges from the analysis of agents and drivers explaining the increasing trend: use approach a. STEP 5. Analysis of deforestation constraints This step only applies if the conclusion of Step 3 is that the regional rate of unplanned baseline deforestation is likely to be constant or increasing. If the conclusion was decreasing continue with Step 6. A continuation or increase of the annual deforestation rate compared to past trends is credible only if there are no increasing constraints to the further deforestation of the forest in the reference region. This will be the case if it can be demonstrated that forest land suitable for the conversion to non-forest land is still available in sufficient quantity. The following sub-steps perform this demonstration: 5.1 Identification of land-use constraints 5.2 Identification of forest land that is suitable for deforestation 5.3 Stratification in deforestation suitability classes 5.1 Identification of land-use constraints Identify the biophysical and infrastructural constraints (soil, climate, elevation, slope, distance to roads, etc.) that limit the geographical area where deforestation agents could expand their land-use activities in current forest land. Consider the constraints as they are perceived by the main groups of deforestation agents. The criteria and maps developed to model the spatial location of deforestation can be used in this step (see Module Location and quantification of the threat of unplanned baseline deforestation - BL-UL) 5.2 Identification of forest land that is suitable for deforestation Using the constraints identified in Step 5.1, map the forest land that is suitable for the further expansion of non-forest land in the reference region. If the area that is suitable for conversion to non-forest land is more than 100 times the average area annually deforested within the reference region during the historical reference period, conclude that there is no constraint to the continuation of deforestation and continue with Step 6; otherwise continue with Step Stratification in deforestation suitability classes The methodology assumes that deforestation happens first in optimal areas and that in these areas it can continue at the historical or even higher rate. Once all optimal areas are exhausted, 11

12 Using the constraints identified in Step 5.1, define criteria and thresholds that delineate optimal, average and sub-optimal 11 conditions for each of the main land-uses implemented by the main agent groups (e.g. by defining ranges of distance to road, slope, rainfall, etc). Select thresholds that are relevant from the point of view of the deforestation agents. Using the selected criteria and thresholds stratify the Maximum Potential Deforestation Map in three broad suitability classes representing optimal, average and sub-optimal areas for non-forest uses. When available from other sources, use existing maps. Calculate the following parameters: A optimal A average A sub-optimal Area of optimal forest land suitable for conversion to nonforest land within the reference region; ha Area of average forest land suitable for conversion to nonforest land within the reference region; ha Area of sub-optimal forest land suitable for conversion to non-forest land within the reference region; ha STEP 6. Estimation of the annual areas of unplanned baseline deforestation in the reference region The method to be used depends on the baseline approach selected. Approach a : Historical deforestation 1) Estimation of the average regional rate of unplanned baseline deforestation that applies to the reference region during the first T optimal years: A BSL,RR,unplanned,t = A RR,unplanned,hrp / T hrp * DF (1) A BSL,RR,unplanned,t A RR,unplanned,hrp T hrp DF t Regional rate of unplanned baseline deforestation in the reference Total area deforested during the historical reference period in the reference region; ha Duration of the historical reference period in years; yr Discount factor (0.5 the confidence interval of the mean historical deforestation in case of inconclusive evidence about future deforestation trends and 1.0 in case of conclusive evidence). A year of the proposed project term. 11 deforestation will become slower as only average and sub-optimal areas will remain available. When all sub-optimal areas have been cleared, deforestation must be set to zero. More or different suitability classes can be used, depending on the information that is available and the specific project circumstances. 12

13 2) Determination of the number of years during which the calculated regional rate of unplanned deforestation is applicable (T optimal ): T optimal = A optimal / A BSL,RR,unplanned,t (2) T optimal A optimal Number of years since the start of the REDD project activity in which A optimal is deforested; yr Area of optimal forest land suitable for conversion to non-forest land within the reference region; ha A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference If: If: T optimal > Project term: The regional rate calculated with Equation 1 is applicable during the entire project term. T optimal < Project term: The regional rate calculated with Equation 1 is applicable only to the first T optimal years. For the following T average years use A BSL,RR,unplanned,t * 0.5 3) Determination of T average : T average = A average / (A BSL,RR,unplanned,t * 0.5) (3) T average A average Number of years in which A average is deforested; yr Area of average forest land suitable for conversion to non-forest land within the reference region; ha A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference If: If: T optimal + T average > Project term: After T optimal years since the start of the REDD project activity and until the end of the project term the regional rate of unplanned baseline deforestation will be A BSL,RR,unplanned,t * 0.5. T optimal + T average < Project term: For T sub-optimal years after T optimal + T average years since the start of the REDD project activity use A BSL,RR,unplanned,t * After T optimal + T average + T sub-optima years the regional rate of unplanned baseline deforestation will be zero. 4) Determination of T sub-optimal : T sub-optimal = A sub-optimal / (A BSL,RR,unplanned,t * 0.25) (4) 13

14 T sub-optimal A sub-optimal Number of years in which A sub-optimal is deforested; yr Area of sub-optimal forest land suitable for conversion to nonforest land within the reference region; ha A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference Approach b : Linear extrapolation 1) Estimation of the regional rate of unplanned baseline deforestation in the reference region at year t during the first T optimal years: A BSL,RR,unplanned,t = a + b * t (5) A BSL,RR,unplanned,t a Regional rate of unplanned baseline deforestation in the reference Estimated intercept of the regression line; ha b Estimated coefficient of the time variable; ha yr -1 t A year of the proposed project term. 2) Determination of the number of years during which the calculated regional rate of unplanned baseline deforestation is applicable (T optimal ): If: b < 0 If: b > 0 T optimal is the period of time during which Equation 5 yields a positive value. After that period of time, A BSL,RR,unplanned,t = 0. T optimal is the period of time between t=1 and t=t optimal, the latter being the year at which the following condition is satisfied: A optimal,,, (6) Area of optimal forest land suitable for conversion to non-forest land within the reference region; ha A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference t optimal Year at which T optimal ends; yr 14

15 If: If: T optimal > Project term: The regional rate calculated with Equation 5 is applicable during the entire project term. T optimal < Project term: The regional rate calculated with Equation 5 is applicable only to the first T optimal years. For the following T average years use the following equation: A BSL,RR,unplanned,t = a + b * t optimal (7) A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference a Estimated intercept of the regression line; ha b Estimated coefficient of the time variable; ha yr -1 t optimal Year at which T optimal ends; yr 3) Determination of T average : T average is the period of time between t = t optimal and t = t average, the latter being the year at which the following condition is satisfied:,,, (8) A average Area of average forest land suitable for conversion to non-forest land within the reference region; ha A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference t optimal t average Year at which T optimal ends and T average starts; yr Year at which T average ends, yr If: If: T optimal + T average > Project term: The regional rate calculated with Equation 7 is applicable during the period of time between t = t optimal and t = t average. T optimal + T average < Project term: The regional rate calculated with Equation 7 is applicable only to the first T average years following T optimal. For the following years use the following equation: A BSL,RR,unplanned,t = a - b * t (9) 15

16 A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference a Estimated intercept of the regression line; ha b Estimated coefficient of the time variable; ha yr -1 t A year of the proposed project term Note: If A BSL,RR,unplanned,t calculated with Equation 10 is < 0, use A BSL,RR,unplanned,t = 0. Approach c : Modeling The regional rate of unplanned baseline deforestation in the reference region is estimated using a statistical model, such as a multiple regression: A = a + b V + b V b BSL. RR, unplanned, t n i n (10) A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference a; b 1 ; b 2 ;... ; b n Estimated coefficients of the model V 1 ; V 2 ;...;V n t Variables included in the model V A year of the proposed project term. The model and its rationale must be explained by the project participants using logical arguments and verifiable sources of information. Typically, the regional rate of unplanned baseline deforestation will be modeled as a function of several independent variables (e.g. population density, gross regional product, exports, agricultural product prices, etc. consistently with Step 3). Good historical data and credible future projections are required for each of these variables 12. To avoid an overestimation of the deforestation rate, variables that constrain deforestation (such as limited availability of forest land that is suitable for conversion to non-forest land as per Step 5) should be included in the model. Several models based on different combinations of independent variables should be tested. The model with the best fit with historical deforestation data should be used. No specific procedures are prescribed here to validate a model or test its goodness of fit with historical data 13. However, project proponents must demonstrate that such 12 To determine the future values of the variables included in the model use official projections, expert opinion, other models, and any other relevant and verifiable source of information. Justify with logical and credible explanations any assumption about future trends of the driver variables and use values that yield conservative estimates of the deforestation rate. 13 If sufficient historical data are available, divide them in two sub-periods: an earlier calibration period and a following validation period. Use the data of the calibration period to estimate the 16

17 tests have been realized and that a credible and conservative model has been chosen to project the baseline rate of unplanned deforestation. Seek assistance from an expert statistician as necessary. STEP 7. Analysis of the location of the risk of unplanned baseline deforestation To perform this analysis, project proponents shall use Module Location and quantification of the threat of unplanned baseline deforestation (BL-UL). STEP 8. Estimation of the annual areas of unplanned baseline deforestation The project rate of unplanned baseline deforestation in the Project Area is estimated as follows: A BSL,PA,unplanned,t = A BSL,RR,unplanned,t * P PA,t (11) A BSL,PA,unplanned,t Project rate of unplanned baseline deforestation in the Project Area at year t; ha yr -1 A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference P PA,t t Proportion of the regional area deforested in year t that is within the boundary of the Project Area at year t; % A year of the proposed project term. The annual area of unplanned baseline deforestation in the leakage belt is estimated as follows: A BSL,LK,unplanned,t = A BSL,RR,unplanned,t * P LK,t (12) A BSL,LK,unplanned,t Project rate of unplanned baseline deforestation in the Leakage Belt Area at year t; ha yr -1 A BSL,RR,unplanned,t Regional rate of unplanned baseline deforestation in the reference P LKt Proportion of the regional area deforested in year t that is within the boundary of the Leakage Belt Area at year t; % coefficients of the model and then evaluate how well each model predicts the deforestation observed for the validation period. Select the model that yields the best fit. If two models yield about the same accuracy, the most conservative one should be chosen (the one that predicts less deforestation). In selecting the final model, also consider the conceptual logic underlying the choice of the independent variables. Once the best model has been chosen, recalculate its coefficients using data from the entire historical reference period (calibration + validation period). 17

18 t A year of the proposed project term. Refer to the VCS-approved Module Estimation of emissions from activity shifting for avoided unplanned deforestation (LK-ASU) to estimate the area of the Leakage Belt and calculate P LKt. 18

19 III. Data and parameters used and generated in this module Data/parameter Unit Used in equations Descripiton a ha 5, 7, 9 Estimated intercept of the regression line a; b 1 ; b 2 ;...; b n number 10 Estimated coefficients of the model Source of data Measurement procedure (if any) Comments A average ha 3, 8 Area of average forest land suitable for conversion to non-forest land within the reference region A BSL,LK,unplanned,t ha yr Project rate of unplanned baseline deforestation in the Leakage Belt Area at year t A BSL,PA,unplanned,t ha yr Project rate of unplanned baseline deforestation in the Project Area at year t A BSL,RR,unplanned,t ha yr -1 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,11, 12 Regional rate of unplanned baseline deforestation in the Reference Region at year t A optimal ha 2, 6 Area of optimal forest land suitable for conversion to non-forest land within the reference region A RR,unplanned,hrp ha 1 Total area deforested during the historical reference period in the reference region A sub-optimal ha 4 Area of sub-optimal forest land suitable for conversion to non-forest land within the reference region b ha yr -1 5, 7, 9 Estimated coefficient of the time variable 19

20 Data/parameter Unit Used in equations Descripiton Source of data Measurement procedure (if any) Comments P LK,t % 12 Proportion of the regional area deforested in year t that is within the boundary of the Leakage Belt Area at year t P PA,t % 11 Proportion of the regional area deforested in year t that is within the boundary of the Project Area at year t t 1, 5, 9, 10, 11, 12 1, 2, 3 t * years elapsed since the start of the REDD VCS project activity t average yr 8 Year at which T average ends T average yr 3 Number of years in which A average is deforested T hrp yr 1 Duration of the historical reference period in years T optimal yr 2 Number of years since the start of the REDD project activity in which A optimal is deforested t optimal yr 6, 7, 8 Year at which T optimal ends T suboptimal yr 4 Number of years in which A sub-optimal is deforested V1; V2;...;Vn 10 Variables included in the model 20