Forests in the Balance Sheet: Lessons from Developed Country Land Use Change and Forestry Accounting and Reporting Practices

Florence Daviet and Lauren Goers, with Kemen Austin

World Resources Institute Working Papers and forest degradation (REDD). contain preliminary research, analysis, In order to ensure that a REDD mechanism is both effective and findings, and recommendations. They are credible in protecting and restoring forests and reducing carbon circulated to stimulate timely discussion and dioxide emissions, a range of accounting and methodological critical feedback and to influence ongoing challenges will have to be solved. In assessing the scale of those debate on emerging issues. Most working challenges many commentators have looked to pilot projects. papers are eventually published in another However, while instructive, these projects do not get at the form and their content maybe revised. larger issues involved in national accounting. It is interesting therefore to look for lessons from the experience of Annex I Acknowledgements (developed) countries in accounting for forest-related A special thanks to Peter Popatov, Andrea emissions and sequestration as part of their national emission Johnson, Fred Stolle, Lars Laestadius, Kelly reduction commitments. This working paper thus analyzes Levin and Polly Ghazi for their help with this developed country experience to date in relation to paper. The views in this paper are the implementation of the LULUCF (land use, land use change and authors’. forestry) provisions of the .

December 6, 2009 Key Findings The paper draws conclusions and recommendations around Suggested Citation: Daviet, F, et al. 2009. Forests in the including the GHG emissions and removals from developing Balance Sheet: Lessons from Developed Country Land Use country forests and other relevant land uses into an Change and Forestry Greenhouse Gas Accounting and international climate architecture. We hope this information Reporting Practices . WRI Working Paper. Washington DC will inform the negotiations both at and beyond the Copenhagen Conference of the Parties to the UNFCCC in December 2009. Lessons learned from Annex I LULUCF Executive Summary accounting point to the need to design an international REDD The world’s forests, both their use and loss mechanism that differs significantly from the current LULUCF have a critical role for international efforts model. to counter . Recognizing this, Parties to the United Nations Framework Lesson 1: Current LULUCF rules have resulted in the significant Convention on Climate Change (UNFCCC) under-reporting of forest-related . included in the 2007 Bali road map a This is because Annex I countries are only required to report on mandate to develop a mechanism that greenhouse gas emissions from deforestation and greenhouse would create incentives for developing gas removals due to afforestation and . countries to reduce emissions from Negotiations for the post-2012 arrangements are debating whether to make inclusion of emissions and removals from

WORLD RESOURCES INSTITUTE • 10 G Street, NE • Washington, DC 20002 • Tel: 202-729-7600 • Fax: 202-729-7610 • www.wri.org

Forests in the Balance Sheet 2 forest management activities mandatory for change and emissions from forest activity. In addition, Annex I countries. Initial research suggests reporting is scant on the biodiversity and human welfare that if only deforestation is considered for impacts of policies designed to reduce emissions. Both of these developing countries, the same under- factors reduce the credibility of, and support for, country reporting that developed countries have actions. experienced will result. Recommendation 3: Parties should undertake full and Recommendation 1: To account more consistent reporting of all relevant data for quantifying accurately for forest-related emissions and emission reductions or greenhouse gas removals. This includes removals, reporting requirements of a data on the basis for setting a reference level and on non- forest mechanism for both developed and carbon factors, such as biodiversity and impacts on forest developing countries should be broader stakeholders, specifically related to tenure. This reporting than the current mandatory requirements should include actions as well as emissions data and under LULUCF. assumptions.

Lesson 2: Current accounting systems under Lesson 4: Inventories of greenhouse gas emissions have been LULUCF rules do not yet provide the critical in helping countries understand where their most accuracy and precision needed for credibly significant land use emissions come from. They also help including land use emissions in emission identify “leakage” between land use activities. trading. Annex I countries are actively seeking to improve quantification and Recommendation 4: Any country that engages in voluntary accounting for emissions and removals in REDD Plus activities as part of its commitments under an relation to forest management activities international climate agreement should undertake a full land under LULUCF, but this has proven use emissions inventory. In accordance with the Convention, challenging. These difficulties will also exist support should be provided to developing countries to produce in a forest mechanism that includes these inventories. activities such as degradation, sustainable management of forests and enhancement of carbon stocks. Such imprecise Section A. Introduction accounting may prove problematic for Context effective functioning of a carbon market Reducing greenhouse gas (GHG) emissions and increasing GHG mechanism. removals associated with forest use activities, combined with reducing land conversion from forests to other uses, could Recommendation 2: A system that relies represent a significant piece of the climate mitigation puzzle if wholly on carbon markets to finance REDD undertaken credibly. Current emissions from land use change would greatly sacrifice accuracy and/or and some forest activities that reduce tree cover are estimated inclusiveness among countries. Parties to account for 12-17 percent of total global emissions (IPCC should implement a range of means for 2007; van der Werf et al. 2009). Many of these emissions occur financing REDD. This should include a fund in developing countries, and their reduction has been proposed to finance activities that cannot be precisely as a cost-effective approach that could be supported by accounted for. In some cases changing the developed countries. definition of deforestation in order to capture significant land conversion activities While many expect forward movement on these discussions in may address some of the accuracy issues. Copenhagen, two fundamental questions remain:

Lesson 3: Significant improvements in a) Where will the financing come from? A market or a fund transparency and consistency are needed in are regarded as the two basic options. the way countries report on land use

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 3 b) What activities will be included in the from deforestation and forest degradation in developing final mechanism? Deforestation, forest countries; and the role of conservation, sustainable degradation, sustainable management management of forests and enhancement of forest carbon of forests, conservation of forests, stocks in developing countries” (UNFCCC 2007; see Decision enhancement of carbon stocks, 1/CP 13, p 3). Such a mechanism has been termed “REDD afforestation and reforestation are all Plus.” activities that have been mentioned in the plethora of documents that the However, as negotiators work out how to make a forest Parties have issued since Bali. mechanism for developing countries operational, many of the same issues of emissions estimation and accounting rules that While seemingly not related, these pertain to Annex I countries will become relevant to developing questions are closely linked. The decision to country activities. As a result, a key part of the discussion will use a market mechanism—such as a carbon be to consider which elements of the existing LULUCF rules and market—requires that emission reductions quantification guidance might be relevant for the new under the mechanism can be quantified mechanism. precisely enough to ensure the credibility of the mechanism, since Annex I Parties will be Depending on the final scope of the forest mechanism that allowed to increase their emissions above emerges from the climate negotiations, different LULUCF rules their emissions cap by the amount of and experiences may be more or less relevant. Options under offsets purchased through the market. A discussion for the forest mechanism range from a narrow fund approach, on the other hand, could “RED” scope that would address only emissions from include activities where the quantification deforestation activities, to the broadest “REDD Plus” scope of the exact emission reduction is less incorporating forest degradation, the role of conservation, precise, but which would create a sustainable management of forests and enhancement of forest mechanism that more accurately captures carbon stocks. In addition, following negotiations in Barcelona what goes into the atmosphere. in November 2009, the latter approach might also include afforestation and reforestation initiatives by developing To date, negotiations among parties to the countries (UNFCCC 2009). UN Framework Convention on Climate Change (UNFCCC) around land use and About This Report forest issues have proceeded on two This working paper seeks to inform the Copenhagen separate courses. On the one hand, rules negotiations by trying to capture the main lessons from the are being negotiated for how developed LULUCF accounting and reporting rules that are likely to be (Annex I) countries should account for their relevant for a decision on REDD Plus. Our analysis focuses on GHG emissions resulting from land use, the following issues: land-use change and forestry (LULUCF). On the other, negotiators are discussing how to 1. Estimation uncertainty and inaccuracies resulting from: develop a forest mechanism1 that would a. Definitions of activities and types of land use. provide policy approaches and positive b. Data requirements. incentives “relating to reducing emissions c. Reference levels. 2. Clear and credible reporting of changes in behavior and 1 This document uses the term “forest their broader impacts. mechanism” as a neutral term that does not make any assumptions about the scope of activities Looking forward, the paper considers options that could included in the final agreement on a positive incentive mechanism for developing countries to provide a more complete interim solution than addressing reduce GHG emissions and increase GHG removals emissions from deforestation only, should such an approach be due to forest conversion and forest use activities – taken. In the event that a broader, “REDD Plus” approach is a mechanism often called REDD Plus.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 4 adopted, the paper concludes that limited to emissions from deforestation only (RED), may be alternatives to existing LULUCF accounting important for the following reasons: It would enable credible and reporting practices should be measurement of emission reductions, allow reporting and considered so that REDD Plus can be a more verification (MRV) of emission reduction actions with high credible and manageable mechanism. levels of certainty, and focus financing where significant emissions are occurring. To address the range of possible outcomes, the remainder of this paper is divided into A decision to focus the scope on emissions from deforestation three broad sections. The first section only would mimic the decision by Annex I countries during the focuses on lessons from LULUCF if a narrow Kyoto Protocol negotiations to limit mandatory accounting of focus on reducing emissions from emissions from LULUCF to deforestation only (see Decision deforestation only is adopted by Parties to 9/CP.4 on Article 3.3). the UNFCCC. The next section considers lessons from the LULUCF discussions on However, there would be some differences in applying this “forest management” in an attempt to approach to a forest mechanism for developing countries. One identify potential concerns for REDD Plus if important difference might involve the inclusion in the a broader scope is adopted. The final accounting of GHG removals due to afforestation and section provides conclusions and reforestation. While the Kyoto Protocol accounts for such recommendations. removals as part of Article 3.3, such activities might not be included in a RED mechanism because positive incentives to Each section assesses the treatment of the undertake these actions are already included in the UNFCCC’s topic under current LULUCF rules and Clean Development Mechanism (see Figure 1). However, if practices, and then establishes potential “net” rather than “gross” emissions from deforestation are lessons regarding the application of such being considered for REDD Plus, then Parties will need to rules for a developing country forest decide what should happen to the afforestation and mechanism. We conclude each section with reforestation provisions in the Clean Development Mechanism recommendations for the new mechanism. (see the discussion about this in Box 1). Given that many of these issues are heavily intertwined, some of the sections may seem artificially distinct. We have taken this approach in an attempt to keep the discussions relatively clear, and the indulgence of the reader is requested. Concerns about LULUCF have also been raised in other papers, and we point readers to those papers that may be useful to consider.

Section B. A Narrow Forest Mechanism: Considerations and Concerns Figure 1: Potential Similarities between LULUCF and REDD Plus Many UNFCCC Parties and stakeholders Accounting Rules have noted that a “narrow” forest mechanism for developing countries,

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 5

Box 1. Key Definitions Discussing the potential scope for REDD Plus is difficult as the elements or activities included in REDD Plus have been shifting. After Bali, REDD Plus was thought to mean, “Reduced Emissions from Deforestation and Forest Degradation; the role of conservation, sustainable management of forests and enhancement of carbon stocks.” Subsequent decisions and deliberations have seen the semi colon changed in 1b (iii) of Decision 1/CP 13 into a comma, and added additional activities such as afforestation and reforestation. As none of the terms in the “REDD Plus” sentence have themselves been defined for this mechanism, this creates additional confusion. Bearing in mind that definitions may change over time, this paper tries to identify issues that may occur when current LULUCF definitions are used, as well as the risk of LULUCF definitions being applied to REDD Plus, given the similarities in the types of activities that are being proposed. The following definitions are taken from: a) Decision 16/ CMP 1, which adopted the definitions defined in Marrakesh; b) the IPCC Good Practice Guidelines (2003) that served to interpret and make operational some of the language in the Marrakesh. Forest: Land of a minimum area of 0.05-1.0 ha with tree crown cover (or equivalent stocking level) of more than 10- 30 per cent with trees with the potential to reach a minimum height of 2-5 m at maturity. A forest may consist either of closed forest formations, where trees of various stories and undergrowth cover a high proportion of the ground, or open forest. Young natural stands and all plantations which have yet to reach a crown density of 10-30 per cent or tree height of 2-5 m are included under forest, as are areas normally forming part of the forest area which are temporarily unstocked as a result of human intervention (such as harvesting or natural causes), but which are expected to revert to forest. This definition of forest is essentially a minimum “tree cover” definition, defined by the minimum height and area and crown cover. This is often confusing for people who think of forests more as ecosystems and therefore would define a forest and a plantation differently, even if the “tree cover” is similar. Deforestation: The direct human-induced conversion of forested land to non-forested land. This emission activity definition, along with the forest definition, means that emissions from significant tree cover change and conversion from a natural forest ecosystem to a plantation area are not captured. The temporary destocking rule in the forest definition means that the immediate emissions from clearing forest areas are not captured when countries account for the emissions from this activity, as long as the tree cover is expected to return to the “forest” threshold. Forest Management: A system of practices for stewardship and use of forest land aimed at fulfilling relevant ecological (including biological diversity), economic and social functions of the forest in a sustainable manner. This emission activity definition means that all emissions or removals in areas that can be described as ‘forest management’ areas are supposed to be calculated using the “forest remaining forest” quantification methods in the IPCC 2003 Guidelines. How strictly this definition is applied by countries in defining the areas to be captured means emissions in forest areas that are not caused by deforestation – e.g., unplanned harvesting or fires – may be overlooked. Managed Forests: All forests subject to some kind of human interactions (notably commercial management, harvest of industrial round-wood (logs) and fuelwood, production and use of wood commodities, and forest managed for amenity value or environmental protection if specified by the country), with defined geographical boundaries. (Definition from IPCC Good Practice Guidance: Penman et al. 2003a) This introduction of the term “managed forests” was to help countries define the “forest management” area where emissions and removals would be considered as anthropogenic. Emissions and removals in areas outside the “managed forest” area are considered non-anthropogenic.

Box Continued on next page WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 6

Box 1 Continued In addition, other terms are used when discussing REDD Plus which have no formal definitions. Some of these may overlap. Although not all of these terms are used consistently, the following are some of the main ones that are relevant: Forests Remaining Forests: This term is used to describe areas where emissions and removals will be calculated to capture “periodic or ongoing human interventions and that they include the full range of management practices from commercial timber production to stewardship in noncommercial purposes” (Penman et al. 2003a) but do not result in a change in the definition of the area as “forest”. Gross Deforestation/ Gross Emissions from Deforestation: These terms are often defined in different ways. Some use them to refer simply to the immediate impacts of tree cover change– either in terms of area or emissions. For others, they refer more specifically to the total area of conversion of natural forests, whether to another type of forest or another land use (e.g., agriculture). For yet others, they only represent the emissions from areas where tree cover has been diminished below the forest definition threshold for the foreseeable future. Net Deforestation, Net Emissions from Deforestation: There are several interpretations of these terms. One is that they represent the deforestation definition list above. The other is that it includes afforestation and reforestation as well as deforestation as defined for Annex I countries. Degradation: There are several potential definitions discussed, including in an IPCC 2003 document on degradation for Annex I accounting. One of those definitions is “The overuse or poor management of forests that leads to long-term reduced biomass density (carbon stocks).” (Penman et al. 2003b) The Role of Conservation of Forests: Many opinions exist on the potential link between this activity and carbon. Some believe this is linked to accounting for carbon stocks, others think it defines activities that would be helpful to reduce emissions from deforestation and forest degradation.

Sustainable Management of Forests: There are numerous definitions, although few related to emissions and removals. The following description was one offered by an Indian delegate in Accra – linking the activity to “forest remaining forest” quantification, and potentially afforestation or reforestation (India 2008): - Improvement in existing [forest] cover - Increase in forest/tree cover

Enhancement of Carbon Stocks: This is also a term that has, to date, been poorly defined. There appear to be at least two potential interpretations: Afforestation/ Reforestation, or increasing forest stocks in forests remaining forests

Lessons from LULUCF number of experts (Schlamadinger et al. 2007a). They note that a. Emissions are Underestimated the current UNFCCC definition of “forest” is very broad, including everything from largely undisturbed complex forest The impact of the decision by Annex I ecosystems to intensely managed monoculture plantations, or countries in the Kyoto Protocol negotiations clear cut areas where the intent is to allow trees to re-grow at to limit mandatory accounting of emissions some later date. In contrast, the definition of “deforestation” is from LULUCF to emissions from quite narrow and only covers conversion from forested to non- deforestation has been analyzed by a forested land. Together with the reference to temporary

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 7 destocking in the definition of forests, this through deforestation, such as hardwoods, are not recaptured terminology has an unfortunate, real world through afforestation/reforestation practices that focus on impact. It effectively enables large areas of softwood species (see Table 1). tree cover to be clear cut without counting the resulting greenhouse gas emissions, as One would expect that the impact of these choices would be long as the intent is the eventual return of captured in the quantification of emissions and removals under this area to some type of forest cover. the UN-led LULUCF accounting system, since hardwood species generally store more carbon (Brown and Schroeder 1999; b. The importance of GHG Inventories Neilson et al. 2007). However, country reporting may not incorporate all impacts unless very specific reporting and The extent to which emissions to the detailed data are available (see more on this problem in section atmosphere are underestimated by the C on quantification of emissions and removals). In addition, current LULUCF rules is only known today, these impacts may be drown out by the myriad of other however, as a result of Annex I countries emission and removal impacts in the inventory. developing national LULUCF inventories, even though they are not required to Table 1: Afforestation Practices in Annex I Countries account for all LULUCF emissions. Although Country Predominant / Predominant Secondary the quality of the data in some of these Secondary Species, % of Species, % inventories still needs improvement and the Species Planted Planted Area of Planted manner in which it is collected and reported Area is not always as transparent as could be Australia* Pine/Eucalyptus 44 25 hoped for (Schlamadinger et al. 2007b), New Pine/Fir 89 4.4 nevertheless these inventories provide all Zealand* Parties with a better sense of the key United Pine/Pine 65 16 sources of emissions and their relative size. States* United Spruce/Pine 35.6 10.9 This information has also provided the basis Kingdom* for discussions on how to account for GHG Canada** Spruce/ Pine 63 16 emissions from land use, land use change * Source: FAO. 2006. Global planted forests thematic study: results and and forestry in the next Kyoto Protocol analysis. Planted Forests and Trees Working Paper 38. Rome. commitment period. http://www.fao.org/forestry/plantedforests/10368/en/ ** Source: White, T.M. and W.A. Kurz. 2005. Afforestation on private land in Canada from 1990 to 2002 estimated from historical records. The c. More Data is Required to Reflect Non- Forestry Chronicle 81: 491-497. Carbon Impacts Applying LULUCF Lessons to REDD Plus Reviewing Annex I national inventory Looking at LULUCF, it is clear that accounting for deforestation reports, one clear concern is that the only–with the current rules–is not accurate. Not including impacts of the “plus” from afforestation forest management in the accounting results in overlooking and reforestation, compared to the “minus” significant emissions in forest areas that remain forests (see from emissions from deforestation, are not Box 1). However, if Parties decide to adopt this approach reported in a manner that allows several aspects of the current LULUCF rules could be adjusted stakeholders to understand the impacts on to address shortcomings regarding underestimating emissions biodiversity, which is one of the principles and reporting non-carbon impacts. Some initial agreed to by the Parties to Kyoto Protocol recommendations are provided below. (UNFCCC 2005). Only by gathering outside Undertaking the production of inventories however, is a information about planting practices (from precedent that non-Annex I (developing) countries should the FAO, for example), is it possible to ascertain that some species of timber lost

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 8 follow, as it will build their capacity for result in approximately 50 percent of all emissions from tree accounting for greenhouse gas emissions cover change in Indonesia—some 224 million tons of carbon and removals. dioxide (CO2) per year. These emissions, calculated in Table 2 below, would not be counted if current LULUCF forest a. Broadening the Definition of definitions were applied with a narrow REDD Plus scope looking Deforestation to Capture Significant only at emissions from deforestation. Emissions Table 2. Emissions from Non-Deforestation Tree Cover If LULUCF definitions of “forest” and Change in Indonesia “deforestation” are applied in the context of a forest mechanism where only Concession Tree Cover Assumed CO2 Emissions Annual CO2 emissions from deforestation are Type Change Carbon 2000 –2005 (t) Emissions (t) considered, significant emissions from 2000 – Loss* 2005 (ha) converting natural forests to intensely managed forest areas or plantations will Timber 671,125 195 t C / 479,767,128 95,953,425 not be captured. As a result, incentives to Plantation ha address emissions from land conversions Selective 247,625 100 t C / 90,779,325 18,155,865 will not be created. While the scale of this Logging ha problem will differ from country to Oil Palm 312,200 190 t C / 246,113,191 49,222,638 country, in places where conversion of ha natural forests to areas for massive timber extraction or plantations is the 2 Production 839,550 100 t C / 307,779,030 61,555,806 main driver of tree cover change , the Allocation ha overlooked emissions will be Total 2,070,500 1,124,438,674 224,887,734 considerable. *Based on assumptions of 254 t C / ha in intact forests in Indonesia Indonesia is illustrative of this problem. (Murdiyarso et al. 2002), 150 t C / ha in selectively logged areas (loss of 100 t Tree cover change is often the result of C / ha, Putz et al. 2008), 60 t C / ha in timber plantations (Van Nordwijk et al. conversion for timber plantation 2005), and 40 t C / ha in OP plantation (Henson 2009). development, oil palm plantation development, or selective logging (Curran Accounting for these emissions is possible – at least in an et al. 2004). So extensive is this practice approximate manner. As can be seen in Table 2, historical data that the results can be seen using could be used to create a reference emissions level. Such data inexpensive remote sensing technologies could be captured both by low-cost remote sensing methods like LandSat and Modis (Hansen et al. 2008). and by “intent” information extracted from paperwork granting An analysis using data on the allocation of forest concessions to various uses. concessions in Indonesia (FWI/GFW 2002), compared with data on tree cover change Without expanding beyond RED, the following three changes for the period 2000-2005 (Hansen et al. could enable more accurate emissions data capture: 2008), shows that these three activities 1. Remove the “temporary destocking” clause from the 2 The term “tree cover change” is used to talk deforestation definition, and include as deforestation about emissions from areas where trees are being either any significant clearing of forests (e.g. X% of cut down without consideration for the final use of forest cover and Y% area), or a shift in management the land. Knowledge of the intended or final use regime (e.g. from a natural forest to an intensively of forest land is currently required to make a managed forest). Such a change could ensure that all distinction between deforestation and forest significant activities would be captured. management in LULUCF rules, or deforestation and degradation in REDD plus discussions.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 9 2. Add forest type definitions and b. Making National Inventories Part of a Narrow Forest broaden the deforestation Mechanism Approach definition to include all “conversion” activities. The The Indonesia example demonstrates the urgent need to build addition of a simple “plantation” capacity and understanding about additional emission sources definition to the existing forest in order to ensure that accounting practices become definition would allow countries to increasingly complete and robust over time. However, the data capture large-scale conversion of collection and analysis required to develop GHG Agriculture natural forests – both to non-forest Forest and Other Land Use (AFOLU) inventories are not uses and to intensively managed inexpensive (Schlamadinger et al. 2007a). Regardless of what plantation uses. A minor the eventual forest mechanism will account for, non-Annex I modification to the “deforestation” countries will need financing assistance to build the capacity to definition would then allow these develop robust AFOLU/LULUCF inventories. major emission activities to be captured. c. Reporting Non-Carbon Impacts

3. Ensure that at least these At the national level it may not be possible to give funders and conversion activities are included buyers of emission reductions full information about the in an interim “degradation” benefits of activities that promote biodiversity and the rights approach, if a full degradation and participation of local communities and indigenous peoples approach is judged too difficult in in achieving RED. However, a system where some of this data is the near term due to data provided could be achieved by: limitations. The inclusion of at least these emissions that result from i. Identifying the types of forests – using FAO or large-scale “degradation” activities, nationally appropriate definitions – that have been in the short term, would likely be cleared (e.g., intact, primary, natural, intensively preferable to having a managed areas, etc.). “degradation” definition that could ii. Identifying changes in management practices and not be implemented in the near tenure to reduce deforestation. An example might be term. While ultimately these the granting of tenure in frontier forests to indigenous conversion activities do not reflect people for maintenance of non-forest product use, with the full complement of actions that very limited timber removal. may cause emissions in “forests iii. Identifying the types of forests that have been remaining forests” (i.e., forest areas replanted and how they will be managed. Examples that remain in tree cover after might include single-species plantations with clear significant forest activities have harvest intervals, or restoration of sensitive ecological occurred), they could be an areas with only non-timber forest products to be important first step. See Section B gathered. This approach will be most relevant if net for more discussion on degradation deforestation is included, or afforestation/reforestation and the data required. is brought under the umbrella of the forest mechanism.

These same practices should also be applied to LULUCF reporting.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 10 Section C. Considerations for a The challenge of high-quality data collection and analysis to credibly quantify carbon emission and/or Broader REDD Plus removal changes over time. Agreement The challenge of setting credible reference emission levels. For some Parties, a forest mechanism that The challenge of developing methods to credibly is limited to deforestation only, even if capture non-anthropogenic emissions. expanded to capture conversion of natural The challenge of reporting emission reductions and forests to intensively managed forest areas increased removals in a way that not only provides and plantations, does not fully capture the confidence that significant changes in behavior are range of activities they could take to reduce linked to the emission reductions and increased GHG emissions or increase carbon removals, but that also communicates the impacts that sequestration on their land base to mitigate this has had on biodiversity and stakeholders in the climate change. For this reason, countries country. Often, these considerations are as important such as India suggested additional actions to domestic stakeholders as the mitigation of such as “the enhancement of carbon emissions. stocks” and “sustainable management of forests“ for inclusion in the forest Figure 2: One Option for How REDD Plus Activities Add Up mechanism. As a result of such suggestions, the concept of REDD Plus was introduced by UNFCCC Parties in the 2007 Bali Action Plan (India 2007).

Although there are no formal definitions of REDD Plus “activities,” indicative definitions provided as part of IPCC guidance or contained in country submissions for the various activities such as “degradation,” “sustainable management of forests,” and “enhanced carbon stocks.” These resemble the “forest management” activities undertaken by Annex I countries under the LULUCF regime (see Figure 2). Like forest management, the accounting for these These challenges, taken together, have created major activities will most likely use the “forest difficulties for LULUCF countries in precisely quantifying and remaining forest” quantification guidelines accounting for emissions and removals. As a result, electing to contained in the IPCC 2003 or 2006 finance REDD Plus activities only through a market mechanism guidelines. Therefore, lessons drawn from approach may mean that activities such as degradation, the discussions about accounting for sustainable management of forests, and enhancement of emissions and removals from “forest carbon stocks might be difficult to include credibly as part of a management” activities initiated under forest mechanism for developing countries, at least in the near LULUCF are likely to be quite relevant to term. This is especially true as some developing countries have REDD Plus. less capacity and information at their disposal than developed countries. There are no easy answers to these challenges. In the following sections, the main However, by identifying them, the authors seek to raise the challenges facing LULUCF accounting and questions that negotiators should be asking to ensure that reporting practices are identified. These REDD Plus rules follow a more effective path. include:

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 11 Lessons from LULUCF Canadian National Inventory—an inventory that uses a Tier 3 approach, and provides significant information about data a. Data Considerations collection practices. Also Canada, like many of the potential developing countries participating in a forest mechanism has The first lesson learned from analyzing large expanses of relatively untouched forests. LULUCF rules and practices is that the importance of high-quality data cannot be Large or Small? Defining the Managed Area overestimated. In the LULUCF discussion, data, or its absence, have played a critical The “managed area” concept under LULUCF allows countries to role on a number of fronts. New data were exclude from their forest management accounting any important for helping industrialized emissions and removals that occur in areas outside direct countries to decide whether or not to elect human influence, and that therefore are likely non- forest management accounting rules for the anthropogenic (Penman et al. 2003a). The “managed land” first commitment period under the Kyoto concept was not approved as part of the Kyoto Protocol or Protocol (Kurz et al. 2008). In addition, lack related Decisions, but is part of the Good Practice Guidance for of data resulted in a decision by UNFCCC LULUCF implementation developed by the IPCC (Fry 2007). In Parties to limit the credits and debits that 2003, experts noted that depending on how the “managed would be accounted for from “forest forest” area is defined, significant emissions may be management” activities for any given Annex overlooked, for example, if the forest management definition is I country in the first commitment period taken too literally to include areas with active forestry activities (UNFCCC 2001; Kurz et al. 2008). Almost a only, rather than including all areas of forest where emissions decade later, concerns persist about both and removals could occur as the result of human impacts. data quality and the reference emission Conversely, if countries include areas devoid of management or levels that are created from that data. As a other human activity, it raises the question of whether carbon result, developing countries have proposed removals accounted for are simply the result of the natural that Annex I countries again have limited process of growth (Penman et al. 2003b; Fry 2007; Cadman access to credits from forest management 2008). activities undertaken domestically (PNG et al. 2009). Countries have taken different approaches in response to these challenges. Canada excluded a significant portion of its The second lesson is that it is imperative to northern, sparsely populated forests from its managed forest provide a clear understanding of the area (Environment Canada 2009) in its inventory, while challenges to gathering the type and quality Australia reports on its entire forest area in their inventory, of data needed to credibly quantify including areas with minimal human impacts (Government of 3 emissions and removals at the national Australia 2009) . In both cases uncertainty about the emissions level, namely data that records: and removals remain (Fry 2007; Cadman 2008), which speaks to problems of both precision and accuracy that are difficult to the area that the country will address. designate as “managed”; tree cover change ; This discussion about a lack of symmetry between the management practices ; and accounting of emissions and removals on the one hand, and carbon stocks and flows linked to what is anthropogenic or not on the other, has become even data that accurately reflect the management regime of the area.

3 The following subsections draw heavily As neither has elected to report “forest management” in the first from data and analysis provided around the CP, this may not exactly represent their final “managed area” definition.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 12

more acute in the context of rules for Table 3. Areas of Fires, Harvest, and Restocking in Canadian LULUCF in the second commitment period. Managed Forests (in 1000 ha) Countries have discussed factoring out Year Total Fire Area Harvested Total Total emissions that they believe may be non- Managed (Wild and Areas Immediate Planted anthropogenic even in managed areas Area, Controlled) Emission and (Government of Canada 2009), but seem Area Seeded less interested in ensuring that all removals Area included are anthropogenic. For some 2000 229865 170.4 1068.8 1239.1 479.1 countries, such as Canada, the asymmetry seems all the more stark, given that the 2001 229824 263.5 1015.3 1278.8 491.8 areas planted are much smaller than those 2002 229783 1251.4 1000.8 2252.2 469.8 harvested, and similar or smaller to those 2003 229741 823.2 995 1818.3 481.3 impacted by fires (see Table 3). 2004 229701 832.9 1010.6 1843.4 453.1

2005 229661 727.1 1057.7 1784.8 459.8 Tree Cover Change 2006 229613 639 823.9 1462.9 483.9 The ability to determine extensive tree 2007 229565 659.8 733.8 1393.6 485.7 cover loss using inexpensive, remote Source: Canadian National Inventory 2009 and National Forestry Data sensing technology is already globally Base. http://nfdp.ccfm.org/silviculture/national_e.php available (Hansen et al. 2008). Though remote sensing data is usually not Instead, they are relying primarily on data from forest considered sufficient for determining inventories and management plans, which may be incomplete. whether tree cover change equates to In Canada, for example, the authors of the latest national deforestation, it is viewed as a good initial inventory report note that the use of new harvesting data have step. Such data can also be extremely useful resulted in a significant shift in emission calculations: in helping countries that still contain large expanses of natural forests to record where “Carbon emissions associated with the removal of harvested wood tree cover change occurs in areas that were were recalculated upwards in the early years of the time series, by as previously relatively untouched by much as 43% in 1990 (over 12 Mt carbon). Areas harvested were deforestation or forest management recalculated downwards in the early 1990s, and upwards after 1999. (Mollicone et al. 2007). Overall, the amount of carbon transferred from forests to the forest product sector (treated as immediate harvest emissions in the Unfortunately, however, a review of the inventory) has increased significantly, primarily because of recalculations of the carbon harvested. The apparent trend in techniques used to collect tree cover harvesting activities has also been altered. Over the last 10 years change data for calculating forest harvesting activities accounted for annual average emissions of 184 management emissions and removals Mt, a 21% increase since 1990 levels; this contrasts with the previously shows that remote sensing data is not estimated 54% increase. (Environment Canada 2009, p 171-172).” commonly used by Annex I countries, except for some specific types of data. For In their discussion about uncertainty in the quantification of example, Canada uses spatially explicit fire forest management emissions and removals, the Canadian data gathered by remote sensing authors identify accuracy and completeness of management (Environment Canada 2009). data as long-standing issues, whose full impact in terms of uncertainty is unknown (Environment Canada 2009). Also, data The lack of remote sensing data, however, from management plans will not capture all emissions in means that industrialized countries may not countries where there is significant unapproved timber be capturing all emissions or removals.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 13 extraction or uncontrolled fires that might removals as well as changes from historical practices. be set. Remote sensing data could be used as a vital tool for checking the veracity of Carbon Stocks and Flows: Quantifying Emissions and Removals country management plan data and identifying potential emissions that have The experience of Annex I Parties’ collection of “forest been overlooked. remaining forest” inventory data underscores the challenges inherent in capturing necessary data to credibly quantify all Management Plans relevant emissions reductions and increased removals for forest management activity. For example, in a study of the As noted above, for Annex I countries the carbon stored in natural forests in Australia, researchers found use of management plan data is, in practice, that the data used in NCAS (the system used in Australia to at least as significant as remote sensing develop the national inventory, “underestimated the carbon data for determining the tree cover change carrying capacity of natural forests with high biomass stocks” likely to occur in managed forest areas, and (Mackey et al. 2008, p. 7) The researchers note that “NCAS was therefore for calculating the emissions from designed to model biomass growth in plantations and these activities. afforestation/reforestation projects using native plantings. The empirically based functions within NCAS were calibrated using Management plans generally indicate the data appropriate for that purpose. But, this meant that NCAS number of trees being removed, their was unable to accurately estimate the carbon carrying capacity species, site, and where any replacement of carbon-dense natural forests in south eastern Australia planting will occur (Environment Canada (Mackey et al. 2008, 7).” This has led to concerns by experts 2008). While there are clearly uncertainties and NGOs looking at the Australian inventory that emissions associated with using only management from native forest harvesting is under-reported. Such under- plans to capture emission and reporting is potentially significant due to assumptions made in sequestration that is occurring, they are the soil and litter carbon quantification methods used for very useful for determining historical native forests and, as a consequence, could affect the veracity management regimes and how they have of a country’s overall inventory for managed forest areas been changed to increase removals and (Mackey et al. 2008; Cadman 2009). sequestration, especially where practices may be too subtle to be captured by all but Such questions could also be raised in the context of other the most advanced remote sensing inventories, especially where large natural forest areas remain methodologies, e.g., single tree extraction and CO2 emission and removal assumptions are derived from as opposed to clear cutting practices. merchantable timber volume figures that may or may not accurately capture the carbon being stored in undisturbed soils. Additionally, remote sensing practices to data have not focused as much on capturing As can be seen in Tables 4 and 5, the variation that can occur removals – and changes in rates of even in areas that are being actively managed can still be quite removals. Therefore, management plan large, both in terms of removals and the age at which the data is helpful where historical remote forests stop sequestering carbon. Given the size of some of the sensing data is not available for quantifying mostly untouched areas that are being included in some removals that are the result of additional inventories as managed forests, using data that do not human actions. Therefore, given the accurately reflect the situation in these forests could result in difficulty involved in capturing increased large uncertainty in the inventory. For example, in the removals or emissions that are the result of Canadian inventory half of the managed forest area falls in the 4 silviculture activities, data from realm of “intact forest fragments, ”, and approximately 37 management plans will likely continue to be needed to help quantify emissions and 4 Intact forest fragments are “a remnant of an intact forest landscape. It is a contiguous mosaic of naturally occurring ecosystems, including forest, bog, water, tundra, and rock outcrops, that is within a forest ecozone, and that is

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 14 percent of the areas that are being Table 4. Regional Estimates of Timber Volume for Douglas Fir harvested are in areas that were previously Stands on Forest Land After Clearcut Harvest untouched5 (see Table 6). It is not possible to know what assumptions were used to Age Pacific Pacific Pacific NW, Rocky Rocky derive the emissions from these areas NW, NW, West (high Mountain, Mountain, looking at the Canadian inventory; East West production) North South therefore these numbers are merely to note 3 3 3 3 3 the size of the problem that could exist if m /ha m /ha m /ha m /ha m /ha the Canadian inventory had issues similar to 125 465.2 1364.7 1544.4 369.5 242.3 those pointed out for the Australian Source: USDA 2006 inventory.

b. Reference Emission Levels and Non-Anthropogenic Table 5. Timber Volume Yields Used to Emissions Estimate Above-Ground Biomass in

Managed Douglas Fir Stands Translation of inventories of a country’s GHG emissions and

Source: Wulder et al. 2008 removals into emission reductions and increased carbon sequestration that can be counted as a contribution to Timber Site Age of mitigating climate change requires two key determinants. The volume Class Maturity m3/ha first is to set a national reference (baseline) GHG emission level and the second is to establish a means of addressing 10 110 200 anthropogenic (human-induced) versus non-anthropogenic emissions in national accounting. 15 300 200

20 542 200 Table 6. Non-Fire Activity in "Intact Forest Fragments" and Other Forest Areas in Canada 25 797 200 Managed Area Total Area, Forest Loss (Not Subcategories 1,000,000 ha Burned) 1,000,000 ha 30 1048 200 Intact Forest 193.8 2.5 35 1262 165 Fragments Other Forest Area 40 1343 115 182.7 4.3 Sources: Lee et al. 2006; Hansen et al. 2008 45 1,361 80 Negotiations on the quantification of LULUCF emissions associated with Annex I forest management activities have been contentious from the start. A key area of discussion and essentially undisturbed by significant human disagreement has centered around the setting of a baseline influence visible on Landsat satellite images.” (Lee et al. 2006) point or number against which to calculate emissions and 5 Data layers of intact forest fragments (Lee et al. removals - referred to in this paper as the “reference emission 2006) and tree cover change derived using Modis and level.” Linked to this question is whether and how emissions LandSat (Hansen 2005) were used to calculate the and removals from non-anthropogenic events should be areas to give an idea of the scale of natural forests that are being cleared. This is likely to be a relatively addressed in national accounting. These sources often result in conservative number, given that the definition of an some of the national emission fluctuations that worry intact forest fragment requires that a large area be developed countries considering whether to undertake forest untouched by human activity. Note that the management accounting rules. definitions used to classify these areas are not the same as used in the Canadian inventory therefore they cannot be compared directly.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 15

Box 2. The Starting Point: Methods for Defining Reference Emission Levels In a GHG inventory, it is sufficient to simply track the changes of emissions and removals over time. However in order to quantify emission reductions or increased sequestration that result from actions taken in a given country, there needs to a starting point against which current emissions and removals are compared. The decision about which starting point to use can make a big difference in how much a country is seen as “doing” in terms of reducing emissions or increasing removals.

For Annex I countries, the emission or removal changes that are quantified from afforestation, reforestation and deforestation (ARD) – and also from forest management (FM) if elected— is then added to the accounts, where the country has a target. If emissions are reduced or more carbon is sequestered, then the country has less to do to meet its emission reduction target. If removals diminish or emissions increase against the reference level, the country must reduce emissions in other sectors.

Various proposals have been put forward for a new methodology that would provide a reference emission level for forest management activities that is more palatable than what was agreed to for the first commitment period. The approach currently applicable to most Annex I countries is “gross-net.” All countries that were sequestering carbon overall in 1990 adopted this approach. This approach means that the reference emission level is zero. Using gross-net accounting, if a country is sequestering carbon overall –regardless of whether this is more or less sequestration than before or resulted from a change of behavior or not – the country gets a credit for the amount being sequestered.

Countries that were emitting CO2 in 1990, such as Australia, were allowed to use “net-net” accounting. This means that as long as they are emitting less than they were in 1990, they would get credits for reducing emissions.

There have been several papers written about the benefits and limitations of these two approaches (see for example, Bernard Schlamadinger et al. 2007a.) There have also been proposals from Parties for variations on the net-net approach, such as taking a base-period of five or ten years to smooth out some of the annual variability in emissions (EU 2008).

In addition to the more historical “net-net” approach, or the absolute, “gross-net” approach, some countries including Austria, Canada and others have proposed adoption of a “forward -looking” reference emission level. This approach requires identifying likely future levels of activities that will cause emissions for setting the reference emission level, rather than only comparing current emissions and removals to historical emissions data. Often called the BAU (business as usual) approach, this is linked conceptually to project-level accounting practices, such as in the Clean Development Mechanism.

Undertaking this approach at the national level is unprecedented and could allow countries to emit more than they had historically emitted and still get credits. This would run directly counter to the directives in the 1997 UN Framework Convention on Climate Change which set the aim “of returning individually or jointly to [their] 1990 levels these anthropogenic emissions of carbon dioxide and other greenhouse gases not controlled by the Montreal Protocol.” It would also run counter to the clear instruction of the Kyoto Protocol, which states: “That the aim stated in Article 3.1 of the Kyoto Protocol not be changed by accounting for land use, land-use change and forestry.” The aim referred to here is to reduce Parties’ “overall emissions of such gases by at least 5 per cent below 1990 levels in the commitment period 2008 to 2012.”

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 16 During the negotiation of the Kyoto emission levels, in order to make electing forest management Protocol, discussions on setting reference more palatable. levels for Annex I countries centered around the pros and cons of two Questions that have recently surfaced in the negotiations approaches: “gross-net,” based on a include the following: historical counting of emissions; and “net- net,” based on an absolute approach to Should emissions that stem from direct human emissions accounting (Fry 2007) (see Box 2 activities resulting from “business-as-usual” practices – for more detail). The final compromise such as increased use of wood for energy – be required a gross-net approach from all “factored out” (i.e. excluded) so as to not cause debits, countries electing to count emissions even if associated emissions from these “business-as- reductions from forest management that usual” practices are greater than they had been were overall carbon sinks in 1990 and net- historically (see Box 3)? net for those Parties that were an overall How should risks less directly attributable to human source of greenhouse gas emissions.6 In actions be dealt with, if at all? How should increased addition, a cap was set on the potential emissions from fires and insects be “factored out”? emissions or removals that could derive Should they be completely removed so that the from forest management activity, in order country receives no debits for those emissions? Or only to control the amount of credits at could be removed past a certain threshold that is greater than generated (Hohne et al. 2007). the historical variability that would be captured by having a multi-year base? As LULUCF became operational, Parties such as Canada and Australia, realizing that There are a number of different proposals on the table that electing forest management would likely contend with these thorny questions for Annex I Parties. result in carbon dioxide emissions rather However, one question that is not well addressed is how than sequestration, did not sign up to parties will maintain symmetrical accounting practices using include this activity, despite their extensive these approaches. At present, proposals focus on “factoring forest cover (Kurz et al. 2008). out” only emissions and not removals that may be non- anthropogenic or the result of “business as usual” practices. If However, in part to address the concern negotiators conclude that only the more complex approaches that significant sources of emissions are for setting reference emission levels will create a positive being overlooked by countries that are not incentive for countries to address GHG emissions and increase including forest management activity in removals, this raises significant questions about the data their Kyoto Protocol accounting, many required to develop these credibly. Box 4 provides an initial list Parties and other stakeholders are hoping of data required for the types of adjustments suggested in to make accounting for such activity country proposals. mandatory under the post-2012 international climate regime. However, vigorous discussions have taken place in the UN negotiations over the past two years on how to change the current forest management rules, especially reference

6 “This is widely known as the ‘Australia clause’ as it was introduced by Australia at around 2.00 am of the final night of COP-3 in Kyoto to give Australia an additional allowance for their Kyoto target (personal observation)” (Fry 2008, page 344).

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 17

Box 3. Emission from Forest Management Areas in Canada

Figure 3: Emissions and Removals in Forests Remaining Forests, Canada Source: Natural Resource Canada 2007

Emissions and removals in the “forest remaining forest” category in Canada are large and fluctuate significantly. In 2006, total Canadian greenhouse gas emissions were 721 Mt of carbon dioxide equivalent (CO2e). For the same year the net LULUCF emissions (all activities) was 31.3 Mt of CO2e. On average, the Canadian government finds that harvesting activities account for annual average emissions of 122 Mt of CO2e, but that annual wildfire emissions fluctuated between 14 and 342 Mt of CO2e (Natural Resource Canada 2007). This, combined with emissions resulting from insect infestations, causes the huge changes in emissions year by year seen above. The Canadians have put forward reference emission level proposals that try to help them address the risks that such large fluctuations represent when setting a reference level using historical data.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 18

Box 4. Data Requirements for Adjusting Reference Levels for Age Class, Natural Disturbances, and Business-As-Usual Practices

The following list of data requirements was compiled by a group of experts during the Barcelona AWG-LCA meeting in November 2009:

Age class adjustment An age class adjustment would “factor out” reductions in sequestration that may be occurring due to an increasingly old forest age class structure, or an increase of removals from a young forest. In such cases, the following information would be required of countries: A description of the methodology applied for the adjustment Current age class data for forest types in the managed forest area by: o Species o Region and site class o Dominant management regime Methodology for collecting age class data (e.g., spatially explicit, inventory-based, sampling, etc.) and related estimates for standard error Assumptions for projecting future age class structure Data supporting future projections, such as: o Current or future rotation intervals by species and region in intensively managed forests, including the method for collecting data or generating projections o Current harvest/management plans and trajectories by management regime o Natural succession assumptions, including literature references Country/region-specific species growth curves and allometric equations, including literature references

Natural disturbance adjustments Where countries are seeking to apply a natural disturbance adjustment to “factor out” or “diminish” emissions from events like fires, pest outbreak, landslides, etc., the following information would be required: A description of the methodology applied for the adjustment Historical, region-specific, and natural disturbance data, including: o Types of disturbance (fires, insects), return intervals, extent of areas affected and proximity to human influence and infrastructure o For fires, ignition sources, where known, as well as intensity of fires Correlations with environmental factors (e.g. El Niño/Southern Oscillation (ENSO), , and other disturbance events) Information about carbon stocks at the starting point, effects on individual carbon pools, carbon stock recovery trajectories Post-disturbance management practices Methodologies used to collect disturbance data and estimates resulting in emissions and/or removals, including specifying where spatially explicit data has been used, literature sources, etc.

Box continued on next page.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 19 c. Transparency in Reporting

One of the most important lessons that can Box 4: Continued be learned by looking at the LULUCF negotiations and processes is that the National circumstance adjustments active withholding of important data by Where countries are seeking to adjust their reference countries has resulted in a lack of trust level for anthropogenic activities to forward looking between various Parties and stakeholders baselines – for example, to reflect increased use of regarding the credibility of countries’ bioenergy, changes in harvesting practices, or other expressed concerns, particularly their national circumstances – the following information concerns about reference emission levels. would be required: For example, many countries have not A description of the methodology applied for the provided the underlying data needed to adjustment and justification for the removal of verify their claims about hardships resulting the anthropogenic emissions and/or removals from different reference level approaches. The dominant management regime of the areas This lack of trust has led to renewed to be impacted and their historical management proposals in Parties’ submissions to cap the regimes numbers of “credits” that a country could Increases or decreases of emissions in areas claim from forest management activities where anthropogenic disturbances are under LULUCF (PNG 2009). undertaken New management regime being implemented The second lesson is that, given the Changes in growth curves resulting from the concerns over data quality and reference anthropogenic disturbance levels in LULUCF accounting, more Regional information on forest types impacted information about what countries are For each region/forest type impacted, the actually doing to change carbon dynamics – following information should be collected: in terms of their policies and measures - is o Data relevant to all forest carbon pools necessary to build confidence that the o Country- or region-specific emissions “credits” being generated are real and factors from activities verifiable. o Growth curves and rotation intervals o Harvest data (by region, forest type and The information provided by Canada is volume) illustrative. A review of Canada’s third o Effects of harvest on all carbon pools national communication to the UNFCCC— which includes its inventory data, its national inventory report, its national communication report and a compilation of ensuring that a positive incentive has actually been created and policies and measures the country has beneficial change has occurred. This challenge has been taken—reveals a good deal of information recognized by Parties, and has resulted in some of the on reasons why emissions are going up and reference emission level proposals put forward by countries down year to year. However, with a few like Canada (Government of Canada 2009). exceptions, these are related to data and quantification issues or natural A third lesson is that part of the discussion about credibility disturbances, rather than changes in forest revolves around a country’s willingness and ability to report the management. This disconnect between the impacts on biodiversity and local communities and shifts in emissions and removals and actual stakeholders of the activities it takes to reduce emissions or activities undertaken remains a barrier to increase removals. In some cases, these are the issues -- more

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 20 than climate change impacts -- that concern a. Alternative Accounting Rules stakeholders. By not reporting such impacts, countries are undermining their Annex I countries have faced considerable difficulties in setting ability to present the case for why such up accounting rules that would quantify credible “forest activities are important for national well- remaining forest” emission reductions and increased removals. being in addition to climate change One solution would be that Parties recommend that an IPCC or mitigation. other expert group consider in more depth the different approaches to accounting for emissions and/or removals from Applying LULUCF Lessons to REDD the “forest remaining forest” activities in a REDD Plus Parties and experts have done extensive mechanism. There are two major options, beyond the forest work to try to address some of the management activity approach, that Parties could consider in complicated issues around forest addition to the LULUCF approach: management accounting since the Kyoto Protocol was signed. However, the 1. Keeping the activities that increase removals and protracted LULUCF negotiations continued reduce emissions separate, each with their own to experience difficulties in the lead up to reference level, and classifying these areas differently. the Copenhagen Conference of the Parties (COP) in December 2009, highlighting some 2. Undertaking land-based or full carbon accounting clear lessons for Parties. These lessons raise practices. This would ostensibly7 mean quantifying all the following questions that Parties may emissions and removals in a country first, and then wish to consider as they make decisions determining whether any emissions or removals would about a forest mechanism for developing be excluded when accounting for emission reductions countries in a new international climate or increased removals. Normally this approach would agreement: not allow for factoring out “non-anthropogenic emissions” (Schlamadinger et al. 2007b). Are there alternative ways to look at “forest remaining forest” activities that These options present both opportunities and risks. Under can reduce some of the reference option 1, keeping the activities that reduce emission and emission level issues and data quality increase removals separate from each other may create issues that Annex I countries have loopholes and asymmetries. For example, if there are not clear faced? guidelines and criteria on the activities that countries must Is financing from a fund needed, at least account for if they wish to participate in a REDD Plus in the short term, to create a positive mechanism, they may cherry pick those activities that make incentive, given the difficulties in them look “good.” While creating criteria is an obvious getting precise data? solution, these too would likely require negotiation, which may How can MRV practices for REDD Plus itself create more loopholes. improve on those used in LULUCF? How can we ensure that countries On the other hand, separating activities that generate removals adopt the practice of conducting GHG and emission reductions would allow countries to focus on inventories so that, regardless of the gathering more credible and precise data in relation to the final mechanism’s scope of activities, it activities most relevant for generating credits for changing can be progressively improved with behavior under their national circumstances. Some of those better information? data are already available in many developing countries, although still very rough. Refining that data may prove easier than capturing all forest-related emissions and removals. In

7 Though land-based accounting is often mentioned, a clear description of what this would mean is still unclear.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 21 addition, where historical data is available, credibly in a market mechanism. However, the exclusion of such an approach might enable reference these emissions and removals means a much less accurate emissions levels to be created for near term mechanism is being created. actions that are significant for reducing emissions. One option is to establish a REDD Plus fund that includes all such activities, making the full range of potential emission The advantages of option 2, land based reductions and increased removals possible. While Parties accounting, are that - as long as the would need to ensure that quality of data and reference levels assumptions countries use to estimate were continuously updated, the environmental integrity of removals and emissions are complete and Annex I caps would not be compromised. credible - all changes to the land base would be captured. If countries then still Another option – if a fund is not possible – is to use an created rules to “factor out” certain intermediate approach that would allow for quantifying the emissions and/ or removals, this process largest sources of emissions in “forest remaining forest” areas might be more transparent than at present (options are listed in Section B). under LULUCF practices where different countries may capture more or less of their Another approach is to have a hybrid where all activities that emissions and removals depending on how can be accounted for in a precise manner are considered for a they apply things like the definition of market, and the others are compensated through a fund. IPCC forest management and managed land. experts, independent experts and SBSTA should be mandated to further consider these questions carefully as the REDD Plus On the other hand, the ability for mechanism is developed. developing countries to have such information in the short term might be c. Alternative Reporting Methods limited. One way around this might be if developing countries were willing to Given the discussions about credibility and the questions about account for all emissions and removals, biodiversity and social impacts, good reporting at the national without the same considerations about level will be vital for REDD Plus. There are tools currently anthropogenic versus non-anthropogenic available and steps that countries can take that could help all emissions applied to Annex I countries Parties report better information to stakeholders about their under LULUCF. This would provide a much actions. For example, countries could build on the IPCC 2006 more accurate approach, though not guidance that recommends developing subcategories in the necessarily a more precise one. Still, “forest remaining forest” area, which some countries have especially if historical baselines are used already used to develop their inventory; for example, in and significant financial and human defining the “managed area” (Environment Canada 2006). resources made available, this may be a way forward. Advantages of the subcategory approach include helping to determine: b. Different Financing Approaches how to undertake field sampling to develop credible Unless completely new approaches are emission and removal estimates across a variety of developed for gathering data and different management regimes and ecosystem types; developing inventories, as well as for the level of monitoring and ground checking that might be defining reference levels and addressing needed; and questions about non- anthropogenic the type of data required to support the development of emissions, activities that essentially look credible reference levels. like the Annex I“forest management” activity will likely be difficult to include

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 22

In addition, subcategory definitions would of various tenure arrangements. For example, definitions might help countries communicate at both the be given for private lands, state or national parks, timber national and international levels how their plantation concessions, and indigenous peoples’ lands. actions are impacting other important aspects of forest management, such as Finally, definitions linked to the integrity of the forest biodiversity and tenure arrangements. ecosystem - in terms of levels of human intervention, combined size of the ecosystem, and/or other factors - might be Three potential, yet not mutually exclusive, embraced to classify the level of forest health. (See Box 5). options for a subcategory approach would be to report emissions and removals using Although currently not required, countries do selectively report the subcategories “management regime,” some of this information already. A brief survey of Table 5A of “tenure arrangement,” and “forest the Common Reporting Format (CRF) from forested Annex I ecosystem integrity.” For example, countries reveals that some nations already subdivide their “management regime” definitions might forested land into categories based on information such as include the following information: management status, species, or biome. For example, Australia, Japan and New Zealand break down their forested land by “non-intensive managed forest”: A separating out natural forest (which could also be called managed forest in which silvicultural “native” or even “primary” forest) from plantations. Australia practices do not define the structure and and Japan also separate out managed forest from natural composition of forest stands. The forest areas. Australia has also some “management” information may be a natural forest or a forest related to each subcategory. For example, “Harvest Native plantation; a management plan may or may Forests” represents areas of native forest that have been not exist. In a “non-intensive managed harvested at some time and may still be available for harvest, forest,” human interventions have been or may now be in a reserve or other land tenure unavailable for applied to perform production, ecological, harvest (Government of Australia 2009). While this definition or social functions, excluding the following: and others may remain too broad to address some of the fellings that result in greater than 10 transparency concerns delineated above, it does contain many percent reduction in growing stock, of the elements noted. These examples demonstrate how production of wood fuel, or interventions countries could begin to report different activities being that create a condition in which total conducted in their forests, even if accounting rules do not biomass growth is impaired by more than distinguish between these areas. an average of 10 percent over 50 years. MRV activities must be sufficient to show that these exclusions are not violated and to account for changes in total biomass (Funk 2009)

Definitions could be decided by countries and, if necessary, could differ across countries to account for various management regimes. If standardized definitions prove impossible, best practice methods could be developed.

Management definitions might also be linked to, or supplemented by, definitions

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 23

Box 5. Understanding the Landscape, Mapping Forest Ecosystem Types

Spatially explicit data can help countries understand of the impacts of their management practices on biodiversity. For example, the following data are the result of layering together Canada’s Intact and Fragmented Forest Landscapes (see Lee 2006 for more discussion about the methodology) and tree cover change data using Modis and LandSat (see Hansen et al. 2008). Both Intact Forest Landscapes and Fragments represent unbroken expanse of natural ecosystems within the zone of current forest extent, showing no signs of significant human activity, and large enough that all native biodiversity, including viable populations of wide-ranging species, could be maintained (IFL website, Lee 2006). These landscapes represent some of the last large, mostly untouched boreal ecosystems on the planet and are believed to store as much carbon as tropical forest ecosystems in their soils (Malhi et al. 1999, Carlson 2009). Given the importance of these ecosystems for both biodiversity and carbon storage, it would be helpful to understand the impact that forest management activities are having on these areas. By overlaying tree cover change data on the “managed” forest layer provided by Natural Resource Canada staff (Environment Canada 2009), information about what is occurring can more clearly be seen. For example: 1. While there is significant tree cover change in what Canada classifies as “unmanaged forests”, it is mostly by fires, which represent about 99 percent of the total change. Of what little non-fire change there is, however, this appears to be in forest areas that in 2000 had been classified as having minimal human impacts (intact forest areas or intact forest area fragments). 2. Fire levels in unmanaged areas are high, and these fires occur in both areas near and far from managed forest areas, making it difficult to always judge whether all fires in this area are non- anthropogenic or not. However, some areas are so far from areas with significant human intervention that they most likely are not caused by human activity. 3. Fires occur in a slightly larger percent of the unmanaged intact forest areas than in managed intact forest areas, possibly as a result of successful fire suppression activities in these managed areas, which is a criterion used by the Canadian government to define managed forest areas (Environment Canada 2009). 4. In “managed areas” about 2,500,000 ha, or 47 percent of the tree cover change from 2000 to 2005, is from non-burn related impacts to Intact Forest Areas. This represents about 37 percent of total tree cover change from non-burn impacts in the managed area.

If such information were collected and reported systematically, Parties would then have baseline information with which they could easily monitor the impacts of activities on critical ecosystems, such as intact forest areas. They could articulate clear goals related to these areas. For example, parties could commit to reducing emissions from non-fire activities in managed intact forest areas by 50%, or eliminating deforestation altogether in intact forests, in an effort to consider biodiversity impacts (more relevant in intact forests, which harbor high levels of biodiversity), following the principles listed in Decision 16/ CMP.1 which states, “That the implementation of land use, land-use change and forestry activities contributes to the conservation of biodiversity and sustainable use of natural resources” (UNFCCC 2005).

In addition, as noted above, knowing where harvesting activities are occurring in previously undisturbed, e.g. intact forest, areas can help ensure that countries are using more appropriate growth curves and soil carbon assumptions.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 24

Other approaches, including better The Indonesia example in Section B demonstrates the urgent reporting of policies and measures in need to build capacity and understanding about additional national communications around these emission sources in order to ensure that accounting practices issues, and making better links between become increasingly complete and robust over time. However, actions and their impact on emissions and the data collection and analysis required to develop AFOLU removals, may also be important and have inventories is not inexpensive (Schlamadinger et al. 2007b). been recommended in other articles and Regardless of the accounting requirements contained in the papers (Benndorf et al. 2007; Daviet 2009; eventual forest mechanism, non-Annex I countries will need Karsenty 2008; Karsenty 2009). Figure 4, a financial assistance to build the capacity to develop complete reproduction of a graph included in a paper and robust AFOLU/LULUCF inventories. by the Schlamadinger et al. (2007a), presents a conceptual framework that could Section D: Conclusions and help parties think about reporting “cause- Recommendations oriented” activities. This type of reporting with subcategories of different Negotiation of a forest mechanism or mechanisms that create a management regimes reported in Common positive incentive for developing countries to reduce forest- Reporting Format tables, could also be related greenhouse gas emissions will be an important helpful. component of a successful Copenhagen summit or its follow-on

meetings. While these proposals may seem to present a greater burden in tracking activities in It is vital that such a forest mechanism find a path that, various areas, they may help improve the compared with LULUCF, is less convoluted and fraught with quantification of the emission reductions issues that generate distrust between countries around the and sequestration achieved. This would quantification of GHG emissions reductions and increased allow all Parties to better understand where policies and measures are resulting in positive outcomes. Figure 4. Matrix of Options for Linking LULUCF Commitments to Wider

International Arrangements d. GHG Inventories

No matter how accounting takes place , all countries should develop GHG inventories that faithfully and regularly capture all agriculture, forest and other land use (AFOLU) GHG emissions and removals occurring over time as a requirement of participation in a REDD mechanism. Only by doing this will countries be able to capture all the data necessary to assess the impacts of decisions on their land base with the increased accuracy and precision required to build confidence that declared emission reductions and increased removals from land-based activities are credible.

Source: Schlamadinger et al. 2007a

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 25 carbon sequestration. removals, reporting requirements of a forest mechanism for Without such an outcome, the international both developed and developing countries should be broader community runs the risk that support for than the current mandatory requirements under LULUCF. emission reductions and trading under a REDD mechanism may Recommendation 2: A system that relies wholly on carbon prove to be very limited. markets to finance REDD would greatly sacrifice accuracy and/or inclusiveness among countries. Parties should In devising such a pathway Parties will need implement a range of means for financing REDD. This should to emphasize the long-term health of a include a fund to finance activities that cannot be precisely forest mechanism over short-term benefits accounted for. In some cases changing the definition of offered by approaches not considered deforestation in order to capture significant land conversion credible by all. In doing so, Parties will need activities may address some of the accuracy issues. to balance the following: Recommendation 3: Parties should undertake full and a) The accuracy of determining what consistent reporting of all relevant data for quantifying the atmosphere contains – which emission reductions or greenhouse gas removals. This includes requires the development of a data on the basis for setting a reference level and on non- broader scope – with the precision carbon factors, such as biodiversity and impacts on forest of the actual emissions and stakeholders, specifically related to tenure. This reporting removals calculated. should include actions as well as emissions data and b) The desire to report GHG inventory assumptions. information only, with the need to provide more information about Recommendation 4: Any country that engages in voluntary the policies and measures taken, REDD activities as part of its commitments under an including their local impacts on international climate agreement should undertake a full land biodiversity and tenure rights, in use emissions inventory. In accordance with the Convention, order to ensure the credibility of support should be provided to developing countries to produce the mechanism moving forward. these inventories. c) Embracing differences between the REDD Plus mechanism and the LULUCF mechanism where appropriate, without throwing out relevant and useful LULUCF rulemaking processes and without ignoring the possibility that LULUCF could adopt practices developed for REDD Plus to make LULUCF more credible.

Based on lessons from Annex I rules and practices for quantifying and accounting for emissions from forest loss and forest management activities, we recommend four points for the development of a REDD Plus mechanism.

Recommendation 1: To account more accurately for forest-related emissions and

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 26

References Daviet, F. 2009. Beyond Carbon Financing. World Resources Institute, Washington DC. Benndorf, R., S. Federici, C. Forner, N. Pena, http://pdf.wri.org/beyond_carbon_financing.pdf E. Rametsteiner, M.J. Sanz, and Z. Somogyi. 2007. Including land use, land-use change, Environment Canada. 2009. National Inventory Report: and forestry in future climate change, Greenhouse Gas Sources and Sinks in Canada: 1990-2007. agreements: thinking outside the box Greenhouse Gas Division, Environment Canada. United Nations Environmental Science & Policy 10: 283-294 Framework Convention on Climate Change. R. Environment Canada. 2008. National Inventory Report: Brown, S.L. and P.E. Schroeder. 1999. Greenhouse Gas Sources and Sinks in Canada. 1990-2006. Spatial patterns of aboveground production Greenhouse Gas Division, Environment Canada. United Nations and mortality of woody biomass for eastern Framework Convention on Climate Change. U.S. forests. Ecological Applications 9: 968- 980. Environment Canada. 2006. National Inventory Report 1990- 2004. Greenhouse Gas Division, Environment Canada. United Cadman, S. 2008. Submission to the AWG- Nations Framework Convention on Climate Change. KP on LULUCF Definitions, Modalities, Rules and Guidelines in the Second Commitment European Union. 2008. Submission by France on Behalf of the Period. The Wilderness Society. Available European Community and its member states. Voluntary EU online at: submission on the assessment and implications of future rules http://unfccc.int/resource/docs/2008/smsn for the LULUCF Annex I. /ngo/090.pdf FAO. 2006. Global planted forests thematic study: results and Cadman, S. 2009. Personal communication analysis. Planted Forests and Trees Working Paper 38. Rome. on the Australian GHG Inventory and its www.fao.org/forestry/site/10368/en potential limitations. Forest Watch Indonesia (FWI)/Global Forest Watch (GFW). Canadell, J.G et all. 2007. Factoring out 2002. The State of the Forest: Indonesia. Bogor, Indonesia: natural and indirect human effects on Forest Watch Indonesia, and Washington DC: Global Forest terrestrial carbon sources and sinks. Watch. Environmental science & policy 10 (2007) 370–384. Fry, I. 2007. More Twists, Turns and Stumbles in the Jungle: A Further Exploration of Land Use, Land-Use Change and Forestry Carlson, M., Wells, J., Roberts, D. 2009. The Decisions within the Kyoto Protocol. Review of European Carbon the World Forgot: Conserving the Community and International Environmental Law 16: 341-355. Capacity of Canada’s Boreal Forest Region to Mitigate and Adapt to Climate Change. Funk, J. 2009. Personal Communication on potential Boreal Songbird Initiative and Canadian management regime definitions for LULUCF. Boreal Initiative, Seattle, WA, and Ottawa. 33 pp. Government of Australia. 2009. Australia’s National Greenhouse Accounts. National Inventory Report 2007. Volume Curran, L.M., S.N. Trigg, A.K. McDonald, D. 2. Department of Climate Change, Australian Government. Astiani, Y.M. Hardiono, P. Siregar, I. United Nations Framework Convention on Climate Change. Caniago, and E. Kasischke. 2004. Lowland forest loss in protected areas of Indonesian Government of Canada. 2009. Informal Submission to the Borneo. Science 303: 1000--1003. AWG-KP. Information and Data on Land Use, Land-Use Change and Forestry (LULUCF).

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 27 http://unfccc.int/meetings/ad_hoc_working Sustainable Management of Forests and Enhancement of _groups/kp/items/4907.php Forest Carbon Stocks. http://unfccc.meta- fusion.com/kongresse/AWG_08_Ghana/downl/080822_Ghana Government of Canada. 2006. Canada’s _%20AWG_P1_1130_India.pdf Fourth National Report on Climate Change. http://unfccc.int/national_reports/annex_i_ Karsenty A., 2008. The architecture of proposed REDD schemes natcom/submitted_natcom/items/3625.ph after Bali: facing critical choices. International Forestry Review p 10: 443-457.

Hansen M.C., S.V. Stehman, P.V. Potapov, Karsenty, A. 2009. What the carbon market cannot do. French T.R. Loveland, J.R.G. Townshend, R.S. Agricultural Research Centre for International Development DeFries, K.W. Pittman, B. Arunarwati, F. (CIRAD). Stolle, M.K. Steininger, M. Carroll and C. Kurz, W. A., G. Stinson, G. J. Rampley, C. C. Dymond, and E. T. DiMiceli. 2008. Humid tropical forest Neilson. 2008. Risk of natural disturbances makes future clearing from 2000 to 2005 quantified by contribution of Canada's forests to the global carbon cycle using multitemporal and multiresolution highly uncertain. Proceedings of the National Academy of remotely sensed data, Proceedings Of The Sciences of the United States of America 105: 1551-1555. National Academy Of Sciences Of The United States Of America 105 :9439-9444. Lee P, J.D. Gysbers, and Z. Stanojevic. 2006. Canada’s Forest Landscape Fragments: A First Approximation. Global Forest Henson, I. 2009. Modeling Carbon Watch Canada Report. Edmonton, Alberta: Global Forest Watch Sequestration and GHG emissions Canada. associated with OP cultivation and land use change in Malaysia: A Re-evaluation and a Mackey et al. 2008. Green Carbon: The role of natural forests in Computer Model. Malaysian Palm Oil carbon storage. Part 1. A green carbon account of Australia’s Board Technology No. 31. south-eastern Eucalypt forests, and policy implications. Australian National University Press, Canberra; p36; Hohne, N, S. Wartmann, A. Herold, A. http://epress.anu.edu.au/green_carbon_citation.html Freibauer. 2007. The rules for land use, land use change and forestry under the Kyoto Malhi, Y., Baldocchi, D.D., and P.G. Jarvis. 1999. The carbon Protocol – lessons learned for the future balance of tropical, temperate and boreal forests. Plant, Cell climate negotiations. Environmental and Environment 22: 715-740. Science & Policy 10 (2007) 353-369. Mollicone, D., F. Achard, S. Federici, H. D. Eva, G. Grassi, A. Intact Forest Landscape website. Belward, F. Raes, G. Seufert, H. J. Stibig, G. Matteucci, and E. D. http://www.intactforests.org/concept.html Schulze. 2007. An incentive mechanism for reducing emissions from conversion of intact and non-intact forests. Climatic India. 2007. Views on issues related to Change 83: 477-493. further steps under the Convention related to reducing emissions from deforestation in Murdiyarso, D., M. Van Noordwijk, U. R. Wasrin, T. P. Tomich, developing countries: Approaches to and A. N. Gillison. 2002. Environmental benefits and stimulate action. Submissions from Parties. sustainable land-use options in the Jambi transect, Sumatra. Addendum. Journal of Vegetation Science 13:429-438. http://unfccc.int/documentation/document s/advanced_search/items/3594.php#beg Natural Resource Canada. 2007. Is Canada’s Forest a or Source. Science-Policy Notes. October 2007. India. 2008. Policy Approaches and http://cfs.nrcan.gc.ca/news/544 Positive Incentives for Conservation,

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 28

Neilson, E.T., MacLean, D.A., Meng, F.-R., A. Labat, G. Marland, A. Michaelowa, L. Montanarella, P. and P.A. Arp. 2007. Spatial Distribution of Moutinho, D. Murdiyarso, N. Pena, K. Pingoud, Z. Rakonczay, E. carbon in natural and managed stands in an Rametsteiner, J. Rock, M. J. Sanz, U. A. Schneider, A. Shuidenko, industrial forest in New Brunswick, Canada. M. Skutsch, P. Smith, Z. Somogyi, E. Trines, M. Ward, and Y. Forest Ecology and Management 253: 148- Yamagata. 2007a. A synopsis of land use, land-use change and 160. forestry (LULUCF) under the Kyoto protocol and Marrakech accords. Environmental Science & Policy 10: 271-282. Penman, J., M. Gytarsky, T. Hiraishi, T. Krug, D. Kruger, R. Pipatti, L. Buendia, K. Miiwa, T. Schlamadinger, B., T. Johns, L. Ciccarese, M. Braun, A. Sato, A. Ngara, K. Tanabe, and F. Wagner. 2003a. Senyaz, P. Stephens, M. Takahashi, and X. Q. Zhang. 2007b. Good Practice Guidance for Land Use, Land- Options for including land use in a climate agreement post- Use Change and Forestry: Glossary. IPCC 2012: Improving the Kyoto protocol approach. Environmental National Greenhouse Gas Inventories Science & Policy 10: 295-305. Programme. UNFCCC. 2009. AWG LCA Non-Paper 39: Contact Group on Penman, J., M. Gytarsky, T. Hiraishi, T. Krug, Enhanced Action on Mitigation and its Associated Means of D. Kruger, R. Pipatti, L. Buendia, K. Miiwa, T. Implementation. Subgroup on paragraph 1(b) (iii) of the Bali Ngara, K. Tanabe, and F. Wagner. 2003b. Action Plan. IPCC Report on Definitions and http://unfccc.int/meetings/ad_hoc_working_groups/lca/items/ Methodological Options to Inventory 5012.php Emissions from Direct Human-induced Degradation of Forests and Devegetation of UNFCCC. 2007. Decision 1/CP.13 Bali. Report of the Conference Other Vegetation Types. IPCC National of the Parties on its thirteenth session, held in Bali from 3 to 15 Greenhouse Gas Inventories Programme. December 2007 Addendum Part Two: Action taken by the http://www.ipcc- Conference of the Parties at its thirteenth session. nggip.iges.or.jp/public/gpglulucf/degradatio http://unfccc.int/documentation/decisions/items/3597.php?su n_contents.html ch=j&volltext=/CP.13#beg

PNG, Cambodia, Cameroon, Costa Rica, UNFCCC. 2005. Decision 16/CMP.1 Montreal. Land use, land- Democratic Republic of Congo, Dominican use change and forestry. United Nations Framework Republic, Ecuador, Gabon, Ghana, Guyana, Convention on Climate Change. Myanmar, Mexico, Papua New Guinea, http://unfccc.int/resource/docs/2005/cmp1/eng/08a03.pdf#pa Vietnam. 2009. Submission to the Kyoto ge=3 Protocol. http://unfccc.int/meetings/ad_hoc_working UNFCCC. 2001. Decision 11/CP.7 Marrakesh. Land use, land- _groups/kp/items/4907.php use change and forestry. United Nations Framework Convention on Climate Change. Putz, F.E., P.A. Zuidema, M.A. Pinard, R.G.A. http://unfccc.int/resource/docs/cop7/13a01.pdf#page=54 Boot, J.A. Sayer, D. Sheil, P. Sist, J.K. Vanclay, and Elias. 2008. Improved tropical UNFCCC. 1998. Decision 9/CP.4 Buenos Aires. Land use, land- forest management for carbon retention. use change and forestry. United Nations Framework PLoS Biol 6:e166. Convention on Climate Change. http://unfccc.int/resource/docs/cop4/16a01.pdf Schlamadinger, B., N. Bird, T. Johns, S. Brown, J. Canadell, L. Ciccarese, M. van der Werf, G. R., D. C. Morton, R. S. DeFries, J. G. J. Olivier, P. Dutschke, J. Fiedler, A. Fischlin, P. S. Kasibhatla, R. B. Jackson, G. J. Collatz, and J. T. Randerson. Fearnside, C. Forner, A. Freibauer, P. 2009. Co2 emissions from forest loss. Nature Geoscience 2: Frumhoff, N. Hoehne, M. U. F. Kirschbaum, 737-738.

WORLD RESOURCES INSTITUTE • December 2009

Forests in the Balance Sheet 29 van Noordwijk M., A. Widayati, B. Lusiana, K. Hairiah and B. Arifin. 2005. What can a Clean Development Mechanism do to enhance trees in the landscape? Experience with rubber, coffee and timber-based agroforestry systems in Indonesia. In: Murdiyarso D. and Herawati H, eds. 2005. Carbon forestry: Who will benefit? Proceedings of workshop on carbon sequestration and sustainable livelihoods. Bogor, Indonesia. Center for International Forestry Research (CIFOR). pp. 92-111.

White, T.M, Kurz, W. A. 2005. Afforestation on private land in Canada from 1990 to 2002 estimated from historical records. The Forestry Chronicle 81: 491-497.

Wulder, M.A., J.C. White, G. Stinson, T. Hilker, W.A. Kurz, N.C. Coops, B. St-Onge, and J.A. Trofymow. 2009. Implications of differing input data sources and approaches upon forest carbon stock estimation. Environmental Monitoring and Assessment.

WORLD RESOURCES INSTITUTE • December 2009