Reference Levels


For the purposes of this webpage, reference levels are defined as benchmarks for assessing performance.

For standards or initiatives focused on carbon accounting or reporting, reference levels provide a quantitative way to measure the performance of a country, programme, or project in reducing emissions or increasing removals. For standards focused on social and environmental issues, reference levels also provide a benchmark for assessing the social and environmental impact of activities relative to what would have occurred in their absence.

REDD+ negotiations in the UNFCCC have discussed “forest reference emission levels and/or forest reference levels” (REL/RLs) without a clear distinction negotiated between the two. It has been agreed that REL/RLs are expressed in tonnes of carbon dioxide equivalent per year and defined as “benchmarks for assessing each country’s performance” in implementing REDD+ activities. In REDD+ negotiations, baseline methods that rely on extrapolating historic rates of deforestation have been seen as problematic, particularly by “high forest cover, low deforestation” (HFLD) countries that are under increasing pressure from economic growth or agricultural expansion and, absent additional policies or measures, would expect deforestation to increase. For this reason, it has been agreed that REDD+ REL/RLs should “take into account historic data” but also can “adjust for national circumstances”. Information to substantiate such adjustments must be provided. The guidance provided to date also suggests an approach for REL/RLs that is flexible (allowing for some choice in pools, gases and activities), step-wise (allowing for improvements over time in data and methodologies), and transparent (countries submit information and a rationale). There is no agreement yet, however, on how REDD+ REL/RLs relate to financing.

Under the Kyoto Protocol, historic data—in most cases greenhouse gas (GHG) emissions from the year 1990—have been used to establish emissions targets across all sectors. Recent submissions from developed countries for economy-wide commitments to 2020 continue to use historic data but refer to a variety of “base years” (including 1990, 1992, 2000, and 2005). Voluntary submissions from developing countries of “nationally appropriate mitigation actions” often take a more forward looking approach using a range of methods that include reductions in GHG intensity (i.e. per unit GDP), reductions against business-as-usual projections, GHG emission reductions based on whether international support is received, or even long-term climate neutrality. All such submissions can be found on the UNFCCC website (see Recommended Reading).

The Kyoto Protocol also allows flexibility for Annex B countries when setting their reference levels for forest management. This flexibility was in recognition that using historic reference levels for land use may not capture countries’ relative level of effort within a commitment period (given that forest rotation periods may be longer than KP commitment periods), or if a historically distant (e.g. 1990) base year is used given the age-class structure of forests. Under an agreement made at COP-17 in 2011, Annex B Parties may set their forest management reference level using a variety of approaches, including a projection.

Emerging REDD+ reference levels are taking various approaches. Brazil has chosen to use a historic average, whereas Guyana (an HFLD country) in its agreement with Norway, has created a unique “Combined Incentives” model.

Domestic programmes that recognize forest-related offsets and project-based mechanisms (e.g. voluntary carbon markets standards and CDM) tend to favor the development of “business as usual” reference levels to measure performance. Such standards have developed the most detailed set of guidance for creating baselines through their associated methodologies which often include complex spatial modeling for estimating future deforestation (e.g. under the Verified Carbon Standard (VCS)).

Emerging guidelines for jurisdictional approaches tend to favor a strong reliance on historical data and conservative approaches, and only allow for adjustments based on national circumstances in more exceptional circumstances and when such adjustments can be justified. The use of projections at a jurisdictional or national scale are seen as more difficult to estimate given the complex dynamics and multiple drivers of deforestation and forest degradation that may be hard to model over a 5 – 10 year period (e.g. future commodity prices, impacts of government policy, economic development) and require robust data to develop reliable models and accurate assessments of future emissions and removals.

Finally, social and environmental standards use reference levels, or scenarios, based on what expected conditions would be like absent the project activities.