The flux of CO2 due to land use, land use change and forestry (LULUCF) is estimated in a variety of science- and policy-based assessments. Each approach uses different data and methods as they are designed for different purposes. Results are not directly comparable with each other and are subject to high degrees of uncertainty. This could lead to policy not achieving its desired outcomes.
Estimates directly developed though the policy process (UNFCCC reporting and Kyoto Protocol accounting) focus on the direct consequences of human activities that could be changed for a desired outcome: “activity-based accounting”. Scientific approaches ultimately aim to account for all the fluxes the atmosphere sees to gain better understanding of the system and more robust estimates of past and future change: “full atmospheric accounting”.
The IPCC Fourth Assessment Report estimated global LULUCF emissions uncertainty during the 1990s to be 5.9 ± 4.0 GtCO2/yr (±70%). For reference, this compares to fossil fuel emissions in 1990 of around 23 GtCO2/yr. With increasing use of satellite data and more comprehensive inclusion of processes, estimates of LULUCF emissions have been revised downwards both in terms of the mean and uncertainty. A mean across several recent global model results gives 4.14 ± 1.83 GtCO2e/yr (± 45%) for the 1990s, and 4.03 ± 1.83 GtCO2e/yr for 2000-2009 (Houghton et al., in press). This does not include the peatland flux, which may account for 1.7 ± 1.0 GtCO2e/yr (Joosten, 2009).
For Annex I countries only, the results of a global model show a net source from LULUCF of 0.2 GtCO2/yr in the 1990s switching to a net sink of 0.2 GtCO2/yr in 2000-2009 (Houghton 2003, updated in Friedlingstein et al., 2010). This is in contrast to the UNFCCC which finds a net sink due to forest area change only of 0.2 GtCO2/yr in both decades, and a much larger sink of 1.6 and 2.1 GtCO2/yr for all LULUCF. A small part of the difference is due to the way legacy emission fluxes from forest area change prior to the start of reporting in 1990 are dealt with.
However the main difference is due to forest management. Under UNFCCC, carbon stock change is reported on all areas of forest classed as “managed”. Much of this stock change will be regrowth due to recovery from past disturbance or forest planting, and due to environmental change (fertilising effects of increasing atmospheric CO2 and N concentrations, and climate change such as longer growing seasons). Global models specifically model actual management effects only under LULUCF. Environmental effects are modelled as the residual terrestrial flux, and this is similar in magnitude (with a high degree of uncertainty) to what is reported under “Forest Land Remaining Forest Lands” in UNFCCC.
The results under the Kyoto Protocol are very different again as they serve a very specific purpose of accounting for activity that occurs after 1990 only. Sinks in forests planted prior to 1990 will be counted under Forest Management (FM). Therefore fluxes due to afforestation and reforestation will be much lower than reported in UNFCCC or by global models, while deforestation fluxes will be similar. Therefore the net flux from forest area change under Kyoto Protocol accounting is a source of 0.1 GtCO2/yr in the 1990s and 2000-2009. The sink due to FM is similar to that reported under UNFCCC as data and methods are similar. Under the Kyoto Protocol, the amount that can be credited for FM is capped to represent the fact that much of it is due to past activity and environmental change rather than activity undertaken since 1990 to address climate change. The difference between the model results and uncapped FM clearly indicates that it is correct to have a strong cap.
THE UNEP GAP report (UNEP, 2010) uses similar approach to the Kyoto Protocol in assessing the outcome of pledges on emissions reduction made by countries under the Copenhagen Accord. Emissions targets are based on % reductions in emissions from fossil fuels and energy, but LULUCF credits can be counted towards meeting the target. Accounting for FM is optional and the amount that can be credited is capped. The GAP Report treats FM credits as “lenient” as even the capped amount may not represent real additional emission reductions. If “lenient” 2
emissions are credited then fossil fuel emissions can be higher to reach the same pledged target. If they do not represent real reductions then targets will fail to meet the necessary reduction in emissions. This report finds that the 0.8 GtCO2e/yr “lenient” LULUCF emission used in the report is too high, but some of it does represent genuine emission reduction as it (incorrectly) includes ARD. However since it remains uncertain how much of capped FM is due to real additional activity, it is reasonable to treat this as a “lenient” flux The outcome of emissions pledges analysed in the UNEP GAP report (UNEP, 2010) were compared to Integrated Assessment Model (IAM) trajectories of emissions reduction necessary to meet a 2 degrees stabilisation target. The gap between the pledges and modelled trajectories gives an indication of the amount of further action necessary to meet the target. The gap ranges from 3 to 18 GtCO2/yr depending on low and high ambition (usually “conditional” on other factors such as financing), inclusion or otherwise of “lenient” sinks, and also uncertainty.
Both pledge analyses and modelled trajectories use very different approaches to dealing with LULUCF. Most of the IAMs could not model LULUCF explicitly so all but one used the LULUCF projections from the IMAGE model. This model includes changes in land cover (not just forests), land management (including forests) under LULUCF emissions. The effects of CO2 and climate on vegetation and soils are also modelled but not as part of LULUCF. Unfortunately it remains impossible to separate out direct forest management from other effects using measurements of stock change. Thus unless the methodologies underlying country level reporting and accounting change (e.g. using forest models and reference-level accounting), the approach taken by the GAP report to separate real LULUCF fluxes from “lenient” ones remains appropriate, as it is useful to be consistent with the on going policy process. However LULUCF should be treated more consistently both among pledge analysis groups, and in the way it is described in the report.
The GAP report quotes the IPCC uncertainty in LULUCF estimates of ± 4.0 GtCO2/yr (Denman et al., 2007) as contributing most to the difference across the pledges in each particular case of conditionality and “lenient” or “strict” rules. As stated above, this uncertainty is too high and more recent expert opinion puts it at ±1.8 GtCO2/yr. However this uncertainty is also not fully represented in the pledge analyses as LULUCF was not calculated for any Annex I countries, and only for a few non-annex I countries other than Brazil and Indonesia. It is also not fully accounted for in the modelled trajectories as all but one of the models used the same LULUCF projection. Initial analyses suggested possible anomalies in the reporting of LULUCF in the models. If LULUCF emissions are higher than assessed, it means fossil fuel emissions would have to be even lower to meet the same mitigation target. Uncertainty in LULUCF will also affect the size of the gap. These possibilities should be explicitly assessed in any follow on of the GAP Report.