9.4.3.2 Global Forest sectoral modelling

Currently, no integrated assessment (Section 9.4.3.3) and climate stabilization economic models (Section 3.3.5) have fully integrated a land use sector with other sectors in the models. Researchers have taken several approaches, however, to account for carbon sequestration in integrated assessment models, either by iterating with the land sector models (e.g., Sohngen and Mendelsohn, 2003), or implementing mitigation response curves generated by the sectoral model (Jakeman and Fisher, 2006). The sectoral model results described here use exogenous carbon price paths to simulate effects of different climate policies and assumptions. The starting point and rate of increase are determined by factors such as the aggressiveness of the abatement policy, abatement option and cost assumptions, and the social discount rate (Sohngen and Sedjo, 2006).

Since TAR, several new global assessments of forest mitigation potential have been produced. These include Benitez-Ponce et al, (2004; 2007), Waterloo et al. (2003) with a constraints study, Sathaye et al. (2007), Strengers et al. (2007) Vuuren et al. (2007), and Riahi et al. (2006). Global estimates are provided that are consistent in methodology across countries and regions, and in terms of measures included. Furthermore, they provide a picture in which the forestry sector is one option that is part of a multi sectoral climate policy and its measures. Thus, these assessments provide insight into whether land-based mitigation is a cost-efficient measure in comparison to other mitigation efforts. Some of these models use a grid-based global land-use model and provide insight into where these models allocate the required afforestation (Figure 9.11).

Figure 9.11: Comparison of allocation of global afforestation in various studies

(A) Location of bioenergy and carbon plantations

(B) Additional sequestration from afforestation per tropical country per year in the period 2008-2012 (MtC/yr),

(C) Percentage of a grid cell afforested

(D) Cumulative carbon sequestration through afforestation between 2000 and 2012 in Central and South America).

Source: (A) Strengers et al., 2007; (B) Waterloo et al., 2003; (C) Strengers et al., 2007; (D) Benitez-Ponce et al., 2007.

The IMAGE model (Strengers et al., 2007) allocates bio-energy plantations and carbon plantations mostly in the fringes of the large forest biomes, and in Eastern Europe. The Waterloo study only looked at tropical countries, but found by far the largest potential in China and Brazil. Several models report at the regional level, and project strong avoided deforestation in Africa, the Amazon, and to a lesser extent in Southeast Asia (where land opportunity costs in the timber market are relatively high). Benitez-Ponce (2004) maps geographic distribution of afforestation, adjusted by country risk estimates, under a 50 US$/t carbon price. Afforestation activity is clustered in bands in South-Eastern USA, Southeast Brazil and Northern South America, West Africa, north of Botswana and East Africa, the steppe zone grasslands from Ukraine through European Russia, North- Eastern China, and parts of India, Southeast Asia, and Northern Australia. Hence, forest mitigation is likely to be patchy, but predictable using an overlay of land characteristics, land rental rates, and opportunity costs, risks, and infrastructure capacity.

Several models produced roughly comparable assessments for a set of constant and rising carbon price scenarios in the EMF 21 modelling exercise, from 1.4 US$/tCO₂ in 2010 and, rising by 5% per year to 2100, to a 27 US$ constant CO₂ price, to 20 US$/tCO₂ rising by 1.4 US$/yr though 2050 then capped. This exercise allowed more direct comparison of modelling assumptions than usual. Caveats include: (1) models have varying assumptions about deforestation rates over time, land area in forest in 2000 and beyond, and land available for mitigation; and (2) models have different drivers of land use change (e.g., population and GDP growth for IMAGE, versus land rental rates and timber market demand for GTM).

Global models provide broad trends, but less detail than national or project analyses. Generally global models do not address implementation issues such as transaction costs (likely to vary across activities, regions), barriers, and mitigation programme rules, which tend to drive mitigation potential downward toward true market potential. Political and financial risks in implementing afforestation and reforestation by country were considered by Benitez-Ponce et al. (2007), for example, who found that the sequestration reduced by 59% once the risks were incorporated.

In the last few years, more insight has been gained into carbon supply curves. At a price of 5 US$/tCO₂, Sathaye et al. (2007) project a cumulative carbon gain of 10,400 MtCO₂ by 2050 (Figure 9.12b). The mitigation results from a combination of avoided deforestation (68%) and afforestation (32%). These results are typical in their very high fraction of mitigation from reduced deforestation. Sohngen and Sedjo (2006) estimate some 80% of carbon benefits in some scenarios from land-use change (e.g., reduced deforestation and afforestation/reforestation) versus some 20% from forest management.

Benitez-Ponce et al. (2007) project that at a price of 13.6 US$/tCO₂, the annual sequestration from afforestation and reforestation for the first 20 years amounts to on average 510 MtCO₂/yr (Figure 9.12a). For the first 40 years, the average annual sequestration is 805 MtCO₂/yr. The single price of 13.6 US$/tCO₂ used by Benitez-Ponce et al. (2005) should make afforestation an attractive land-use option in many countries. It covers the range of median values for sequestration costs that Richards and Stokes (2004) give of 1 US$ to 12 US$/tCO₂, although VanKooten et al. (2004) present marginal cost results rising far higher. Sathaye et al. (2007) project the economic potential cumulative carbon gains from afforestation and avoided deforestation together (see also tropics, Section 9.4.3.1.). In the moderate carbon price scenarios, the cumulative carbon gains by 2050 add up to 91,400 to 104,800 MtCO₂.

Figure 9.12: Comparison of carbon supply curves globally from various studies

(A) Cumulative carbon supply curves: afforestation and reforestation by year and price scenario. At a price of 100 US$/tC after 70 years, some 40 Gt carbon will have been supplied cumulatively from afforestation.

(B) Annual cost-supply curves for abandoned agricultural land in the B2 scenario. For example, at a price of 100 US$/tC, in 2075, some 250 Mt carbon will have been supplied annually from afforestation and reducing deforestation.

(C) Annual marginal cost curves for carbon sequestration in forests: estimates for boreal, temperate, and tropical regions. For example, at a price of 100 US$/tC, some 1400 Mt carbon will have been supplied annually from afforestation and reducing deforestation in 2100.

Sources: (A) Benitez-Ponce et al., 2005; (B) Strengers et al., 2007; (C) Sohngen and Sedjo, 2006.

The anticipated carbon price path over time has important implications for forest abatement potential and timing. Rising carbon prices provide an incentive for delaying forest abatement actions to later decades, when it is more profitable (Sohngen and Sedjo, 2006). Carbon price expectations influence forest investment decisions and are, therefore, an important consideration for estimating mitigation potential. Contrary, high constant carbon prices generate significant early mitigation, but the quantity may vary over time. Mitigation strategies need to take into account this temporal dimension if they seek to meet specific mitigation goals at given dates in the future (US EPA, 2005).

Some patterns emerge from the range of estimates reviewed in order to assess the ratio between economic potential and technical potential (Sathaye et al., 2007; Lewandrowski et al., 2004; US EPA, 2005; Richards and Stokes, 2004). The technical potential estimates are generally significantly larger than the economic potential. These studies are difficult to compare, since each estimate uses different assumptions by different analysts. Economic models used for these analyses can generate mitigation potential estimates in competition to other forestry or agricultural sector mitigation options. Generally, they do not specify or account for specific policies and measures and market penetration rates, so few market potential estimates are generated. Many studies do not clearly state which potentials are estimated.

The range of economic potential as a percentage of technical potential is 2% to 100% (the latter against all costs). At carbon prices less than 7 US$/tCO₂, the highest estimate of economic potential is 16% of the technical potential. At carbon prices from 27 US$/tCO₂ to 50 US$/tCO₂, the range of economic potential is estimated to be 58% or higher of the technical potential, a much higher fraction as carbon prices rise. Table 9.3 summarizes mitigation results for four major global forest analyses for a single near-term date of 2030: two forest sector models - GTM (Sohngen and Sedjo, 2006; and GCOMAP (Sathaye et al., 2007), one recent detailed spatially resolved analysis of afforestation (Benitez-Ponce et al., 2007), and one integrated assessment model with detail for the forest sector (IMAGE 2.2, Vuuren et al., 2007). These studies offer roughly comparable results, including global coverage of the forest sector, and land-use competition across at least two forest mitigation options (except Benitez-Ponce et al., 2007). All but the Benitez-Ponce et al. study have been compared by the modelling teams in the EMF 21 modelling exercise (see Sections 3.2.2.3 and 3.3.5) as well.

These global models (Table 9.3) present a large potential for climate mitigation through forestry activities. The global annual potential in 2030 is estimated at 13,775 MtCO₂/yr (at carbon prices less than or equal to 100 US$/tCO₂), 36% (~5000 MtCO₂/yr) of which can be achieved under a price of 20 US$/tCO₂. Reduced deforestation in Central and South America is the most important measure in a single region with 1,845 MtCO₂/yr. The total for the region is the largest for Central and South America with an estimated total potential of 3,100 MtCO₂/yr. Regions with a second largest potential, each around 2000 MtCO₂, are Africa, Centrally Planned Asia, other Asia, and USA. These results project significantly higher mitigation than the regional largely bottom-up results. This is somewhat surprising, and likely, the result of the modelling structure, assumptions, and which activities are included. Additional research is required to resolve the various estimates to date using different modelling approaches of the potential magnitude of forestry mitigation of climate change.