Short- and long-term warming effects of methane may affect the cost-effectiveness of mitigation policies and benefits of low-meat diets – Nature.com

Description of the models

The CAPRI (Common Agricultural Policy Regionalised Impact) modelling system is an economic large-scale, comparative-static, partial equilibrium model focusing on agriculture and the primary processing sectors. CAPRI comprises two interacting modules, linking a set of mathematical programming models of EU regional agricultural supply to a spatial multicommodity model for global agrifood markets. The regional supply models depict a profit-maximizing behaviour of representative farms in the European Union and candidate countries, taking into account constraints related to land availability, nutrient balances for cropping and animal activities and policy restrictions40. The market module consists of a spatial, non-stochastic global multicommodity model for about 60 primary and processed agricultural products, covering 77 countries in 40 trading blocks. Bilateral trade flows and attached prices are modelled based on the Armington assumption of quality differentiation41. The behavioural functions in the market model represent supply and demand for primary agricultural and processed commodities (including human and feed consumption, biofuel use, import demand from multilateral trade relations), balancing constraints and agricultural market policy instruments (that is, import tariffs, tariff rate quotas, producer and consumer support estimates, and so on). Depending on scenarios, behavioural functions are shifted (for example, to reflect productivity shocks or preference shifts) and the model solves for the new market equilibrium.

With regard to GHG accounting, CAPRI calculates EU agricultural GHG emissions for the most important N2O and CH4 emission sources based on the inputs and outputs of agricultural production activities, following to a large extent the 2006 IPCC guidelines. It also takes into account detailed technical and management-based GHG mitigation options for EU agriculture. GHG emissions for the rest of the world are estimated on a commodity basis in the market model42,43 GHG mitigation in non-European countries is represented by a change in emission factors and a matching change in output prices to reflect the increase in cost, derived from mitigation cost functions from the literature25. In terms of the database the European data are mostly sourced from Eurostat, while the international data are mostly from the Food and Agriculture Organization, for both model parts supplemented by topic-related sources.

The Global Biosphere Management Model (GLOBIOM)44 is a partial equilibrium model that covers the global agricultural and forestry sectors, including the bioenergy sector. Commodity markets and international trade are represented at the level of 35 economic regions in this study. Prices are endogenously determined at the regional level to establish market equilibrium to reconcile demand, domestic supply and international trade. The spatial resolution of the supply side relies on the concept of simulation units, which are aggregates of 5–30 arcmin pixels belonging to the same altitude, slope and soil class, and the same country45. For crops, livestock and forest products, spatially explicit Leontief production functions covering alternative production systems are parameterized using biophysical models such as EPIC (Environmental Policy Integrated Model)46, G4M (Global Forest Model)47 or the RUMINANT model48. For the present study, the supply-side spatial resolution was aggregated to 2° (about 200 × 200 km at the equator). Land and other resources are allocated to the different production and processing activities to maximize a social welfare function which consists of the sum of producer and consumer surplus. The model includes six landcover types: cropland, grassland, short rotation plantations, managed forests, unmanaged forests and other natural vegetation land. Depending on the relative profitability of the production activities of primary products, by-products and final products, the model can switch from one landcover type to another. Spatially explicit land conversion over the simulation period is endogenously determined within the available land resources and conversion costs that are taken into account in the producer optimization behaviour. Land conversion possibilities are further restricted through biophysical land suitability and production potentials, and through a matrix of potential landcover transitions. GLOBIOM covers major GHG emissions from agricultural production, forestry and other land use including CO2 emissions from above- and belowground biomass changes, N2O from the application of synthetic fertilizer and manure to soils, N2O from manure dropped on pastures, CH4 from rice cultivation, N2O and CH4 from manure management, and CH4 from enteric fermentation. For this study, only results for non-CO2 emissions were reported.

GLOBIOM explicitly covers different mitigation options for the agricultural sector. Technical mitigation options such as anaerobic digesters, livestock feed supplements, nitrogen inhibitors, and so on, are based on ref. 49. Structural adjustments are represented through a comprehensive set of crop and livestock management systems parameterized using biophysical models, that is, transition in management systems, reallocation of production within and across regions44 and consumers’ response to market signals50. Detailed information on the parameterization of the different mitigation options for the agricultural sector is presented in ref. 38. For more information on the general model structure we refer to refs. 44,51.

The Modular Applied GeNeral Equilibrium Tool (MAGNET) model is a multiregional, multisectoral, applied general equilibrium model based on neoclassical microeconomic theory52,53. It is an extended version of the standard GTAP model54. The core of MAGNET is an input–output model, which links industries in value-added chains from primary goods, over continuously higher stages of intermediate processing, to the final assembly of goods and services for consumption. Primary production factors are employed within each economic region, and hence returns to land and capital are endogenously determined at equilibrium, that is, the aggregate supply of each factor equals its demand. On the consumption side, the regional household is assumed to distribute income across savings and (government and private) consumption expenditures according to fixed budget shares. Private consumption expenditures are allocated across commodities according to a non-homothetic constant difference of elasticity expenditure function, and government consumption according to Cobb–the Douglas expenditure function.

The MAGNET model, in comparison to GTAP, uses a more general multilevel sector-specific nested constant elasticity of substitution production function, allowing for substitution between primary production factors (land, labour, capital and natural resources) and intermediate production factors, and for substitution between different intermediate input components (for example, energy sources and animal feed components). MAGNET includes an improved treatment of agricultural sectors, examples include: various imperfectly substitutable types of land; the land-use allocation structure; a land-supply function; substitution between various animal-feed components53,55, agricultural policy (such as production quotas and different land-related payments) and biofuel policy (capital-energy substitution, fossil fuel–biofuel substitution56). On the consumption side, a dynamic constant difference of elasticity expenditure function is implemented which allows for changes in income elasticities when purchasing-power-parity-corrected real gross domestic product per capita changes. Segmentation and imperfect mobility between agriculture and non-agriculture labour and capital are introduced in the modelling of factors markets.

MAGNET calculates absolute non-CO2 GHG emissions resulting from agricultural production which depends on demand (gross domestic product, population, diet and bioenergy use) and productivity. Emission intensities (that is, emissions per unit of production) are determined through model-specific emission factors. In addition, emission intensities change in the SSP2 baseline scenario due to the following assumptions on technological improvements: (1) nitrogen fertilizer substitution with labour, capital and land; (2) yield increases due to exogenous technological improvements (adopted from IMAGE) and endogenous improvements due to substitution of land with fertilizer and land–fertilizer bundle with labour and capital; and (3) exogenous feed use efficiency by livestock (adopted from the IMAGE model57) and endogenous substitution between different feed components.

In MAGNET most of the CH4 emissions scale with the output of the agricultural sector and so taxing the emissions is equivalent to a tax on output. This is also the case with N2O emissions from the livestock sectors. For the crop sectors, however, N2O emissions come mostly from the application of synthetic fertilizer which can be substituted for land. If the land price rises (declines) the crop sectors will have an increased incentive to apply more (less) fertilizer and use relatively less (more) land. Marginal abatement cost curves are exogenously implemented based on calculations per sector and region from the IMAGE model. For every period the CO2 price would correspond to a particular level of emission abatement by technical means (that is, farmers would have an incentive to invest in abatement technology) which would be reflected in a reduction of the emission coefficient for a particular agricultural sector. The additional cost of this abatement would be added to the effective carbon price applied to the sector.

Scenario construction

The scenarios considered are counterfactual to a long-term ‘business as usual’ projection of agricultural commodity markets and are presented to provide a more comprehensive perspective of how global mitigation policies and dietary policies could contribute to the temperature target set by the Paris Agreement under GWP100 and GWP* metrics. Focus is on the reduction of agricultural CH4 emissions over time and their effective contribution to climate change, differentiating between sources (for example, ruminant, dairy and rice production) and world producing regions.

To analyse the economic impact of global climate mitigation policies we use a global carbon price path as a proxy for a global mitigation effort15. The impacts of this global carbon price on CH4 emissions depends on the emission metric applied, and will differ from the standard GWP100 if the GWP* metric is applied due to the introduction of time dynamics in its calculation. This is achieved through the following equation, presented as in the simplified rearrangement from ref. 11:

$$E_{{{{\rm{CO}}_2 \mbox{-} {\rm{w.eq.}}}}_{({\mathrm{CH}_{4}})}} = {\mathrm{GWP}}_{100} \times \left( {4 \times E_{{\mathrm{CH}_{4}}\left( t \right)} – 3.75 \times E_{{\mathrm{CH}_{4}}\left( {t – 20} \right)}} \right)$$

(1)

where ‘CO2-warming-equivalent’ emissions ($$E_{{\mathrm{CO}_{2} \mbox{-} {\mathrm{w.eq.}}}}$$) have a large initial effect at the time of release (four times the conventional GWP100 valuation), but much of this (3.75 times the conventional GWP100 valuation) is considered reversed 20 yr later. Consequently, the reported CO2e valuation of CH4 emissions can be higher for those sources where emissions have increased over time and can be negative for those cases where emissions have decreased.

Added temperature from CH4 emissions is computed in equation (2):

$${\mathrm{AW}}_{{\mathrm{CH}_{4}}\left( t \right)} = E_{{\mathrm{CO}_{2}} \mbox{-} {\mathrm{w.eq.}}_{(\mathrm{CH}_{4})}} \times {\mathrm{TCRE}}$$

(2)

where $${\mathrm{AW}}_{{\mathrm{CH}_{4}}\left( t \right)}$$ is the added warming (that is, temperature increase or decrease) in year t relative to year t − 20. For TCRE (that is, the transient climate response to cumulated carbon emissions)28, we use the observationally constrained best estimate of 1.8 °C per TtC13, which is converts into 0.49 °C per TtCO2.

‘Carbon pricing’ is widely acknowledged as an efficient means to achieve the ambitions set out in the Paris Agreement17,18,19,20; however, it requires a transparent, predictable and practicable monitoring system that reports emissions at their source. Emissions from agriculture differ from emissions from standardized industrial processes due to their biological nature, diverse land-use techniques and different farm-management practices, leading to large variations in emission intensities for identical products21,22,23. In addition, the spatial dispersion of farming renders the accurate monitoring of agricultural emissions at their source almost impossible. Carbon prices have therefore been applied in agricultural economic models as an approximation of other policies that incentivize farmers to implement mitigation options (or penalizes them for not adopting them), while the transaction costs caused by those policies have been neglected15,24.

Notwithstanding the fact that monitoring emissions in agriculture involves high transaction costs, the change in the GHG accounting metric to GWP* makes the implementation of a carbon price in agriculture more complicated because the warming impact of CH4 emissions needs to be based on two points in time, 20 yr apart. In effect, the full impact of emissions is delayed by 20 yr, raising concerns about who can be made responsible for paying the carbon price. GWP100 accounts for emissions only in the year in which they originate. On the contrary, GWP* accounting requires the year emissions occur, and 20 yr before, to reflect that short-lived CH4 is rapidly destroyed in the atmosphere via natural processes. A carbon price that requires a time span of 20 yr to be calculated is difficult to administer. Neither the farmer nor the firm responsible for the emissions will necessarily be the same after 20 yr. (This 20-yr time span is suggested in ref. 13, which has the effect of reducing the volatility in CO2-w.eq. emissions and improving the correspondence with temperature response.)

For this paper, we consider two options for the computation of the carbon price: a ‘short term’ (MEF-ST) and a ‘long term’ (MEF-LT) one, separating out the two components of the GWP* equation.

The MEF-ST option focuses on the strong impact of changing CH4 emission rates, computing the carbon price based on the initial valuation at the point an emission occurs, and neglecting the subsequent reversal of most of the emission’s impact in the years ahead. Consequently, the resulting carbon price will be four times higher than a GWP100-based carbon price, leading to strong incentives to mitigate CH4 emissions. The MEF-ST option could reflect motives of decision-makers that prioritize reducing overall GHG emissions fast, that is, almost independent from any metric.

The MEF-LT option aims at implementing a GWP*-based carbon price assuming that the CH4 dynamics are perfectly understood by economic actors and credibly enforced by regulators. Since we apply static models to a dynamic planning problem, we simplify the planning problem by assuming Hotelling’s rule58. This rule states that the optimal price path of a non-renewable, durable resource follows the discount rate. We assume that every CO2 emission causes the same damage, and, consequently, no tipping points are considered. In our case, that resource would be the CO2-absorbing capacity of the atmosphere. With constant carbon prices in real terms, the carbon price will be understood as a regime that charges emissions four times the GWP100-based carbon price the year they accrue, but rewards a rebate of 3.75 times that price 20 yr later. The net effect is a price of (4 – 3.75 =) 25% of the GWP100-based carbon price. The MEF-LT option thus reflects the fact that ‘after 20 years much of the warming caused by an individual CH4 emitter is automatically reserved’11. In our models, the MEF-LT carbon price is implemented with its net effect in the year emissions occur.

The two options above only regard the pricing of CH4. The computation of the carbon price for CO2 and N2O is not affected by GWP* and follows GWP100 in both options. Although N2O is less durable than CO2, it is not considered a short-lived climate gas like CH4. These globally uniform carbon prices are used to estimate the cost-efficient mitigation potential and its distribution across sectors and regions rather than a real-world policy15. We apply two global carbon price trajectories, US$150 and US$500 per tCO2e at 2070, consistent with earlier publications, as higher carbon prices cannot stimulate more technical options, but decrease consumption. In addition to CH4 emissions, N2O emissions from agricultural production are priced according to the GWP100 N2O price. In line with the focus of this study on non-CO2 emissions from agriculture, CO2 emissions from deforestation or other land-use change are not priced.

The second mitigation option is a shift towards consumption of diet containing less animal protein. The adoption of such a diet has been identified in the literature as a promising strategy to curb GHG emissions from the agriculture and global food systems29,30. This is in line with recommendations by the EAT-Lancet Commission, which proposes a healthier diet where whole grains, fruits, vegetables, nuts and legumes comprise a greater proportion of foods consumed. This diet includes calorie intake targets by food group and a total calorie intake target of 2,100 kcal (refs.31,59,60,61,62,63).

Calculation of methane prices

The models are run for two options for the computation of the carbon price: MEF-ST (‘short-term’) and MEF-LT (‘long-term’). Both options use the same carbon price. We distinguish two price paths: US$150 and US$500 per tCO2e in 2005. Following Hotelling’s rule58, the optimal carbon price path follows a discount rate which is set to 5%. This results in the carbon price rates shown in Table 3.

We base our computation of the carbon price for CH4 on the GWP* equation in ref. 11, p. 3:

$${\mathrm{GWP}}_t^ \ast = \left( {4 \times E_{{\mathrm{SLCP}}\left( t \right)} – 3.75 \times E_{{\mathrm{SLCP}}\left( {t – 20} \right)}} \right) \times {\mathrm{GWP}}_{100}$$

where $${\mathrm{GWP}}^*_t$$ measures the global warming potential of CH4 depending on emitted CH4 in year t and emitted CH4 20 yr before (t - 20). Emitted CH4 in t appears a second time in the calculation of GWP* in t + 20:

$${\mathrm{GWP}}_{t + 20}^ \ast = \left( {4 \times E_{{\mathrm{SLCP}}\left( {t + 20} \right)} – 3.75 \times E_{{\mathrm{SLCP}}\left( t \right)}} \right) \times {\mathrm{GWP}}_{100}$$

Assuming a carbon price rate CPRt, a given CH4 emission ESLCP(t) in year t is taxed twice and the total amount CPt of that emission is given by:

$${\mathrm{CP}}_t = \left( {4 \times E_{{\mathrm{SLCP}}\left( t \right)} \times {\mathrm{CPR}}_t – 3.75 \times E_{{\mathrm{SLCP}}\left( t \right)} \times {\mathrm{CPR}}_{t + 20}} \right) \times {\mathrm{GWP}}_{100}$$

Since the path of the carbon price follows the discount rate, the carbon price rate in real prices remains the same in all years, CPt+20 = CPt.

The MEF-ST option of the carbon price disregards the flow term, that is, the price reward 20 yr after the emission occurred. Therefore, the carbon price of a CH4 emission in year t defined in US$per tCH4 becomes: $${\mathrm{CP}}_t\left({\mbox{}}{{{\rm{short}} \mbox{-} {\rm{term}}}}{\mbox{‘}}\right) = \left( {4 \times E_{{\mathrm{SLCP}}\left( t \right)} \times {\mathrm{CPR}}_t} \right) \times {\mathrm{GWP}}_{100}$$ The long-term option of the carbon price regards both terms, and the carbon price of a CH4 emission in year t defined as US$ per tCH4 is computed as:

$${\mathrm{CP}}_t\left({\mbox{}}{{{\rm{long}} \mbox{-} {\rm{term}}}}{\mbox{‘}}\right) = \left( {0.25 \times E_{{\mathrm{SLCP}}\left( t \right)} \times {\mathrm{CPR}}_t} \right) \times {\mathrm{GWP}}_{100}$$

Assuming GWP100 = 25 (AR4), the carbon prices for the two options are shown in Table 4. For comparison, the price rates based on GWP100 are also shown.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Source of this news: https://www.nature.com/articles/s43016-021-00385-8

IP Rotating Proxy Onsale

SPECIAL LIMITED TIME OFFER

00
Months
00
Days
00
Hours
00
Minutes
00
Seconds
First month free with coupon code FREE30