Energy demand - REMIND-MAgPIE: Difference between revisions
Laura Delsa (talk | contribs) No edit summary |
Laura Delsa (talk | contribs) No edit summary |
||
| Line 24: | Line 24: | ||
Second, the model can endogenously improve end-use efficiency by investing in more efficient technologies for the conversion of final energies into energy services. For example, three vehicle technologies with different efficiencies are implemented in the light duty vehicle (LDV) mode of the transport sector, including internal combustion engine vehicles, battery-electric vehicles, and fuel cell vehicles. | Second, the model can endogenously improve end-use efficiency by investing in more efficient technologies for the conversion of final energies into energy services. For example, three vehicle technologies with different efficiencies are implemented in the light duty vehicle (LDV) mode of the transport sector, including internal combustion engine vehicles, battery-electric vehicles, and fuel cell vehicles. | ||
<references/> | |||
Revision as of 15:09, 3 February 2017
| Corresponding documentation | |
|---|---|
| Previous versions | |
| Model information | |
| Model link | |
| Institution | Potsdam Institut für Klimafolgenforschung (PIK), Germany, https://www.pik-potsdam.de. |
| Solution concept | General equilibrium (closed economy)MAgPIE: partial equilibrium model of the agricultural sector; |
| Solution method | OptimizationMAgPIE: cost minimization; |
| Anticipation | |
Baseline final energy demand projections are provided exogenously to REMIND. REMIND is then calibrated to match these projections in the baseline case. This approach based on EDGE (Energy Demand GEnerator) version 2, a simple econometric model combined with long-term convergence assumptions, helps combining historical trends with stylized facts and provides simple insights to the baseline energy demand trajectories.
These projections cover several energy carriers and sectors (residential, commercial, industry, non-energy use, agriculture and fisheries, others).
Energy demand projections are calculated with the model EDGE. For the scenario SSP2, EDGE 2 relies on a mixture of econometric projections and convergence assumptions to produce baseline energy demand projections. The econometric projections play an important role in the short term while convergence assumptions rather influence the long-term behavior. Other SSPs are calculated as variations from SSP2 long-term assumptions. In the econometric part, projections are derived for different energy carriers biomass, coal, electricity, liquids, gas, district heat in several economic sectors residential, commercial, industry, agriculture and fishery, other, non-energy use. The regressions draw on the historical relationship between the per capita energy carrier demand in each given sector and the GDP or sectoral value added per capita. The specification of the econometric model differs from one energy carrier to the other depending upon the observed relationship in historical data between the explained and the explanatory variables, or upon the regional heterogeneity. Each sectoral energy carrier is treated individually, which allows for a better control of the econometric fit, but has the disadvantage of ignoring the interdependencies between them. However, these interdependencies are partly reflected in the historical data.
In the convergence part, a global convergence line is first computed from regional econometric projections, which relates the per capita demand for the energy carrier and the per capita value added level. Each region is then assumed to converge towards this line in the long term without necessarily reaching full convergence within the time horizon of the model. The convergence assumption differs across energy carriers and sectors. Typically, demand for electricity will assume greater convergence than demand for gas, liquids or district heat, which reflects the diverse regional heating requirements.
The resulting demands were then user-adjusted to ensure that aggregated demand for energy carriers used to provide heat lies within a band of expected per-capita heat demand at a given per capita income. The projections show agreement with several energy stylized facts [1]. In line with the energy-ladder concept [2], the share of solids decreases widely, most notably due to the phase-out of traditional biomass in developing countries. By contrast, the share of grid-based energy carriers, in particular electricity, is projected to increase across all regions over the century. Following GDP per capita and population projections, developing regions’ demands grow fast, while developed regions experience a slower increase. In line with other studies, we find that currently least-developed countries will account for the bulk of global energy demand in the long-term.
SSP1 and SSP5 are derived from SSP2 intensity trajectories. There are three sets of assumptions used to derive the other SSP energy demand pathways. The first one describes how the sectoral energy intensity will evolve, the second one addresses the energy carrier intensities, and the third one the regional convergence of trajectories. The two first sets of assumptions are intended to represent scenario choices about the energy intensity development of a sector on the one hand side, and the choice of energy carriers on the other side. The convergence assumption describes the reduction of the regional spread to the global mean by 2100. Once these projections are calculated, they are aggregated to the sectoral and energy carrier levels present in REMIND. Then, the macro-economic production function of REMIND is calibrated to these energy demand pathways in the baseline scenario by adjusting the efficiency parameters at each CES level in each time step.
In policy cases, REMIND can reduce energy intensity energy service input per unit of economic output through two mechanisms.
First, the CES production function allows for price-dependent substitutions between aggregated energy and capital (substitution elasticity of 0.5). The introduction of additional constraints on the supply side (e.g., carbon taxes, resource, or emission constraints) results in higher energy prices and thus lower final energy consumption compared to the reference trajectories. As a consequence, the share of macro-economic capital input in the production function increases. In absence of distortions, a reduction in final energy results in a lower GDP and, subsequently, lower consumption and welfare values.
Second, the model can endogenously improve end-use efficiency by investing in more efficient technologies for the conversion of final energies into energy services. For example, three vehicle technologies with different efficiencies are implemented in the light duty vehicle (LDV) mode of the transport sector, including internal combustion engine vehicles, battery-electric vehicles, and fuel cell vehicles.