Centre International de Recherche

sur l’Environnement et le Développement

Supervisory authorities

CNRS Ecole des Ponts CIRAD EHESS AgroParisTech

Our partners



Home > Research > Methods and tools > Prospective modeling

Prospective modeling

by Arancha Sánchez - published on , updated on

Integrated modeling: IMACLIM

Models for long-term planning and evaluation of sustainable development policies must be able to integrate knowledge of economists, engineers, climatologists and other actors (from politicy, business, and civil society). To meet this challenge, CIRED develops a modeling architecture, IMACLIM, which integrates "engineers" visions at sectoral level in a computable general equilibrium framework taking into account the second-best mechanisms (imperfect markets, imperfect expectations ...). It comes in a static projection version, IMACLIM-S, and a recursive dynamic version, IMACLIM-R.

IMACLIM-S projects a snapshot of the economy of a country or region at a given time horizon, taking into account on the one hand the constraints due to macroeconomic interdependencies and on the other hand the constraints due to the range of technical possibilities. It allows assessing the macroeconomic and distributive impacts of a carbon constraint (taxes, quotas). Recent work has focused on France, Brazil and South Africa, but other versions are under development (China, Vietnam ...). The development of these tools is done in collaboration with local research teams.

IMACLIM-R projects the economy as a series of annual static equilibrium interlinked by the dynamics of population growth, capital accumulation and technological change. It allows projecting the pathways of evolution of technical systems, and of economic operations, taking into account the interactions between the two. Imaclim-R is used to make long-term scenarios for energy systems and evaluate policies to reduce greenhouse gas emissions. Two versions of the model exist: one on France and the other at the global level (disaggregated into 12 regions).

Researchers: Ruben Bibas, Cyril Bourgeois, Christophe Cassen, Emmanuel Combet, Frédéric Ghersi, Céline Guivarch, Meriem Hamdi-Cherif, Aurélie Mejean, Jean Charles Hourcade, Eoin O’Broin, Julie Rozenberg, Henri-David Waisman
PhD Students: Laurent Faucheux, Gaëlle Le Treut, Florian Leblanc, Julien Lefèvre, Elsa Mosseri, Jules Schers



The RESPONSE model is a compact integrated assessment model which combines an optimal growth macroeconomic model and a stylised climate model. Response was developed at CIRED, and builds on the seminal DICE model, developed by William Nordhaus. Such models combine the scientific and socio-economic dimensions of the climate issue, hence provide decision-makers with a consistent framework to evaluate climate policies and estimate the “social cost of carbon”. The originality of RESPONSE (Dumas et al. 2012) lies in (i) its flexible model structure, which allows accounting for a wide range of modelling choices for the shapes of damages (quadratic vs. sigmoid) and abatement costs (with or without inertia), (ii) its treatment of uncertainty and (iii) the nature of its decision framework (one-shot vs. sequential). For instance, the flexibility of RESPONSE allows assessing the impact of climate uncertainty and the impact of catastrophic climate change (Perrissin Fabert et al. 2012) on the optimal climate policy. RESPONSE is also used to disentangle competing arguments that have been put forward since the early 90s in academic controversies (Espagne et al. 2012) and climate negotiations. Finally, a new version of RESPONSE can be used to assess the performance of contrasted climate policies according to a wide range of social welfare functions.

Researchers: Aurélie Mejean, Patrice Dumas, Antonin Pottier, Etienne Espagne et Baptiste Perrissin-Fabert.


NEDUM-2D has been developed in CIRED to model how city-dwellers inhabitants choose to locate in a city, how policies can influence their choice, and to determine the socio-economic effects associated to these policies. The model has been designed to create long-term scenarios for city expansion, based on scenarios describing future land-use and transport policies in the city, on demographic scenarios on future total population, and on global “techno-economic” scenarios on future income, construction cost and transport cost evolution. These techno-economic scenarios can be produced through global general equilibrium prospective models, such as Imaclim-R developed at CIRED (Rozenberg et al. 2010), or Markal/TIMES developed by IEA ETSAP. NEDUM-2D can therefore be seen as a tool to downscale global scenarios at city scale.

Researchers :

Systemic modeling: SMA, SIG

Multi-agent modeling is particularly well suited to study complexity (Ferber, 1995; Bousquet and Le Page, 2004). It has already been used for issues relating to natural resource management and from an interdisciplinary perspective (Bonnabeau, 2002, Janssen, 2002; Bousquet and Le Page, 2004; Commod Platform; MAELIA project). It is original in that it connects and dovetails multiple theories, something that is necessary when addressing complexity. It is particularly well suited to take into account stakeholders, their behaviors and their interactions within a spatial system of ecologic dynamics. The behaviors of and relationships between stakeholders in the model are described by calling on the various disciplines concerned and the prevalent theories in each of the fields (biology, ecology, sociology, economics, political science) and then translated into algorithms. The Multiple Agent System has an environment and passive objects (the data can be imported from GIS). Agents act on these objects and, as they are active and independent, they also interact with their surroundings (environment, objects and other agents).

On this subject in CIRED : see SAFRAN (Coastal fringe and marine environment development scenarios: an integrated forward looking model for the Golfe du Lion Natural Marine Park Socioecosystem) project funded by Fondation de France (2015-2018)

Researchers: Catherine Boemare, Harold Levrel