Les nouveaux scénarios pour l’analyse des politiques climatiques Stéphane Hallegatte CIRED, Météo-France, Banque Mondiale

Des scénarios socio-économiques pour évaluer le coût du changement climatique Les impacts du changement climatique et les coûts d’adaptation dépendent de l’intensité du changement climatique. Mais aussi des évolutions socio-économiques: Quelle part de la population vit dans l’extrême pauvreté? Quelle est la part de l’agriculture vivrière dans les pays pauvres? Quelle part de la population vit en ville? De quelles technologies d’adaptation disposons-nous? …

Des scénarios socio-économiques pour évaluer le coût des politiques de réduction des émissions Les coûts des réductions d’émissions dépendent aussi des évolutions socio-économiques: Quelle est la population mondiale? Quelles sont les technologies disponibles? Quels modes de consommation? Quels développements économiques? Quelles réserves de pétrole et de charbon? …

Les SRES

Pourquoi de nouveaux scénarios? On a besoin: De scénarios qui tiennent compte des nouvelles évolutions socio-économiques De scénarios avec des baselines et des scénarios de réduction d’émission De scénarios pertinents pour l’étude de l’adaptation et les impacts De scénarios qui vont au delà de 2100 On construit un produit (des scénarios), mais surtout un processus pour organiser la recherche dans le domaine.

Forçage radiatif en W/m2 16/11/2018 Les « Representative Concentration Pathways » ( trajectoires de concentrations représentatives) Forçage radiatif en W/m2 11/16/2018 6

L’architecture “en matrice”

Permet de comparer les conséquences du même changement climatique dans différents scénarios socio-économiques

Permet de comparer le coût de politiques de réduction des émissions qui atteignent la même cible, dans différents scénarios socio-économiques Référence SSP3 Référence SSP1 Shared Policy Assumption Shared Policy Assumption

Les coûts de l’atténuation

Le coût de l’adaptation et les impacts résiduels

Des scénarios pour explorer l’incertitude sur les coûts d’atténuation et les impacts du CC

Couvrir les différents futurs possibles pour explorer l’incertitude

Les SSPs, d’après la réunion de Boulder (Nov 2-4, 2011)

Couvrir les différents futurs possibles pour explorer l’incertitude

SSP1 Haute capacité d’adaptation, Haute capacité d’atténuation Le monde est caractérisé par la coopération internationale, et le développement durable est une priorité Monde fortement régulé, orienté vers l’environnement et le développement.

SSP2 Capacité d’adaptation moyenne, capacité d’atténuation moyenne Un monde caractérisé par la continuation des tendances actuelles Nombreuses vues contradictoires sur les déterminants de ce scénario Plusieurs SSP2???

SSP3 Faible capacité d’adaptation, faible capacité d’atténuation Un monde caractérisé par la compétition entre pays, et par une croissance économique lente. Politiques orientées vers la sécurité, barrières commerciales, peu d’efforts pour protéger l’environnement.

SSP4 Faible capacité d’adaptation, Haute capacité d’atténuation Monde très inégal, à l’intérieur et entre les pays. Une petite élite mondiale est responsable de l’essentiel des émissions de GES. La plus grande partie de la population reste pauvre et vulnérable au changement climatique. Governance et mondialisation sont efficaces, mais contrôlées par les élites, et à leur service.

SSP5 Haute capacité d’adaptation, Faible capacité d’atténuation Développement “conventionnel”, orienté vers la croissance économique et les MDGs. La préférence pour une croissance rapide conduit à une forte dépendance aux carburants fossiles. La croissance permet d’atteindre des MDGs et de fournir les services nécessaires aux populations des pays en voie de développement.

A formal method to build socio-economic scenarios Stéphane Hallegatte Avec Julie Rozenberg, Céline Guivarch, et Robert Lempert CIRED, Météo-France, RAND Corporation

SSP framework

Scenario elicitation methodology Identify drivers Identify a priori the main drivers of future capacity to adapt and mitigate Model scenarios Translate drivers into model parameters Build a large number of model runs Add “quantitative narratives” when necessary Select scenarios Build relevant indicators Select contrasted SSPs Backward approach

Phase 1 Identify drivers Identify a priori the main drivers of future capacity to adapt and mitigate Model scenarios Translate drivers into model parameters Build a large number of model runs Add “quantitative narratives” when necessary Select scenarios Build relevant indicators Select contrasted SSPs

Carbon dependence (high vs. low dependence) Environmental stress (environmentally-stressed vs. environmentally-friendly) Carbon dependence (high vs. low dependence) Use of natural resource Availability of fossil energy Consumption behaviors Localization choices Technologies Urbanization Capacity to adapt Capacity to mitigate Industrial and commercial policies Governance efficiency Extreme poverty reduction Economic structure Population Social protection Capital markets Globalization (convergence vs. fragmented) Labor markets Equity (inclusive growth vs. growth-and-poverty)

Phase 2 4 dimensions 13 drivers Identify drivers 4 dimensions 13 drivers Model scenarios Translate drivers into model parameters Build a large number of model runs Add “quantitative narratives” when necessary Select scenarios Build relevant indicators Select contrasted SSPs

Availability of fossil energy Use of natural resource Consumption behaviors Technologies Urbanization Governance efficiency Industrial and commercial policies Extreme poverty reduction Economic structure Social protection Population Labor markets Capital markets 31

Descriptors translated into input parameters of the IMACLIM-R model Availability of fossil energy Use of natural resource Consumption behaviors Technologies Urbanization Governance efficiency Industrial and commercial policies Extreme poverty reduction Economic structure Social protection Population Labor markets Capital markets 32

This process can be done through a participatory approach Descriptors translated into input parameters of the IMACLIM-R model Availability of fossil energy Use of natural resource Consumption behaviors Technologies Urbanization This process can be done through a participatory approach (see Sandrine’s presentation) Governance efficiency Industrial and commercial policies Extreme poverty reduction Economic structure Social protection Population Labor markets Parameter that is part of the narratives but does not appear in the model Capital markets 33

288 scenarios with ImaclimR

Phase 3 4 dimensions 13 drivers 288 scenarios Identify drivers 4 dimensions 13 drivers Model scenarios 288 scenarios Select scenarios Build relevant indicators Select contrasted SSPs

CO2 emissions in the baseline Choosing indicators CO2 emissions in the baseline GDP per capital of the 20% poorest in a selection of developing countries

288 scenarios & 2 indicators

Definition of 5 SSP spaces with 4 thresholds The definition of SSP spaces is arbitrary. Here we choose thresholds so that 1/3 of the scenarios are below the first threshold and 1/3 are above the second threshold. That’s how Rob defines spaces in his scenario discovery analysis. The PRIM (Patient Rule Induction Method) analysis is used to find out the main drivers that best explain why scenarios are inside a box (density), but also why they are not outside the box (coverage). Density is the fraction of scenarios that are in the box and associated with the SSP drivers. Coverage is the fraction of all scenarios with the SSP drivers and contained in the box

of low C technologies (2 options) Availability of fossil fuels Main drivers according to the PRIM analysis Equity (2 options) Conver-gence (3 options) Energy sobriety Availability of low C technologies (2 options) Availability of fossil fuels Population (3 options) Capital markets Coverage/Density SSP1 (15% of cases) improved Fast or medium high Medium or low 50% / 80% SSP2 (10% of cases) Medium or slow low 30% / 60% SSP3 (14% of cases) worsen High or medium 55% / 90% SSP4 (8% of cases) slow 90% / 85% SSP5 (6% of cases) fast Reduced imbalances 60% / 45% Equity contribute strongly to all the SSPs because this driver has a direct impact on the capacity to adapt axis, since it was used to calculate the indicator the “energy sobriety” driver has a strong impact on the capacity to mitigate, since it directly influences CO2 emissions in the baseline. It also influences the capacity to adapt because energy sobriety leads to higher GDP, i.e. to less poverty. (In scenarios with high energy sobriety, energy prices are lower, accelerating GDP growth) The impact of population on the indicators is ambiguous and not always significant. Indeed, a higher population growth rate implies higher potential economic growth in the model, so that adaptation capacity might increase. Moreover, higher economic growth accelerates capital turnover and increases the share of low-carbon technologies, thus increasing mitigation capacity. The results show, however, that a high population is inconsistent with SSP1 and that a low population is inconsistent with SSP3 The non significant impact of fossil fuel availability is due to two contradictory effects: On the one hand, a constrained oil supply induces substitution toward coal, which emits more CO2 for the same energy service. On the other hand, it also induces higher energy prices, which trigger faster energy efficiency. In the same way, low-carbon technologies contribute to only two SSPs because they tend to slow down energy efficiency through lower energy prices, which lessens their effect on carbon emissions.

The issue of relevant indicators = GDP cost to reach 550ppm The issue of relevant indicators = share of jobs in agriculture

of low C technologies (2 options) Availability of fossil fuels Equity (2 options) Conver-gence (3 options) Energy sobriety Availability of low C technologies (2 options) Availability of fossil fuels Population (3 options) Capital markets Coverage/Density SSP1 (15% of cases) Fast or medium high Medium or low Reduced imbalances 50%/80% SSP2 (10% of cases) Medium or slow low 30%/60% SSP3 (14% of cases) High or medium 55%/90% SSP4 (8% of cases) slow Constant imbalances 90%/85% SSP5 (6% of cases) fast 60%/45% high high medium medium High or medium low low slow Capital markets have a bigger influence on structural change, so on the share of jobs in agriculture. Low carbon tech have a bigger impact on the cost of mitigation policies than on CO2 emissions. fast low

Downscaling at local scale: an application on Paris Stéphane Hallegatte Avec Vincent Viguié et Julie Rozenberg CIRED et Météo-France

Modelling urban form over the very long term? The NEDUM2D model Standard urban economics modelling (Alonso 1964, Mills 1967). Three mechanisms : Household tradeoff: Lower transportation costs and shorter commuting time when living close to the city center; and Larger dwellings and lower rent per square meter in remote areas Investors optimize housing density as a function of rents and construction costs. Different evolution timescales for rents, population density, buildings etc. Simplifying hypotheses : All households have the same income; One trip per day towards the city center; One city center.

VALIDATION on 1900-2010

Paris built area, 2006 Validation process We run the model from 1900 to 2010 using: Data on population; Data on average income; Data on transportation cost, speed, and localization; Construction costs change like income. 16 novembre 2018

Model dynamics validation on the built area, 1900 16 novembre 2018

Model dynamics validation on the built area, 1960 16 novembre 2018

Model dynamics validation on the built area, 1990 16 novembre 2018

Model dynamics validation on the built area, 2006 16 novembre 2018

Validation on local rents

Validation on local density

2010-2100 scenarios

Main hypotheses 4 world scenarios with two dimensions: Tensions on fossil fuel markets (resources, population, technology) Ambitions of world climate policies. High or low demographic scenarios for Paris urban area population

Paris urban area population : 2 scenarios

Main hypotheses 4 world scenarios with two dimensions: Tensions on fossil fuel markets (resources, population, technology) Ambitions of world climate policies. High or low demographic scenarios for Paris urban area population One simple scenarios for transport policies: Infrastructure remains unchanged between 2010 and 2100 Congestion on roads and public transport remains at current levels Two simple options for land use planning: No policy (autonomous or “potential” urbanization) A green belt policy

Example of Paris urban area extension prospective scenario 16 novembre 2018

Paris built area extension – Urban sprawl Increase with high population Increase and then decrease with low population Stable with green belt Small impact of technology and fuel prices scenarios. Only local policies can control urban sprawl (e.g., a green belt).

Input for mitigation policy analysis: Example: impact of a €100/tC carbon tax Monthly rent variation (€/m²/month)

Ttransport-related GHG emissions in Paris Small impact of local policies (green belt or not…) Local policies cannot do much on GHG emissions (unless…)

AC high-density scenario Input for adaptation policy analysis: Adapting to high temperature and air conditioning Vulnerability to the 2003 heat wave, depending on urban forms and the use of AC. Heat stess (outdoor, shadow) in number of hours. Current Paris AC high-density scenario AC low-density scenario 1. Using AC has a negative impact on outdoor heat stress. 2. A high-density city appears more vulnerable to heat wave than a low-density city.

More ? On the scenario architecture: http://www.isp.ucar.edu/socio-economic-pathways