Poster session
Posters on site: Tue, 29 Apr, 16:15–18:00 | Hall X5
In a work published in 2022 (Parey and Gailhard 2022, [1]) a methodology designed to estimate extreme low flow, and based on stochastic modeling has been described and tested. This methodology was suited for a single watershed and involved a single site multivariate stochastic generator of consistent temperature and rainfall timeseries. Since then, methodological issues were raised, linked on the one hand to the hydrological modeling in a cascading basins context and on the other hand to the need of being able to produce and handle an ensemble of climate projections in a reasonable computing time. The first point refers to spatial added to multivariate consistency needed in the sub-basins to obtain coherent streamflow simulations, the second to the computational efficiency of the stochastic weather generator fitting and use.
Further investigations have shown that the multivariate stochastic generation was detrimental for the performance of the extreme events reproduction, especially for long heat waves such as the 2003 event in France. Furthermore, adding spatial consistency, in addition to the multivariate one, in the generator was not straightforward. Therefore, another weather generation strategy has been proposed and tested. It consists in using single variable generators, simple for precipitation and more sophisticated in the case of temperature for the purpose of heat wave projection, used independently and synchronized a posteriori through an empirical copula coupling approach linked with bootstrapping.
After a detailed description of the proposed approach to generate a large number of spatially and mutually consistent temperature and rainfall timeseries, its application to project future low flows in a French watershed of interest for electricity generation will be demonstrated with an example.
Reference:
[1] Parey, S.; Gailhard, J.: Extreme Low Flow Estimation under Climate Change. Atmosphere 2022, 13, 164. https://doi.org/10.3390/atmos13020164
How to cite: Parey, S., Devers, A., and Gailhard, J.: A stochastic simulation strategy designed to study the future of extreme low flows in the context of electricity generation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5674, https://doi.org/10.5194/egusphere-egu25-5674, 2025.
Understanding the long-term impacts of climate change on socio-economic systems requires computationally efficient methods that integrate complex climatological and economic processes. In this study, we employ statistical emulation techniques to project the impacts of climate change on domestic tourism demand in Italy through the year 2100. Using outputs from 22 regional climate models (RCMs) produced by the Coordinated Regional Climate Downscaling Experiment over Europe (EURO-CORDEX) project under RCP 4.5 and RCP 8.5 scenarios, we develop a statistical model that combines economic indicators (e.g., GDP and exchange rates) with climate variables such as temperature, solar radiation, and precipitation.
Non-linear effects of climate on tourism demand are also incorporated. By utilising statistical emulators, we achieve computational efficiency, enabling scenario analyses across diverse emissions pathways.
This research advances impact modelling for the tourism sector, illustrating how parsimonious statistical models can bridge the gap between complex Earth system simulations and practical applications in policy and industry. By quantifying these effects based on empirical evidence and widely accepted climate change projections, this study aims to inform mitigation policies and strategies that enhance sustainability and resilience in tourism destinations facing climate challenges. Ultimately, the findings are intended to influence policymakers and entrepreneurs, emphasising the need to address the long-term impacts of climate change on tourism demand.
How to cite: Duong, N. T. T., Pons, F., D’Attoma, I., and Guizzardi, A.: Statistical Emulation of Climate Impacts on Tourism Dynamics in Italy: Long-term Projections and Policy Implications , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6759, https://doi.org/10.5194/egusphere-egu25-6759, 2025.
Anthropogenic aerosol emissions have historically exerted a net cooling effect which has masked some of the simultaneous warming from greenhouse gases (roughly -0.5°C since pre-industrial times). This mean effect is the result of heterogenous climate forcing through aerosol-radiation and aerosol-cloud interactions both locally close to emission sources and remotely via teleconnections. Future reductions and shifts in aerosol emission patterns due to regional clean air policies and shifting industrial production could therefore unmask additional warming and induce spatially complex climate impacts. Therefore, there is a need for computationally efficient tools to assess the climate impacts of possible future aerosol policy decisions.
We have developed a machine-learning emulator using Gaussian Processes (GP), trained on output from the Norwegian Earth System Model (NorESM), to predict the global spatially resolved surface temperature response to regional aerosol emission perturbations. We use a novel design for our GP model which considers the joint spatial covariance of the outputs. We show the efficacy of the emulator is comparable to that of the parent model NorESM for a fraction of the computational cost, and then use it to assess potential future aerosol emission scenarios that might be relevant to European policy decisions.
How to cite: Dewey, M., Ekman, A., Watson-Parris, D., Lewinschal, A., Samset, B., Wilcox, L., Sand, M., Seland, Ø., Krishnan, S., and Hansson, H.-C.: AeroGP: machine learning how aerosols impact regional climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6287, https://doi.org/10.5194/egusphere-egu25-6287, 2025.
Simple climate models (SCMs) are widely used to simulate global mean temperature (GMT) trajectories across a wide range of emission scenarios by combining simplified representations of the carbon cycle and other Earth system processes. These simulations depend on uncertain Earth system parameters, and ensembles of SCM simulations are created by exploring plausible parameter sets, resulting in scenario-specific distributions of GMT for all considered years.
In this work, we introduce RIME-X (Rapid Impact Model Emulator Extended), a novel emulator approach that extends SCM outputs by translating GMT distributions into distributions of regionally aggregated climate or climate impact indicators. RIME-X uses historical and scenario simulations from climate and impact modeling intercomparison projects, such as CMIP and ISIMIP, to record relationships between global warming levels and indicators. By extracting distributions of indicators at specific global warming levels from those records and combining them with the GMT distributions from SCM ensembles, RIME-X produces scenario-dependent distributions of these indicators over time.
This framework integrates multiple sources of uncertainty along the modeling chain, including model uncertainty (from diverse climate or impact model records), Earth system parameter uncertainty (from SCM ensembles), and internal variability, depending on the indicator’s temporal resolution.
RIME-X is broadly applicable to any indicator whose distribution is predominantly influenced by the global warming level, offering a versatile and efficient tool for assessing climate impacts across a variety of scenarios. We demonstrate the capabilities of RIME-X by emulating a diverse set of regionally aggregated climate and climate impact variables available from ISIMIP3 and beyond for the NGFS (Network for Greening the Financial System) climate scenarios.
How to cite: Schwind, N., Perrette, M., Lejeune, Q., Pfleiderer, P., Högner, A., Werning, M., Byers, E., Zimmer, A., Nicholls, Z., and Schleussner, C.-F.: RIME-X: Emulating regional climate impact distributions using simple climate models and impact models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7013, https://doi.org/10.5194/egusphere-egu25-7013, 2025.
The coupled interactions between components of the human-Earth system – impacts of human activity on the climate, and vice versa via climate impacts – are thought to be crucial determinants of the evolution of this system. However, the representation of these feedback loops is often minimal, or intentionally excluded, in existing integrated assessment modelling approaches.
The new global Integrated Assessment Model FRIDA v2.0 seeks to represent climate impacts as comprehensively as possible, at the global scale, focusing on high-level feedbacks between components of this system. This broad scope and high level of aggregation necessitates a reduced focus on individual impact channel complexity, with impacts simulated as functions of key global climate variables – e.g. temperature, CO2 concentration, and sea level rise.
Through this process, we have implemented key impact channels in FRIDA – on e.g. crops, energy supply and demand, mortality, and human behaviour. These channels generate substantial, complex effects on the evolution of the fully coupled human-Earth system.
In this presentation, we detail the process of collating and modelling climate impact channels within FRIDA v2.0, and present initial results of their overall effects on the system. We discuss the challenges of extracting internally consistent estimates from the literature, dealing with uncertainty across and between studies, conceptualising extremes. Finally, we discuss the need for future work to construct more comprehensive, consistent damage functions, and to coordinate their implementation in IAMs.
How to cite: Wells, C., Smith, C., Blanz, B., Ramme, L., Callegari, B., Adakudlu, M., Rajah, J., Eriksson, A., and Schoenberg, B.: Implementing global climate damage functions in a new Integrated Assessment Model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9384, https://doi.org/10.5194/egusphere-egu25-9384, 2025.
Regionalized climate risk assessments are crucial for understanding impacts on ecosystems and society, and to allow planning for climate change. While existing Earth System Models (ESMs) provide a framework for such assessments, they often lack the critical processes simulated by dedicated Impact Models. However, Impact Models are often driven by output data from ESMs, which may need bias-correcting, and therefore, there is a significant time lag in the modelling chain. Furthermore, reliance on existing ESM data for Impact Models limits our analysis to the handful of scenarios (i.e. SSPs) and models that ran them (an “ensemble of opportunity” bias), while there is a need for multiple model simulations to try to capture uncertainty in future climate.
Over the last few years, we have developed the PRIME framework for producing scenarios of regional impacts for user-prescribed future emissions scenarios. PRIME combines global mean temperature and CO2 concentrations from the emissions driven FaIR simple climate model, as used in the IPCC Sixth Assessment Report, with patterns of climate change from CMIP6 (Coupled Model Intercomparison Project Phase 6) Earth System models to drive the JULES land surface model. This modelling framework projects regional changes to the land surface and carbon cycle. We will describe PRIME for the benefit of a new audience and demonstrate how this powerful and flexible approach answers questions on regional impacts using a range of scenarios. We will also talk about the FASTMIP modelling activity led by ETH Zurich with strong contributions of the UK metoffice and PNNL, which aims to provide a coordinated experiment of regional emulators for a wide range of scenarios. We will discuss how these systems tend to be flexible and fast to run and therefore represent a wealth of future development opportunities. In particular we will focus on how PRIME and similar frameworks will enable rapid probabilistic assessment of novel scenarios emissions scenarios that have not yet been run in ESMs thereby providing a useful insight and the capability to quantify societally-relevant climate impacts.
How to cite: Mathison, C., Burke, E., Munday, G., Smith, C., Jones, C., Huntingford, C., Wiltshire, A., Kovacs, E., Steinert, N., Varney, R., Gohar, L., Windisch, M., Quilcaille, Y., Seneviratne, S., and Hooke, D.: Progress developing the PRIME framework and using it in FASTMIP. , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11014, https://doi.org/10.5194/egusphere-egu25-11014, 2025.
The PRIME emissions-to-impacts framework (Mathison et al. 2025) uses a chain of models, including the FaIR simple climate model and the JULEs land surface model, to simulate spatial resolved climate impacts and carbon cycle processes for policy relevant emissions scenarios. We present multiple updates to this framework, including a new methodology to sample large ensembles of the FaIR simple climate model, using an algorithm which maximises diversity across multiple dimensions (Sexton et al. 2021). The results are a sample with a more thorough representation of both atmospheric CO2 concentration and Global Mean Temperature. We use this sample to simulate the response of the carbon cycle under a more representative range of CO2 and temperature outcomes. In the latest version of PRIME we also include a more sophisticated representation of internal variability, and an updated daily climatology. A third methodological update is use of the PRIME framework with updated versions of JULES which include additional physical processes, such as permafrost physics and explicit representation of fire. This enables evaluation of processes not yet included in coupled Earth System Models. We use the PRIME framework in this configuration to model policy relevant overshoot scenarios, which gives us the opportunity to evaluate climate tipping points over a wide range of uncertainty. Finally, flexibility of the PRIME framework also allows us to provide driving data for other land surface models.
How to cite: Hooke, D., Mathison, C., Sexton, D., Burke, E., Wiltshire, A., Jones, C., and Gohar, L.: Evolution of the PRIME emissions-to-impacts modelling framework., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11101, https://doi.org/10.5194/egusphere-egu25-11101, 2025.
Recent studies suggest that the northern terrestrial permafrost region was a weak CO2 sink during the period 2000-2020. Future model projections remain highly uncertain – will the region remain a sink or become a source of CO2? And, if it becomes a source, when? Here we use a novel probabilistic framework PRIME (Probabilistic Regional Impacts from Model patterns and Emissions) constrained with observations to quantify a range of plausible pathways. Included are uncertainties in the global temperature response to emissions which are combined with uncertainties in spatial climate response to the global temperature change. This information is used to provide driving data for a range of JULES (the Joint UK Land Environment Simulator) configurations all of which include a representation of permafrost carbon to investigate the ecosystem carbon balance in the northern high latitudes.
How to cite: Burke, E., Varney, R., Hooke, D., Steinert, N., Smallman, L., Jones, C., Munday, G., and Mathison, C.: Northern high latitude ecosystem carbon balance under climate change, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11574, https://doi.org/10.5194/egusphere-egu25-11574, 2025.
Efficient and interpretable carbon-cycle emulators (CCEs) as part of climate emulators play a key role in Integrated Assessment Models. We present a framework enabling economists to custom-build purpose-tailored multi-reservoir linear box-model CCEs, accurately calibrated to advanced climate science. Three CCEs are presented for illustration: the 3SR model (replicating DICE-2016), the 4PR model (explicitly accounting for a land biosphere carbon reservoir), and the 4PR-X model, which accounts for dynamic land-use changes like deforestation that impact the reservoir's storage capacity and result in a time dependent CCE. We demonstrate that all three models are in line with benchmark data from comprehensive Earth System Models and exemplify how the dynamic land biosphere in the 4PR-X model impacts atmospheric carbon and temperature. The findings highlight the potential and relevance of use-cased tailored, efficient and interpretable climate emulators for economic studies. We complement our 'build your own CCE' toolbox by another set of statistical tools, commonly known as pattern scaling, that allows to go from the global mean temperature change obtained from the climate emulator to regional temperatures and changes thereof. The regional temperatures may be further translated into regional damages. We discuss the relative importance and uncertainty of each building block of this interpretable climate emulator chain, from (dynamic) CCE, to climate emulator, to pattern scaling.
How to cite: Folini, D., Eftekhari, A., Friedl, A., Kübler, F., Scheidegger, S., and Schenk, O.: Build your own! From tailored box-model climate emulators to pattern scaling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10007, https://doi.org/10.5194/egusphere-egu25-10007, 2025.
Climate change impacts are unevenly distributed, with those least responsible often bearing the brunt of its effects. This study quantifies how greenhouse gas emissions from high-income groups have influenced present-day global mean temperature levels and the frequency of temperature and potential drought extremes worldwide. We deploy an emulator-based modeling framework to systematically attribute changes in regional climate extremes to emissions from different wealth groups.
Our results show that the wealthiest 10% globally contributed about 6.5 times the global average to warming (0.40°C ± 0.16°C), while the top 1% contributed 20 times the average (0.12°C ± 0.05°C). These disproportionate contributions are further amplified for extreme events, with the top 10% contributing about 7 times more to the emergence of 1-in-100 year heat and potential drought events than the global average. Emissions from the wealthiest 10% in the United States and China are associated with a two- to three-fold increase in the frequency of heat and drought extremes across vulnerable regions. This research provides a quantitative basis for discussions on climate equity and justice by linking wealth disparities to concrete climate change impacts. Our findings have important implications for designing effective and equitable climate policies that address both mitigation and adaptation needs. The study's application of a coupled MAGICC-MESMER-M-TP framework illustrates how emulator approaches can inform policy debates on differential responsibilities and capabilities in climate action, potentially supporting more targeted and just approaches to emissions reduction and climate finance.
How to cite: Schöngart, S., Nicholls, Z., Hoffmann, R., Pelz, S., and Schleussner, C.-F.: Quantifying the Disproportionate Contributions of High-Income Groups to the Emergence of Climate Extremes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13058, https://doi.org/10.5194/egusphere-egu25-13058, 2025.
Extreme precipitation events have recently shown the potential to cause catastrophic flooding and damages, and the increasing effects of climate change are expected to intensify the associated socioeconomic and environmental risks. To facilitate the assessment of future scenarios, we extend the capabilities of the probabilistic emulator MESMER-X to represent extreme precipitation.
MESMER-X, designed to generate spatially resolved realizations of impact-relevant variables — such as annual maximum temperatures, fire weather, and soil moisture — for given global mean temperature trajectories, is adapted to emulate annual maximum daily precipitation. Using data from CMIP6 Earth System Models across various Shared Socioeconomic Pathways, the emulator captures the underlying statistical distributions and spatial patterns, enabling the exploration of customized future scenarios at a fraction of the computational cost of fully coupled Earth System Models. The performance of the emulation process is evaluated through probabilistic skill scores, residual analysis, and quantile comparisons with the original datasets.
MESMER-X outputs can support climate risk models in assessing future damages under policy-relevant scenarios, including those not previously explored with Earth System Models. This extension highlights the flexibility of MESMER-X in emulating a wide range of variables and provides a valuable support for analyzing precipitation-related climate impacts and potential targeted adaptation strategies.
How to cite: Pierini, L., Gudmundsson, L., and Seneviratne, S.: Statistical Emulations of Extreme Precipitation for Future Climate Scenarios, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16962, https://doi.org/10.5194/egusphere-egu25-16962, 2025.
Sea-level rise simulation has previously been limited to Earth system models and global emulators - restricting spatially-resolved sea-level projections to those based on ageing emissions pathways with inflexible and expensive frameworks for updating projections using the latest scenarios. The ProFSea (Projecting Future Sea-level) tool improved on AR5 methods for fast regional sea-level prediction, but was limited to RCP scenarios and a 21st century timescale. We use the FaIR simple climate model to generate an ensemble of global surface temperatures from a range of policy-relevant scenarios, and drive a global sea-level rise simulator. The global projections are then localised using spatial patterns (derived from model estimates and observational evidence) related to key sea-level change drivers. Uncertainty is quantified and propagated throughout the modelling chain. We present the evaluation of this enhanced version of the ProFSea sea-level projections tool, and demonstrate its utility as a policy tool for predicting local sea-level change risk through the 21st century, out to 2300.
How to cite: Munday, G., Palmer, M., Perks, R., Allison, L., Weeks, J., Smith, C., and Gregory, J.: Advancing ProFSea: a spatially-resolved sea-level change emulator for long-term impacts , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14775, https://doi.org/10.5194/egusphere-egu25-14775, 2025.
Introducing METEOR (Multivariate Emulation of Time-Evolving and Overlapping Responses) - a spatially resolved impacts emulator. P. Spatially resolved emulators can produce such data with a fraction of the computational cost required by full Earth system models, allowing the exploration of a much richer scenario space.
METEOR uses Earth system model output to emulate impact response patterns of varying decay timescales to forcing changes. As such, METEOR allows for the projection of future climate changes, including modelling of hysteresis in overshoot scenarios. In-built emissions to forcing mapping enables a full chain emulation of impact variables from emissions scenarios to spatially resolved impacts. METEOR can emulate multiple independent forcer responses, relying on at least one abrupt-CO2-change experiment as training data, and using either more abrupt forcer change experiments or a residual technique to emulate additional responses. This presentation will describe the model and its design philosophy and show results for emulations of CMIP6 model yearly mean temperature and precipitation. The flexibility of the framework allows application to a wide range of other more impact specific variables, and in addition the emulation patterns and timescales in themselves may reveal interesting patterns in the emulated data.
How to cite: Sandstad, M., Sanderson, B., and Steinert, N.: METEOR - a spatially resolved impacts emulator, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3037, https://doi.org/10.5194/egusphere-egu25-3037, 2025.
