Probabilistic Event Attribution Research Articles

Formally combining different lines of evidence in extreme-event attribution

Abstract. Event attribution methods are increasingly routinely used to assess the role of climate change in individual weather events. In order to draw robust conclusions about whether changes observed in the real world can be attributed to anthropogenic climate change, it is necessary to analyse trends in observations alongside those in climate models, where the factors driving changes in weather patterns are known. Here we present a quantitative statistical synthesis method, developed over 8 years of conducting rapid probabilistic event attribution studies, to combine quantitative attribution results from multi-model ensembles and other, qualitative, lines of evidence in a single framework to draw quantitative conclusions about the overarching role of human-induced climate change in individual weather events.

Strong Linkage Between Observed Daily Precipitation Extremes and Anthropogenic Emissions Across the Contiguous United States

AbstractThe results of probabilistic event attribution studies depend on the choice of the extreme value statistics used in the analysis, particularly with the arbitrariness in the selection of appropriate thresholds to define extremes. We bypass this issue by using the Extended Generalized Pareto Distribution (ExtGPD), which jointly models low precipitation with a generalized Pareto distribution and extremes with a different Pareto tail, to conduct daily precipitation attribution across the contiguous United States (CONUS). We apply the ExtGPD to 12 general circulation models from the Coupled Model Intercomparison Project Phase 6 and compare counterfactual scenarios with and without anthropogenic emissions. Observed precipitation by the Climate Prediction Center is used for evaluating the GCMs. We find that greenhouse gases rather than natural variability can explain the observed magnitude of extreme daily precipitation, especially in the temperate regions. Our results highlight an unambiguous linkage of anthropogenic emissions to daily precipitation extremes across CONUS.

Anthropogenic influence on the extremely low September sea ice and hot summer of 2020 over the Arctic and its future risk of occurrence

During 2020, the Arctic is marked by extremely low sea ice coverage and hot climate. September Sea Ice Extent (SIE) was about 2.3 million km2 below the 1979–2014 mean and the 2nd lowest on the 1979–2020 record, while regional summer (June–August, JJA) mean 2 m air temperature (TAS) was about 1.3 °C above the 1979–2014 mean and was the hottest on record at the time. Locally, September Sea Ice Concentration (SIC) was approximately 70% lower and JJA TAS can be as much as 6.0 °C higher than the 1979–2014 mean. Although the proximate cause for the extreme event was the continuously favorable atmospheric circulation patterns, wind conditions and ice-albedo feedback, the main objective of this paper is probabilistic extreme event attribution studies to assess the anthropogenic influence. Based on the CMIP6 multi-model ensemble products, modeled long-term trends of Arctic sea ice and TAS are consistent with observed trends when including anthropogenic forcing or greenhouse gas (GHG) forcing, while cannot exhibit observed trends with only aerosol or natural forcing. Further analysis reveals that human influence including GHG forcing has substantially increased the probability of occurrence of the 2020-like extreme events, which are rare in aerosol-only or natural-only forcing. The frequencies of 2020-like low SIE increase by 19 times with all forcing and 16 times with GHG forcing than with natural forcing. Future climate simulations under different Shared Socioeconomic Pathway (SSP) scenarios of SSP126, SSP245 and SSP585 show that the 2020-like extreme event that is currently considered rare is projected to become the norm and almost occur 1-in-1 year beyond 2041–2060. The probabilities will be approximately in the range of 0.84–1.00 for SIE and 0.76–0.99 for TAS from low emission of SSP126 to high emission of SSP585.

Representing natural climate variability in an event attribution context: Indo-Pakistani heatwave of 2022

Attribution of extreme climate events to global climate change as a result of anthropogenic greenhouse gas emissions has become increasingly important. Extreme climate events arise at the intersection of natural climate variability and a forced response of the Earth system to anthropogenic greenhouse gas emissions, which may alter the frequency and severity of such events. Accounting for the effects of both natural climate variability and the forced response to anthropogenic climate change is thus central for the attribution. Here, we investigate the reproducibility of probabilistic extreme event attribution results under more explicit representations of natural climate variability. We employ well-established methodologies deployed in statistical Earth System Model emulators to represent natural climate variability as informed from its spatio-temporal covariance structures. Two approaches towards representing natural climate variability are investigated: (1) where natural climate variability is treated as a single component; and (2) where natural climate variability is disentangled into its annual and seasonal components. We showcase our approaches by attributing the 2022 Indo-Pakistani heatwave to human-induced climate change. We find that explicit representation of annual and seasonal natural climate variability increases the overall uncertainty in attribution results considerably compared to established approaches such as the World Weather Attribution Initiative. The increase in likelihood of such an event occurring as a result of global warming differs slightly between the approaches, mainly due to different assessments of the pre-industrial return periods. Our approach that explicitly resolves annual and seasonal natural climate variability indicates a median increase in likelihood by a factor of 41 (95% range: 6-603). We find a robust signal of increased likelihood and intensification of the event with increasing global warming levels across all approaches. Compared to its present likelihood, under 1.5 °C (2 °C) of global near-surface air temperature increase relative to pre-industrial temperatures, the likelihood of the event would be between 2.2 to 2.5 times (8 to 9 times) higher. We note that regardless of the different statistical approaches to represent natural variability, the outcomes on the conducted event attribution are similar, with minor differences mainly in the uncertainty ranges. Possible reasons for differences are evaluated, including limitations of the proposed approach for this type of application, as well as the specific aspects in which it can provide complementary information to established approaches.

On the role of climate change in the 2018 flooding event in Kerala

The extreme precipitation during August 2018 in Kerala, India was catastrophic, triggering one of the worst floods in history. There is growing evidence of human-induced climate change in driving hydroclimatic extremes across the globe. However, whether and to what degree the 2018 flooding event in Kerala was influenced by climate change has yet to be fully understood. To this end, we present the first formal attribution analysis of the event, using the probabilistic event attribution (PEA) framework. Three methods using (i) Historical and HistoricalNat runs from CMIP6 (general circulation models-based method), (ii) observed records from 1901–2018 for two periods, split at 1950 (time-slice method) and (iii) observations that are scaled to 1901 and 2018 climates (scaling method), are considered for quantifying the risk ratio (RR) of the event. Using an objective approach, the 2018 precipitation event is defined by the return period of the 4 day cumulative precipitation over the Periyar River Basin (PRB), during 15–18 August, 2018. The subsequent flood event is characterized by the return period of the 1 day maximum streamflow at one of the outlets of the PRB, where maximum impact during the event was reported. The results from multiple methods are consistent, suggesting that the event is exceptionally less likely to have been caused by anthropogenic climate change, with RR for the precipitation and flood events ranging from 0.31 to 0.82 and 0.55 to 0.8, respectively. The role of wet antecedent soil moisture conditions, which is found to be the primary driving factor of floods in the PRB, is also found to be unchanged between simulations with and without climate change. Our results highlight the challenges in unequivocal discerning of the climate change signal on regional hydrological events and emphasize the importance of better consideration of local confounding interventions in PEA studies.

Attribution of the heavy rainfall events leading to severe flooding in Western Europe during July 2021

In July 2021 extreme rainfall across Western Europe caused severe flooding and substantial impacts, including over 200 fatalities and extensive infrastructure damage within Germany and the Benelux countries. After the event, a hydrological assessment and a probabilistic event attribution analysis of rainfall data were initiated and complemented by discussing the vulnerability and exposure context. The global mean surface temperature (GMST) served as a covariate in a generalised extreme value distribution fitted to observational and model data, exploiting the dependence on GMST to estimate how anthropogenic climate change affects the likelihood and severity of extreme events. Rainfall accumulations in Ahr/Erft and the Belgian Meuse catchment vastly exceeded previous observed records. In regions of that limited size the robust estimation of return values and the detection and attribution of rainfall trends are challenging. However, for the larger Western European region it was found that, under current climate conditions, on average one rainfall event of this magnitude can be expected every 400 years at any given location. Consequently, within the entire region, events of similar magnitude are expected to occur more frequently than once in 400 years. Anthropogenic climate change has already increased the intensity of the maximum 1-day rainfall event in the summer season by 3–19 %. The likelihood of such an event to occur today compared to a 1.2 ^{circ }C cooler climate has increased by a factor of 1.2–9. Models indicate that intensity and frequency of such events will further increase with future global warming. While attribution of small-scale events remains challenging, this study shows that there is a robust increase in the likelihood and severity of rainfall events such as the ones causing extreme impacts in July 2021 when considering a larger region.

Intangible Losses, Damages and At-Risk Settlements: The Extent of Causality and Burden of Proof for Climate Related Loss and Damage in the Fiji Islands

In this article, a plethora of social, cultural, legal and policy related remedies for harm caused by climate related loss and damage (L&D) were examined, particularly in rural and remote Fiji Islands. The meaning of loss and damage, and its relationship to climate mitigation and adaptation, was discussed. The concept of causality and attempt are considered to expose some criteria that the law uses to test causation in the hope to subject these to a much-needed discussion of climate change and causation affecting international law, domestic law, and climate science. A probabilistic event attribution (PEA) is explored having crucial implications in the development of PEA. When vulnerabilities and thresholds are known, changing risks can be calculated ex ante and, therefore, changing risks can be forecasted. The improvement of the methods allows geographically very specific events to be anticipated and, thus, appropriate adaptation measures can be designed. It is considered (at a conceptual level) how those harmed by loss and damage in Fiji from human-induced climate change may pursue remedies against those who have contributed to the harm. Finally, this article explores what that evidence needs to be (extent of causality and burden of proof) for loss and the damages to be awarded. It is concluded by highlighting the values of probabilistic event attribution (PEA), and how vulnerabilities in Fijian communities continue to be a deep concern. Further work needs to be done with respect to social, cultural, and biological interconnectivity that concretely underlines the importance of climate change and how it diminishes well-being and cultural integrity of Indigenous people by affecting endemic plant species. Disaster Risk Reduction (DRR) needs to be cognizant of social and cultural implications of forced migration. Causality and burden of proof within the legal context has its built-in complexities and, hence, it needs further research.

Climate change increased extreme monsoon rainfall, flooding highly vulnerable communities in Pakistan

As a direct consequence of extreme monsoon rainfall throughout the summer 2022 season Pakistan experienced the worst flooding in its history. We employ a probabilistic event attribution methodology as well as a detailed assessment of the dynamics to understand the role of climate change in this event. Many of the available state-of-the-art climate models struggle to simulate these rainfall characteristics. Those that pass our evaluation test generally show a much smaller change in likelihood and intensity of extreme rainfall than the trend we found in the observations. This discrepancy suggests that long-term variability, or processes that our evaluation may not capture, can play an important role, rendering it infeasible to quantify the overall role of human-induced climate change. However, the majority of models and observations we have analysed show that intense rainfall has become heavier as Pakistan has warmed. Some of these models suggest climate change could have increased the rainfall intensity up to 50%. The devastating impacts were also driven by the proximity of human settlements, infrastructure (homes, buildings, bridges), and agricultural land to flood plains, inadequate infrastructure, limited ex-ante risk reduction capacity, an outdated river management system, underlying vulnerabilities driven by high poverty rates and socioeconomic factors (e.g. gender, age, income, and education), and ongoing political and economic instability. Both current conditions and the potential further increase in extreme peaks in rainfall over Pakistan in light of anthropogenic climate change, highlight the urgent need to reduce vulnerability to extreme weather in Pakistan.

The role of human-induced climate change in heavy rainfall events such as the one associated with Typhoon Hagibis

Around October 12, 2019, torrential rainfall from Typhoon Hagibis caused large-scale flooding in a large area around the metropole region of Tokyo leading to large-scale destruction including losses of lives, livelihoods, and economic losses of well over $10 bn US dollars. In this paper we use a multi-method probabilistic event attribution framework to assess the role of human-induced climate change in the heavy rainfall event responsible for a large proportion of the damages. Combining different observational datasets and various climate model simulations, we find an increase in the likelihood of such an event to occur of 15–150%. We use this assessment and the calculated fraction of attributable risk (FAR) to further estimate the economic costs attributable to anthropogenic climate change based on the insured economic losses. Our conservative estimate is that ~$4bn of the damages due to the extreme heavy rainfall associated with Typhoon Hagibis are due to human-induced climate change.

Ethics of Probabilistic Extreme Event Attribution in Climate Change Science: A Critique

AbstractThe question whether a single extreme climate event, such as a hurricane or heatwave, can be attributed to human induced climate change has become a vibrant field of research and discussion in recent years. Proponents of the most common approach (probabilistic event attribution) argue for using single event attribution for advancing climate policy, not least in the context of loss and damages, while critics are raising concerns about inductive risks which may result in misguided policies. Here, we present six ethical predicaments, rooted in epistemic choices of single event attribution for policy making, with a focus on problems related to loss and damage. Our results show that probabilistic event attribution is particularly sensitive to these predicaments, rendering the choice of method value laden and hence political. Our review shows how the putatively apolitical approach becomes political and deeply problematic from a climate justice perspective. We also suggest that extreme event attribution (EEA) is becoming more and more irrelevant for projecting loss and damages as socio‐ecological systems are increasingly destabilized by climate change. We conclude by suggesting a more causality driven approach for understanding loss and damage, that is, less prone to the ethical predicaments of EEA.

Quantifying the contribution of an individual to making extreme weather events more likely

Probabilistic event attribution aims to quantify the role of anthropogenic climate change in altering the intensity or probability of extreme climate and weather events. It was originally conceived to calculate the costs associated with any increased likelihood of the meteorological event in question. However, only recently have such studies attempted to divide liability between polluting nations and ascribe a cost. Recent protests indicate a perception that older generations have the greater responsibility for climate change. In this paper, we examine how a portion of the cost of an event can be attributed to any individual person, according to their age and nationality. We demonstrate that this is quantitatively feasible using the example of the 2018 summer heatwave in eastern China and its impact on aquaculture. A relatively simple technique finds sample individuals responsible for between 0.53 and 18.10 yuan, increasing with their age and their country’s emissions over their adult lifetime since the first international consensus on carbon emissions was reached. This provides an illustration of the scale of such responsibilities, and how it is affected by national development and demographics. Such data can support decisions, at national and international levels, on how to fund recovery from climate impacts. It offers a simple quantitative approach for individuals to know their impact on the climate, or for governments to use in making policy decisions about how best to distribute costs of climate change.

Looking Forward and Backward at Extreme Event Attribution in Climate Policy

ABSTRACT How the science of probabilistic extreme event attribution might inform climate change adaptation is hotly debated. Central to these debates is an understanding that event attribution’s backward-looking orientation aligns poorly with the forward-facing goals of adaptation policy. Here, I analyze two new philosophical arguments that challenge this understanding and claim that probabilistic event attribution is not only forward-looking, but has a potentially significant role in risk-pooling adaptive strategies. I argue the purported forward-looking capabilities of event attribution are based on a mischaracterization of the scientific methodology, and one consequence of this mischaracterization is a limited role in adaptive risk-pooling schemes.

A protocol for probabilistic extreme event attribution analyses

Abstract. Over the last few years, methods have been developed to answer questions on the effect of global warming on recent extreme events. Many “event attribution” studies have now been performed, a sizeable fraction even within a few weeks of the event, to increase the usefulness of the results. In doing these analyses, it has become apparent that the attribution itself is only one step of an extended process that leads from the observation of an extreme event to a successfully communicated attribution statement. In this paper we detail the protocol that was developed by the World Weather Attribution group over the course of the last 4 years and about two dozen rapid and slow attribution studies covering warm, cold, wet, dry, and stormy extremes. It starts from the choice of which events to analyse and proceeds with the event definition, observational analysis, model evaluation, multi-model multi-method attribution, hazard synthesis, vulnerability and exposure analysis and ends with the communication procedures. This article documents this protocol. It is hoped that our protocol will be useful in designing future event attribution studies and as a starting point of a protocol for an operational attribution service.

The economic costs of Hurricane Harvey attributable to climate change

Hurricane Harvey is one of the costliest tropical cyclones in history. In this paper, we use a probabilistic event attribution framework to estimate the costs associated with Hurricane Harvey that are attributable to anthropogenic influence on the climate system. Results indicate that the “fraction of attributable risk” for the rainfall from Harvey was likely about at least a third with a preferable/best estimate of three quarters. With an average estimate of damages from Harvey assessed at about US$90bn, applying this fraction gives a best estimate of US$67bn, with a likely lower bound of at least US$30bn, of these damages that are attributable to the human influence on climate. This “bottom-up” event-based estimate of climate change damages contrasts sharply with the more “top-down” approach using integrated assessment models (IAMs) or global macroeconometric estimates: one IAM estimates annual climate change damages in the USA to be in the region of US$21.3bn. While the two approaches are not easily comparable, it is noteworthy that our “bottom-up” results estimate that one single extreme weather event contributes more to climate change damages in the USA than an entire year by the “top-down” method. Given that the “top-down” approach, at best, parameterizes but does not resolve the effects of extreme weather events, our findings suggest that the “bottom-up” approach is a useful avenue to pursue in future attempts to refine estimates of climate change damages.

Climate Change and Causation. Joining Law and Climate Science on the basis of Formal Logic

A strict application of legal tests to find the cause of an event has always been a challenge for a coherent causal analysis. This is again the case for making causal statements in the climate change context, although we are witnessing unprecedented impacts of a changing climate at a global scale. While probabilistic event attribution provides information linking greenhouse gas emission levels to observable characteristics of extreme weather and climate related events, the conventional understanding of causation in law provides a very limited response to this scientific knowledge. We offer a new matrix for developing causal explanations based on formal logic, for a coherent analysis that is compatible with climate science and law. This matrix causally explains the relation between emissions, the increase in global mean surface temperature, the general increase in frequency and severity of climate related events and, most significantly, between emissions and individual climate related events.

The warm and extremely dry spring in 2015 in Tasmania contained the fingerprint of human influence on the climate

Tasmania saw a warm and very dry spring and summer in 2015–16, including a record dry October, which had significant, wide-ranging impacts. A previous study using two probabilistic event-attribution techniques found a small but statistically significant increase in the likelihood of the record dry October due to anthropogenic influence. Given the human signal was less clear amid natural variability for rainfall compared to temperature extremes, here we provided further evidence and context for this finding. An additional attribution method supported the October rainfall finding, and the median attributable risk to human influence in the three methods was ~25%, 48% and 75%. The results suggested that human influence on rainfall was partly through increased sea level pressure in the mid-latitudes associated with fewer rainbearing systems, a circulation driver that was consistent with recent trends that have been attributed to human influence. Dry conditions were also driven by a positive Indian Ocean Dipole and El Niño at the time, but this study could not reliably estimate the effect of human influence on these phenomena, as each model gave a different estimate of the ocean warming pattern. Along with rainfall, attribution modelling showed a role for human influence in higher temperature and evaporation through October 2015, as well as a high drought index throughout spring. Confidence in the attribution of a human signal on this extreme dry event increased as multiple attribution methods agreed, a plausible atmospheric circulation driver was identified, and temperature and evaporation also showed an anthropogenic signal.

Upgrade of the HadGEM3-A based attribution system to high resolution and a new validation framework for probabilistic event attribution

Abstract We present a substantial upgrade of the Met Office system for the probabilistic attribution of extreme weather and climate events with higher horizontal and vertical resolution (60 km mid-latitudes and 85 vertical levels), the latest Hadley Centre atmospheric and land model (ENDGame dynamics with GA6.0 science and JULES at GL6.0) as well as an updated forcings set. A new set of experiments designed for the evaluation and implementation of an operational attribution service are described which consist of pairs of multi-decadal stochastic physics ensembles continued on a season by season basis by large ensembles that are able to sample extreme atmospheric states possible in the recent past. Diagnostics from these experiments form the HadGEM3-A contribution to the international Climate of the 20th Century Plus (C20C+) project and were analysed under the European Climate and Weather Events: Interpretation and Attribution (EUCLEIA) event attribution project as well as contributing to the Climate Science for Service Partnership (CSSP)-China programme. After discussing the framing issues surrounding questions that can be asked with our system we construct a novel approach to the evaluation of atmosphere-only ensembles intended for event attribution, in the process highlighting and clarifying the distinction between hindcast skill and model performance. A framework based around assessing model representation of predictable components and ensuring exchangeability of model and real world statistics leads to a form of detection and attribution to boundary condition forcing as a means of quantifying one degree of freedom of potential model error and allowing for the bias correction of event probabilities and resulting probability ratios. This method is then applied systematically across the globe to assess contributions from anthropogenic influence and specific boundary conditions to the changing probability of observed and record seasonal mean temperatures of four recent 3-month seasons from March 2016–February 2017.

Changing population dynamics and uneven temperature emergence combine to exacerbate regional exposure to heat extremes under 1.5 °C and 2 °C of warming

Understanding how continuing increases in global mean temperature will exacerbate societal exposure to extreme weather events is a question of profound importance. However, determining population exposure to the impacts of heat extremes at 1.5 °C and 2 °C of global mean warming requires not only (1) a robust understanding of the physical climate system response, but also consideration of (2) projected changes to overall population size, as well as (3) changes to where people will live in the future. This analysis introduces a new framework, adapted from studies of probabilistic event attribution, to disentangle the relative importance of regional climate emergence and changing population dynamics in the exposure to future heat extremes across multiple densely populated regions in Southern Asia and Eastern Africa (SAEA). Our results reveal that, when population is kept at 2015 levels, exposure to heat considered severe in the present decade across SAEA will increase by a factor of 4.1 (2.4–9.6) and 15.8 (5.0–135) under a 1.5°- and 2.0°-warmer world, respectively. Furthermore, projected population changes by the end of the century under an SSP1 and SSP2 scenario can further exacerbate these changes by a factor of 1.2 (1.0–1.3) and 1.5 (1.3–1.7), respectively. However, a large fraction of this additional risk increase is not related to absolute increases in population, but instead attributed to changes in which regions exhibit continued population growth into the future. Further, this added impact of population redistribution will be twice as significant after 2.0 °C of warming, relative to stabilisation at 1.5 °C, due to the non-linearity of increases in heat exposure. Irrespective of the population scenario considered, continued African population expansion will place more people in locations where emergent changes to future heat extremes are exceptionally severe.

The Value of Weather Event Science for Pending Climate Policy Decisions

This essay furthers debate about the burgeoning science of Probabilistic Event Attribution (PEA) and its relevance to imminent climate policy decisions. It critically examines Allen Thompson and Friederike Otto’s recent arguments concerning the implications of PEA studies for how the United Nations Framework Convention on Climate Change (UNFCCC) policy framework should be revised during the 2016 ‘review and decision’. I show that their contention that PEA studies cannot usefully inform decision-making about adaptation policies and strategies is misguided and argue that the current UNFCCC treaty, the ‘Paris Agreement’, supersedes their proposed revision.

The social utility of event attribution: liability, adaptation, and justice-based loss and damage

Probabilistic event attribution aims to attribute weather events to anthropogenic forcings. Some claim the development of this methodology is motivated by social utility rather than scientific understanding. I trace the evolution of arguments for probabilistic event attribution’s social usefulness from their origins in private climate change litigation through adaptive decision-making, and end with the methodology’s relevance for addressing loss and damage due to extreme events. I show that probabilistic event attribution is unlikely to substantially contribute to litigation or adaptation, and while it is potentially relevant for addressing loss and damage, securing a lasting role in this context requires answering some key questions regarding event attribution’s capacities and deployment.