Questions: Climate Extremes and Event Attribution

Question 1

A study reports that the 2021 Pacific Northwest heat wave was 'made at least 150 times more likely by climate change.' A news headline reads: 'Climate change caused the Pacific Northwest heat wave.' Is this headline scientifically accurate?

• A. Yes — if an event is 150 times more likely due to climate change, it essentially could not have happened without it
• B. No — attribution science quantifies how much human influence changed the probability of the event; the event remains physically possible without climate change, just far rarer
• C. No — 150 times is too high to be a credible attribution estimate
• D. Yes — but only for temperature extremes; for other event types, the headline would be inaccurate

Answer: (B)

Attribution science is probabilistic, not deterministic. 'Made 150 times more likely' means the factual world (with anthropogenic forcing) has 150x the probability of that event compared to the counterfactual world (with only natural forcings). The event could still occur naturally — just with very low probability. The headline's causal language ('caused') misrepresents the probabilistic framing that attribution science uses. This distinction matters legally and scientifically: attribution assigns increased risk, not sole causation. A fraction of attributable risk near 0.99 indicates that 99% of the risk is anthropogenic, but 1% remains natural.

Question 2

Why are heat extremes more robustly attributable to climate change than drought events?

• A. Heat extremes are more destructive, so they receive more research funding and produce clearer results
• B. Warming directly shifts the entire temperature distribution rightward, making record heat far more probable via a simple thermodynamic mechanism; droughts require modeling complex interactions among precipitation, evaporation, soil moisture, and circulation that models capture less reliably
• C. Climate models are validated for temperature but have never been tested for precipitation or soil moisture
• D. Heat extremes occur globally while droughts are regional, making global model ensembles more applicable

Answer: (B)

The directness and simplicity of the thermodynamic mechanism is the key difference. When mean temperature rises, the entire probability distribution shifts — events that were 3 standard deviations above the mean become only 2 standard deviations above the new mean, and their probability increases dramatically. This signal is large and cleanly attributable. Drought attribution involves multiple interacting variables: did precipitation change? Did evaporation increase due to warming? Did atmospheric circulation shift? Models agree well on temperature but show more spread in how they represent the full soil moisture-evaporation-precipitation feedback chain, making attribution confidence lower.

Question 3

Attribution science can definitively determine whether a specific extreme weather event was caused by climate change, giving a yes-or-no answer.

• T. True
• F. False

Answer: False

Attribution science produces probabilistic statements about changed risk, not causal verdicts. The output is a ratio of probabilities (how much more or less likely the event is with vs. without human forcing) or a change in intensity (how much more severe). Extreme weather events occur in both the factual and counterfactual worlds — attribution quantifies the difference in frequency or magnitude. The fundamental reason for this probabilistic framing is that climate is a statistical system: individual events are realizations of a distribution, and what changes with warming is the shape of that distribution, not whether individual events can occur.

Question 4

The fraction of attributable risk (FAR) framework estimates how much human greenhouse gas emissions changed the probability of an extreme event by comparing large ensembles of model simulations run with and without anthropogenic forcings.

• T. True
• F. False

Answer: True

This is the standard methodology. The 'factual world' simulations include observed greenhouse gas concentrations, aerosols, and other anthropogenic forcings; the 'counterfactual world' simulations include only natural forcings (volcanic eruptions, solar variability). By running many simulations of each scenario (an ensemble), researchers characterize the probability distribution of extreme events under each forcing condition. FAR = 1 − P(counterfactual) / P(factual). Large ensembles are necessary because extreme events are rare — you need thousands of simulated years to estimate their probability with confidence.

Question 5

Explain why attribution studies use large ensembles of model simulations rather than simply comparing the observed extreme event to the historical trend in global average temperature.

Model answer: A single trend line in global average temperature cannot establish how the probability of a specific extreme event changed, for two reasons. First, extreme events are rare — a single observation cannot estimate their probability; you need many simulated realizations to estimate how often an event of that magnitude occurs under different forcing conditions. Second, global average temperature is not what causes local extremes — regional circulation patterns, soil moisture, sea surface temperatures, and other factors interact in complex ways that only model simulations can disentangle. Ensembles provide the statistical power to compare probability distributions under factual vs. counterfactual conditions, which is the actual scientific question.

The key insight is that attribution is fundamentally a question about probability distributions, not about individual events. Individual events don't have probabilities — distributions do. The ensemble is how you access the distribution. Without it, you're limited to saying 'warming made it hotter' in the vague sense; with it, you can say 'this specific event was X times more likely,' which is a quantitative, actionable claim for policy and litigation.