Wütendes Wetter? Wie der Klimawandel Extremwetterereignisse

verändert

Dr Fredi Otto + viele andereEnvironmental Change Institute

University of Oxford

@FrediOtto

Haustein et al. 2017

Wütendes Wetter! Klimawandel?

https://www.worldweatherattribution.org

EG42CH08-Otto ARI 15 July 2017 13:15Li

kelih

ood

Climate variable

Threshold

Actualworld

Counterfactualworld

P1

P0

Figure 3Schematic of the distribution of a climatic variable under current climate conditions (red ) and in acounterfactual climate that might have been (blue). The extreme event is defined by a simple threshold(vertical dashed line), and the different probabilities of such an event occurring (red and blue shading) aremarked as P1 and P0.

2.1. The Risk-Based ApproachThe principle approach behind the probabilistic event attribution methodologies is the assessmentof possible weather events under current and preindustrial or counterfactual climate conditions toestimate the occurrence frequency of the event under different conditions. The idea is comparableto rolling dice, loaded and unloaded over and over again to identify whether and to what extendthe dice are loaded (20).

In essence, every extreme weather event is unique and always the result of a combinationof external drivers, natural and human-induced as well as internal climate variability and noise;it is therefore impossible to say that an event could not have occurred without anthropogenicinfluence. However, in the same way that loading a die can increase the likelihood of rolling a six,the presence of an external driver such as anthropogenic climate change can alter the likelihoodof the occurrence of an extreme weather event. To identify whether and to what extend this hashappened, the risk-based attribution approach simulates possible weather under current climateconditions to identify the likelihood of occurrence of an event in question in today’s climate (P1in Figure 3) and compares this with the likelihood of occurrence of the same kind of event in acounterfactual climate with the human-induced drivers removed (P0 in Figure 3).

Estimating the likelihood of occurrence of an extreme weather event and thus its return time canbe undertaken either on the basis of observed or reanalysis data (21) or on the basis of climate modelsimulations of possible weather in the current climate (22, 23). Each method has advantages anddisadvantages. Observations are less biased compared to necessarily imperfect model simulationsbut records are often short and thus require assumptions about the properties of the underlyingdistribution to be made to infer the return times of rare events. In particular, atmosphere-onlymodels can be used to simulate large ensembles, thus allowing for the statistics of rare events tobe assessed without further assumptions about the statistical properties of the distribution of theevent in question. Sippel et al. (24) compare in perfect model experiments both methods and findthat the empirical model distributions do not always match with statistical modeling on a shortersubset of the data, in particular when simulating rainfall.

2.1.1. Simulating the counterfactual. While statistics of unloaded dice are known, those ofweather and extreme events in a world without human influence are not. Hence, it is necessary

8.6 Otto

Changes may still occur before final publication

Ann

u. R

ev. E

nviro

n. R

esou

r. 20

17.4

2. D

ownl

oade

d fro

m w

ww

.ann

ualre

view

s.org

Acc

ess p

rovi

ded

by U

nive

rsity

of O

xfor

d - B

odle

ian

Libr

ary

on 0

8/28

/17.

For

per

sona

l use

onl

y.

Otto 2017

Es ist unmöglich zu sagen “dieses Ereignis hätte ohne den Klimawandel niemals stattgefunden”, aber, wir können fragen

(& antworten) ob und wenn ja wie sehr sich die Wahrscheinlichkeit für das Auftreten des Ereignisses verändert.

~40% increaseN0N1

5

10

15

20

25

30

Der Klimawandel in realen Extremereignissen

Otto et al. 2017

!"#$ !%&"' = )0)1

Dai

ly p

reci

pita

tion

2015

/16

Return time (years)

mm

per

da

y

By combining information from models and observations, we conclude that the probability of such

an event to occur for France has increased by a factor of at least 10 (see the synthesis in Figure 6).

This factor is very uncertain and could also be two orders of magnitude higher. The change in

intensity of an equally probable heatwave is between 1.5 and 3 degrees.

For Lille, results are similar. The best estimate of the return period is 78 years. The changes in

intensity are similar as for France in the models, but the observation exhibits a best estimate of

3.5°C. Changes in probabilities are also extremely large, at least a factor of ~10 and a range of

intensity increase of about 1.5°C to 3°C as seen from the synthesis in the Figure 6. However models

predict trend estimates that are inconsistent with observation trends, a fact that needs further

investigation beyond the scope of this attribution study.

We conclude that such an event would have had an extremely small probability to occur (less than

about once every 1000 years) without climate change in France. Climate change had therefore a

major influence to explain such temperatures, making them about 100 times more likely (at least a

factor of ten).

Figure 6: Changes in intensity (left panels) and probability ratios (right panels) obtained for all

models and the two stations in France. From top to bottom: France Average, Lille-Lesquin.

Germany

For Germany, we analysed Weilerswist-Lommersum, which has a time series going back to 1937 with

only two missing years. The changes in temperature are, as for France, largely underestimated by

the models compared to observations by all but the HadGEM3-A model. Based on observations and

models, we find that the effect of climate change on heatwave intensity was to elevate

temperatures by 1.5 to 3.5 degrees (synthesis in Figure 7).

Extremer Niederschlag in Texas 2019

Dürre in Somalia 2010

Hitze in Deutschland 2019

Typische Ergebnisse für verschiedene Extreme

Otto et al. 2018; worldweatherattribution.org

Mögliche Ergebnisse einer Attributionsstudie:

Aufgrund des Klimawandels hat sich die Intensität und Auftretenswahrscheinlichkeit erhöht

Aufgrund des Klimawandels hat sich die Intensität und Auftretenswahrscheinlichkeit verringert

Der menschengemachte Klimawandel hat keinen signifikanten Einfluss auf die Auftretenswahrscheinlichkeit des Ereignisses

Auf Basis unseres momentanen wissenschaftlichen Verständnisses und der aktuellen Datenlage ist es nicht möglich die Rolle des Klimawandels zu bestimmen

?

Was wir wissen

https://www.carbonbrief.org/mapped-how-climate-change-affects-extreme-weather-around-the-world

Was wir nicht wissen

Otto et al., 2020

Warum sollten wir es wissen?

… we estimate that around US$67bn of the Hurricane’s overall US$90bn are associated with climate change.

…Nordhaus’s model predicts total economic costs to the US economy in

2017, from climate change, to be around US$20bn.

Frame et al., 2020

Literaturu Haustein, K., et al. (2017) A real-time Global Warming Index. Sci

Rep 7, 15417. https://doi.org/10.1038/s41598-017-14828-5u Otto, F.E.L., et al.(2017) Climate change increases the probability

of heavy rains in Northern England/Southern Scotland like those of storm Desmond - a real-time event attribution revisited. Environmental Research Letters, doi.org/10.1088/1748-9326/aa9663

u Otto, F.E.L. (2017) Attribution of Weather and Climate Events. Annual Review of Environment and Resources, 42.

u Otto, F.E.L.,et al. (2018) Attributing high-impact extreme events across timescales—a case study of four different types of events. Climatic Change, 149(3-4): 399-412.

u Frame, D. J. et al. (2020) The economic costs of Hurricane Harvey attributable to climate change, Climatic Change, doi:0.1007/s10584-020-02692-8

u Otto, F.E.L., et al. (2020) Challenges to understanding extreme weather changes in lower income countries. Bulletin of the American Meteorological Society, doi.org/10.1175/BAMS-D-19-0317.1