How is Climate Change affecting the Weather?

The science of Climate Attribution is used to assign a climate change contribution to general weather trends and individual extreme events, but how does it work?

Saying that either climate change or natural variability caused a particular weather event is often a difficult proposition as there have always been storms, droughts, heatwaves and cold snaps. These have always happened, even before we started burning fossil fuels. So how can we determine what is natural, what is climate change driven or, as is actually the case, what proportion of the two contributed to the event?

Let’s start easily with a phenomena that nobody should disagree with. The average July temperature is warmer in southern England than northern Scotland. Even the proudest of Scotsman would agree, having pointed out that the scenery in the highlands is far superior to the Cotswolds, that Scotland does tend to be a wee bit nippy.

I’ve downloaded historic weather station data from the UK Met Office website¹ for a station near Gloucester at the top of the Bristol Channel, and from a station near Inverness on the Moray Firth 400 miles (650km) to the north.

The data contains monthly averages for temperature, rainfall, sunshine and frost days all the way back to the 1930s. But to keep it simple, I’ve concentrated on July and plotted the occurrence of temperatures for both sites between 2014 to 2024. The two curves confirm our hypothesis that it is generally warmer down south. Note however that there is considerable overlap and that it’s perfectly possible, in fact not unlikely, that sometimes it will indeed be warmer in Scotland. That’s not proof that Scotland isn’t cooler than England, it’s showing a probability that Scotland will be cooler.

We can take this a step further and look at the chances of certain temperatures and average temperature differences.

What are the relative chances of having an average July daily temperature of 21°C? It’s well within the normal area of the Bristol Channel with a likelihood of 0.7 but at the high end of temperatures in the Moray Firth with a likelihood of just 0.2. You are therefore three and a half times more likely to enjoy a nice 21°C July in the Bristol Channel than you are in the Moray Firth.

We can look at the temperatures in the same way. An unusually hot (0.2 chance) July in the Bristol Channel would be 4°C hotter than an equally unusually hot July in the Moray Firth.

Just what you’d expect and hopefully no arguments, even from our proud Scotsman.

OK let’s take a look at climate change. Using the same data source, but this time plotting 2014-2024 July temperatures for the Bristol Channel against the 1965-1975 temperature record yields a similar looking graph. The difference is clear. July temperatures are warmer now than they were 50 years ago.

An unusually hot (0.2 chance) July 50 years ago of 22.5°C is now a 0.52 chance and therefore is 2.6 times more likely today. And an unusually hot July today is 2.5°C hotter than it was 50 years ago. Unlike the geographic shift between the Bristol Channel and the Moray Firth, this shift is down to climate change. We can confidently say that based on this data, very hot days are more likely and hotter today than 50 years ago due to climate change. In addition to the curve shifting to the right, it also has a longer high temperature tail, making extremely hot temperatures more likely than extremely cold ones. The unusually hot day is 2.5°C hotter even though the average increase at the top of the curve is only 1.8°C. The extremely hot 26.3°C July which occurred in 2018 would have been almost impossible 50 years ago.

Downloading the Climate Stripes for annual temperatures in Bristol since 1850 confirms this finding. You can do this for your home town from the Show Your Stripes website².

Interestingly this change is higher than the global average surface temperature (SAT) increase between these dates. The July average global SAT between 1965 and 1975 was 0.26°C above pre-industrial and 1.17°C above pre-industrial between 2014 and 2024. A rise of 0.91°C. This is consistent with the fact that land surface temperatures are warming faster than the global average since ocean surface temperatures warm more slowly due to better heat absorption and mixing with deeper waters. It is also consistent with the fact that Europe is the fastest warming continent globally.

Moray Firth temperatures today have become quite close to Bristol temperatures 50 years ago. No wonder agricultural land prices are increasing in that area. This is also why French Champagne houses are investing in land in the Southeast of England.

This type of analysis can be done for any measure from temperature, rainfall, cloud fraction and so on. Climate Central who helped pioneer this form of analysis use it in their daily Climate Shift Index³ to show regions of the world where temperatures are made much more or less likely due to climate change. The weekend temperatures in Western Europe and the UK, western US, Amazonia, Africa, Alaska and, worryingly Greenland, were up to 5 times more likely as a result of climate change.

Individual Events

When it comes to examining the climate change contribution to individual events a more sophisticated approach is taken. World Weather Attribution (WWA)⁴ is an international consortium of scientists and meteorologists who have been performing this work since 2014. They have performed more than 100 studies on heatwaves, extreme rainfall, drought, floods, wildfires and cold spells around the world.

The method they have developed have been published in a number of peer-reviewed papers as have 26 of the attribution studies.

Once an event has been selected for study, data is collected to determine the geographical boundaries, the duration and the key variables such as maximum temperature or average rainfall. This sets the event definition and parameters.

Next observational and historic data is collected to determine how often similar events can be expected today and if they are more likely in recent decades than the past, using the basic method described above.

The second part of the study uses climate models to isolate the role of climate change in the specific event. Two ‘worlds’ are created, one with today’s warming and one with pre-industrial conditions, so without the present warming. The results are compared to quantify how the event was influenced.

All the results are then combined to provide the overarching quantitative estimate of the change in intensity and likelihood caused by climate change, together with the statistical significance or confidence in the result.

Their website has a full history of their work, Here are some excerpts from a handful of recent studies.

Heatwave

On May 15 2025, Egilsstaðir Airport in Iceland recorded 26.6°C, breaking the previous record for Iceland’s highest May temperature, while regions of the country saw temperatures more than 10°C above average (Icelandic Met Office, 2025). The Ittoqqortoormiit station in Greenland saw temperatures reach 14.3°C on May 19, which is more than 13°C above the May average daily maximum temperature of 0.8°C.

When combining the observation-based analysis with climate models, to quantify the role of climate change in this 7-day heat event, WWA concluded that climate change made the extreme heat about 3°C hotter and about 40 times more likely.

At a global warming level of 2.6°C the likelihood and intensity of such events continue to increase, at least doubling in likelihood and increasing by a further 2°C in intensity.

Flooding

Between April 2nd and 6th 2025, the Central Mississippi river valley experienced catastrophic flooding, with some areas receiving more than 400 mm (16 inches) of rain (figure 1), the worst ever recorded over this region. The ensuing floods were described as ” potentially historic” by the US National Weather Service and caused widespread damage across Mississippi, Arkansas, Missouri, Illinois, Indiana, Kentucky, Tennessee, and Alabama.

The WWA study found that the extreme rainfall event over the region was relatively rare, expected to occur in today’s climate only once every 90-240 years across different observational and reanalysis datasets. However, in a pre-industrial climate, extreme rainfall such as observed would be even rarer. The best estimates for the increase in likelihood for the 2025 event associated with current warming is between a factor 2 to 5, and the increase in intensity for an event of equivalent rarity as observed is 13-26%.

Drought

Sicily and Sardinia, the two largest Italian islands, important centres of agriculture and tourism have suffered from exceptionally low rainfall and very high temperatures over the last two years, culminating in extreme drought conditions from May 2024 onwards.

There are several ways to characterise a drought: meteorological drought is defined only by low rainfall, while agricultural drought combines rainfall estimates with evapotranspiration or directly measures soil-moisture content. As increased evapotranspiration due to regional warming can play a major role in exacerbating drought impacts, They assess agricultural drought in this study by means of the Standardised Precipitation Evapotranspiration Index (SPEI), which calculates the difference between rainfall and potential evapotranspiration to estimate the available water. The more negative the SPEI values are, the more severe the drought is classified.

For both islands, the likelihood of a drought as defined by SPEI12 from August 2023 to July 2024 has increased by about 50% due to human-induced climate change. Without climate change, both would have reduced in intensity being classified as ‘severe’ rather than ‘extreme’.

Storms

Two hurricanes hit Florida in 2024 within weeks of each other. Both intensified extremely quickly due to the high sea surface temperatures in the Gulf of Mexico. Heavy rainfall was highly impactful. The rainfall from Hurricane Helene caused considerable damage far inland. the rainfall was about 10% heavier due to climate change, and equivalently the rainfall totals over the 2-day and 3-day maxima were made about 40% and 70% more likely by climate change, respectively. If the world continues to burn fossil fuels, causing global warming to reach 2°C above pre industrial levels, devastating rainfall events in both regions will become another 15-25% more likely.

In the case of Hurricane Milton, the second of the two storms, three out of the four analysed datasets showed that heavy 1-day rainfall events are 20-30% more intense and about twice as likely in today’s climate, that is 1.3°C warmer than it would have been without human-induced climate change. The fourth dataset shows much larger changes.

By statistically modelling storms in a 1.3°C cooler climate, this model showed that climate change was responsible for an increase of about 40% in the number of storms of this intensity, and equivalently that the maximum wind speeds of similar storms are now about 5 m/s (around 10%) stronger than in a world without climate change. In other words, without climate change Milton would have made landfall as a Category 2 instead of a Category 3 storm.

Looking ahead

The UK Met Office published a paper last week using this approach to assess the future likelihood of the record breaking 40°C temperatures reached in London in 2022. They found that temperatures of this level are now 20 times more likely than in the 1960s giving a 50% chance of a repeat within 12 years. Events could even reach 45°C this decade.

Weather forecast models

World Weather Attribution studies use climate models and often an ensemble of several runs. An alternative approach is to use weather forecasting systems to replay an event from a few days out but constraining temperature data to pre-industrial norms. Alternatively historic storms could be replayed but with modern baseline temperatures to see how much worse they would have been today.

This has been carried out for Storm Ulysses, a severe extratropical storm which hit the UK in 1903. In a proof of principal study⁵, a reanalysis reconstruction was made of the original storm benefiting from rescued pressure readings taken at the time and is a credible reconstruction of one of the most extreme windstorms to occur over the British–Irish Isles in the past 120 years. The storm in the warmer-world reanalysis is constrained to be very similar to the original, except for a 2°C increase in sea surface temperature along the storms development path. Specific humidity and surface air temperatures are also higher.

The warmer-world version of the storm generates stronger maximum winds (middle row) and increased rain-fall (bottom row) compared to the original event, suggesting that the consequences of the same surface circulation pattern would be more severe in today’s climate. The increase in wind footprint represents a significant increase in damage (> 10 %; roughly related to the cube of the wind speed). The total rainfall over land during the storm increases by 26%.

This method could be applied to recent events such as the 2024 hurricane season but in 2° and 3° worlds to inform policy makers and the public of the upcoming threats and probable impacts.

From a policy perspective, a very relevant study subject would be the 1953 North Sea storm that caused devastation from Scotland to Belgium⁶. The scale of the flooding was unprecedented. In England there were 1,200 breaches of sea defences, 140,000 acres of land were flooded, 32,000 people were evacuated, 24,000 properties were damaged, 46,000 livestock were killed, and 307 people died. In the Netherlands, approximately 100,000 people were evacuated, 340,000 acres were flooded, 47,300 buildings were damaged, 30,000 livestock were killed, and 1,836 lives were lost. In addition, there were 17 deaths in Scotland, 22 in West Flanders, Belgium, and 230 in vessels at sea, including the 133 lost on the passenger ferry Princess Victoria. The storm surge peaked at 3.35 meters above the average sea level, and waves of over 4.9 meters were recorded.

Following the storm, the Dutch repaired and improved their dykes which was a huge project which took 40 years to accomplish. The new defences are 40cm higher.

What would the same storm look like in today’s, or tomorrow’s climate? Would the new walls still be high enough? At what point of warming would they now fail, +2°C, +2.5°C? Given current warming rates, how long do the authorities have until a repeat of the 1953 storm would once again be catastrophic?

As climate models and weather forecast models develop and become more closely related, this form of analysis will become more common. The ICON model system for example is being developed for the next round of CMIP and IPCC assessments and will be more vertically integrated between climate and weather applications.

Conclusions

Global warming is real despite it still snowing sometimes. It can be seen on every level from global to local trends to daily anomalies. You can get the Climate Stripes for your local city to see how it has affected your region.

Attribution studies are widely performed, often just days after an extreme event and provide the change in likelihood and intensity of the event as a result of climate change. I’d like to see this carried out routinely for every named storm or weather warning. Imagine if with every yellow or red warning, it was noted that this event is made X% more likely and Y°C-mm-classification due to climate change. It would raise public consciousness and if the public got fed up of hearing it, they may start demanding action to do something about it.

Future studies using weather forecasting systems will be able to replay past events in today’s climate as well as current events in the even warmer climate we are heading into. This could be a big help for both the public and policy makers in planning resilience and adaptation policies and investments. This is happening in some media outlets but not many. The Guardian pointed out that last weekend’s UK heatwave was made 100 times more likely due to climate change⁷. The BBC covered the Met Office study but didn’t refer to it in its coverage of the heatwave.

If we’re not going to mitigate our emissions, we must understand the implications and adapt for them, those of us lucky enough to live in places that can afford to do so that is.

1. UK Met Office Historic Data https://www.metoffice.gov.uk/research/climate/maps-and-data/historic-station-data

2. Show Your Stripes https://showyourstripes.info/l/europe/unitedkingdom/bristol

3. Climate Central - Climate Shift Index https://csi.climatecentral.org/climate-shift-index?lat=7.88515&lng=19.51172&zoom=2

4. World Weather Attribution https://www.worldweatherattribution.org

5. Hawkins, E., Compo, G. P., and Sardeshmukh, P. D.: ESD Ideas: Translating historical extreme weather events into a warmer world, Earth Syst. Dynam., 14, 1081–1084, https://doi.org/10.5194/esd-14-1081-2023, 2023.

6. Hall, Alexander. “The North Sea Flood of 1953.” Environment & Society Portal, Arcadia (2013), no. 5. Rachel Carson Center for Environment and Society. https://doi.org/10.5282/rcc/5181.

7. https://www.theguardian.com/environment/2025/jun/20/england-weekend-heatwave-worse-climate-crisis