October 2026 floods in Turkey likely influenced by both human-driven climate change and natural variability

Contact Authors

• Marco Zanchi, IPSL-CNRS, France, marco.zanchi@lsce.ipsl.fr

• Haosu Tang, University of Sheffield, UK, haosu.tang@sheffield.ac.uk

• Davide Faranda, IPSL-CNRS, France, davide.faranda@lsce.ipsl.fr

Citation

• Zanchi, M., Tang, H., & Faranda, D. (2026). October 2026 floods in Turkey likely influenced by both human-driven climate change and natural variability. ClimaMeter, Institut Pierre Simon Laplace, CNRS. https://doi.org/10.5281/zenodo.23502257

Press Summary

• Weather situations like the October 2026 Istanbul floods now bring about 7% heavier rainfall, about 9% stronger winds and temperatures up to 1.5°C higher than they would have in the past (1950-1987).

• The rainfall increase is significant over western and north-western Anatolia, from the Aegean hinterland to the Ankara region, and along the Black Sea coast north of Istanbul, while winds are stronger over the Aegean Sea and the Sea of Marmara. Over Istanbul itself the rainfall change is not statistically significant.

• The event was linked to a rare weather pattern, a deep upper-level low over the Aegean Sea and western Anatolia.

• Approximately 14.6 million people and 558 billion USD (about €510 billion) in economic activity were exposed within the rainfall-hazard area. Of the resulting rain-linked damages, an estimated 3% is attributable to human-driven climate change.

• We attribute the changes in the rainfall to both human-driven climate change and natural climate variability.

Figure 1. ClimaMeter analysis of the Istanbul floods (14 to 15 October 2026), over the Balkans, the Aegean Sea, the Black Sea and western Anatolia. (a, b) circulation (500 hPa geopotential height) and temperature anomalies; (c) precipitation and (d) wind speed during the event; (e, f, g, h) the present-minus-past changes in circulation, temperature, precipitation and wind speed, coloured only where statistically significant; (i) the months in which similar past events occurred; (j) changes at the selected cities (Istanbul, Gebze and Bursa), with non-significant changes set to zero. The two dials summarise the attribution (left) and how rare the pattern is (right).

Event Description

On 14 October 2026 torrential rain struck Istanbul, triggering 1,277 flooding incidents and 15 landslides across the city. The heaviest rainfall, about 194 mm, was recorded at Ömerli in the Beykoz district on the Asian side, and the governor of Istanbul described it as the heaviest rainfall in 84 years. According to the interior minister, some areas of the city received rainfall equivalent to one seventh of their annual average. Sections of road collapsed in Eyüpsultan and Gaziosmanpaşa, a river burst its banks, cars were submerged in the flooded car park of a shopping centre in Sarıyer, and traffic congestion reached about 90% of the city network during the evening rush hour.

The flooding caused multiple fatalities. A 14-year-old boy was electrocuted when he stepped into floodwater in a park in the Balat neighbourhood of Fatih, and a 34-year-old man slipped and hit his head while crossing a flooded road. The deputy mayor reported that 595 people were evacuated after becoming trapped, while more than 8,000 workers were deployed. The governor suspended in-person education across the province on 15 October, schools remained closed on 16 October as more rain was expected, and 46 planes were unable to land at the airports of Istanbul. The interior ministry also warned of sudden flooding in several other provinces, as heavy rainfall affected other Turkish cities.

The Circulation Anomalies panel (Fig. 1a), the anomaly of the 500 hPa geopotential height (the pressure pattern about 5.5 km up), shows a deep low centred over the south-eastern Aegean Sea and south-western Anatolia, more than 100 m below normal, with higher pressure over central Europe and the eastern Black Sea. The Temperature Anomalies panel (Fig. 1b) shows air 3°C to 5°C colder than usual over the Balkans and the Aegean Sea and warmer than usual over eastern Anatolia. The Precipitation Data panel (Fig. 1c) shows rainfall averaged over the two days of up to 40 to 50 mm/day around the Sea of Marmara and along the Aegean coast of Turkey, and locally more along the southern coast, while winds reached 40 to 60 km/h over the Black Sea and the Aegean Sea (Fig. 1d). At the resolution of the analysis (about 50 km), these daily values are lower than the local rain-gauge totals recorded in Istanbul.

A Summary of How ClimaMeter Works

ClimaMeter attributes an event by comparing it with its closest historical analogues, that is, past days whose large-scale weather pattern most closely resembles the event. The analogues are split into a past period (1950 to 1987) and a present period (1988 to 2025); the difference between the two sets shows how weather situations like this one have changed as the climate has warmed. All fields are detrended and deseasonalised before the search. A change is considered significant when it lies outside the range (2.5th to 97.5th percentile) of the changes obtained by randomly reshuffling the past and present analogues 1000 times. To prevent several consecutive days of the same episode from being counted as separate analogues, the retained analogues are required to be at least seven days apart. The two dials summarise, on the left, how much of the change is due to human-driven climate change versus natural variability, and, on the right, how rare the meteorological pattern is in the historical record.

Climate and Data Background for the Analysis

The IPCC Sixth Assessment Report (AR6, Working Group I, Chapter 11) concludes that heavy precipitation has become more frequent and more intense over most land regions with good observational coverage, and that human-driven warming is the main driver, since a warmer atmosphere holds about 7% more moisture per degree of warming. For the Mediterranean region, observed trends in heavy precipitation are mixed and assessed with low confidence, because of strong year-to-year variability and uneven long-term records.

Working Group I, Chapter 12 projects a drier Mediterranean in the mean, with less total rainfall and more frequent droughts, while the most intense rainfall events are expected to become heavier with further warming in many regions. The Working Group II Cross-Chapter Paper on the Mediterranean identifies flash floods in rapidly growing coastal cities, where impermeable surfaces and limited drainage amplify the effect of intense rainfall, as one of the key climate risks for the region.

ClimaMeter Analysis

We analyse how weather situations similar to the October 2026 Istanbul floods have changed between the past (1950 to 1987) and the present (1988 to 2025) over [15°E to 40°E, 32°N to 48°N], identifying analogues from the 500 hPa geopotential height pattern. Circulation Changes (Fig. 1e) are not significant over most of the domain, indicating that the upper-level low itself is not markedly different between the two periods. Temperature Changes (Fig. 1f) show present-day analogues up to about 1.5°C warmer in the area of heaviest rainfall, with significant warming of up to about 1.8°C over the eastern Mediterranean, the Levant and the southern part of the domain, and no significant change over the Balkans and Istanbul.

Precipitation Changes (Fig. 1g) show about 7% heavier rainfall in the area of heaviest rainfall, relative to the event rainfall. The increase is significant, by 1 to 4 mm/day, over western and north-western Anatolia, from the Aegean hinterland and the southern shore of the Sea of Marmara to the Ankara region, along the Black Sea coast north of Istanbul and over Thrace, eastern Bulgaria and central Greece. Over Istanbul itself the rainfall change is not statistically significant, while conditions are significantly drier over the eastern Mediterranean south of Cyprus and over Romania. Windspeed Changes (Fig. 1h) show winds about 9% stronger in the area of heaviest rainfall, with significant increases of up to about 5 km/h over the Aegean Sea, the Sea of Marmara and the western Black Sea.

Similar Past Events (Fig. 1i) occur from September to November in both periods, but present-day analogues are more frequent in September and October and less frequent in November, suggesting that these situations now occur earlier in the autumn. Changes in Urban Areas (Fig. 1j) show significantly stronger winds in all three cities, by about 3 km/h in Istanbul, 2 km/h in Gebze and 1.5 km/h in Bursa, a significant warming of about 0.7°C in Gebze, and no significant rainfall change in any of the three cities. Finally, we find that sources of natural climate variability, notably the Atlantic Multidecadal Oscillation and the Arctic Oscillation, may have influenced the event. This suggests that the changes we see in the event compared to the past may be partly due to human-driven climate change, with a contribution from natural variability. The right-hand dial points to a rare weather pattern, with clear analogues in the 1950 to 2025 record, so the analogue-based attribution carries medium-high confidence.

Exposure

To estimate the population and economic assets exposed to climate change-influenced conditions, we overlay the ClimaMeter rainfall-hazard map onto gridded datasets of population density and economic activity using the methodology developed in Faranda et al 2026 (ERL). The hazard map retains only the grid cells where present-day conditions during events like the studied one are significantly more extreme than in the historical baseline (1950 to 1987). Onto this area, we overlay gridded population and gridded gross domestic product both at a spatial resolution of about 50 km.

We then aggregate the population and economic value within the hazard area and categorize them into three severity classes. The categories are based on how unusual the local rainfall was during the event relative to the 1950-2025 record:

• Moderate: falling between the local 98th and 99th percentiles (representing the most extreme one to two per cent of days);

• Severe: falling between the 99th and 99.5th percentiles (the most extreme half to one per cent); and

• Extreme: exceeding the 99.5th percentile (the rarest and most intense conditions).

Importantly, these figures measure exposure, the people and assets in harm’s way potentially impacted, not the realized losses.

About 14.6 million people and 559 billion USD of economic activity fall within the rainfall-hazard area, out of a domain total of about 232 million people and 6.1 trillion USD. The hazard area extends mostly over western and central Anatolia, from the Aegean hinterland to the Ankara region and the Black Sea coast near Düzce, with a smaller area in central Greece; Istanbul itself lies outside it, because the rainfall change there is not significant. By severity class, roughly 2.2 million people are in the moderate class, 5.0 million in the severe class and 7.4 million in the extreme class. In economic terms, about 78 billion USD are in the moderate class, 249 billion USD in the severe class and 232 billion USD in the extreme class (Fig. S5).

Impacts and Vulnerability

The methodology for calculating the climate-linked damage share is adapted from Faranda et al. 2026 (QJRMS). In this case, each grid cell represents the geographical spread of the event. We first calculate how much more intense the hazard is today compared to historical analogues, defined as the present-minus-past difference of the analogue composites. Afterwards, we restrict this estimate to the hazard core, representing the cells where the event was most severe and where the change is statistically significant. Applying this relative intensification to the exposed population and assets splits the local impact into two components that sum to 100 per cent in every affected cell: (1) a baseline share, which the weather extreme would have caused anyway, and (2) a climate-change share, the part intensified by the human-driven warming. Because these results are fractions independent of absolute monetary losses, we report percentages only, while the values below are estimates.

We quantify the uncertainty on this attributable share from the analogue permutation test. For every grid cell, randomly reshuffling the past and present analogues 1000 times gives the spread of the present-minus-past change (the delta) expected from the finite number of analogues. We shift the delta by the distances of the 10th and 90th percentiles of that spread from its median and recompute the climate-attributable damage for each, bracketing the central estimate. Because the low and high percentiles are applied at every cell at once, the resulting range (reported in brackets) is a conservative envelope of the uncertainty in the attributed hazard change rather than a strict confidence interval.

Aggregated over the exposed assets, about 3% (10-90th percentile 1 to 4%) of the rain-linked damage is attributable to human-driven climate change (Fig. S6). Crucially, this percentage represents the climate-driven fraction of realized losses, rather than a proportion of the total exposed assets. Urban drainage capacity, soil sealing and the occupation of stream beds, which strongly shape flash-flood damage in Istanbul, are not represented in this rainfall-based estimate.

Conclusion

Weather situations similar to the October 2026 Istanbul floods now bring about 7% heavier rainfall, about 9% stronger winds and temperatures up to 1.5°C higher than in the past (1950 to 1987), with significant rainfall increases over western and north-western Anatolia and the Black Sea coast north of Istanbul, but no significant rainfall change over Istanbul itself. About 14.6 million people and 558 billion USD of economic activity lie within the rainfall-hazard area, and about 3% (10-90th percentile 1 to 4%) of the rain-linked damage is attributable to human-driven climate change.

Additional Information

The supplementary figures below are technical diagnostics. They can be understood by reading the methodology and reference literature provided by the ClimaMeter consortium at https://www.climameter.org/consortium.