Anyone who opens the climate monitor for their own town in the MeteoSwiss app sees the same picture since spring: temperature above the norm, precipitation below, sunshine above, month after month. It feels as if it has to turn at some point, as if the sky owed us a wet month. We recalculated it, using the measurement series of six stations that go back to 1864.
What the measurements show
We use Zurich-Fluntern, the long-term station near Wettingen and the Limmat valley. There, precipitation was below the average in each of the eight months from March to October, lowest in September with 14 per cent of the normal amount and in April with 18. Added up, only 52 per cent of the usual amount fell from March to October. The sun shone longer than usual in every month, in September 59 per cent more.
Show the figures as a table
| Entry | Precipitation | Sunshine duration |
|---|---|---|
| Mar | 71 % | 131 % |
| Apr | 18 % | 141 % |
| May | 41 % | 137 % |
| Jun | 82 % | 128 % |
| Jul | 32 % | 139 % |
| Aug | 71 % | 126 % |
| Sep | 14 % | 159 % |
| Oct | 84 % | 106 % |
Temperature was above the average in every month, most clearly in June with +4.0 degrees:
Show the figures as a table
| Entry | Value |
|---|---|
| Mar | +0.6 °C |
| Apr | +2.0 °C |
| May | +1.7 °C |
| Jun | +4.0 °C |
| Jul | +3.3 °C |
| Aug | +3.9 °C |
| Sep | +3.2 °C |
| Oct | +1.3 °C |
The other stations look similar, only not everywhere the same. Bern and Davos also had less precipitation than usual in every month since March. In Zermatt and Engelberg March was still at the average with 101 per cent, followed by seven months below. On the Säntis the run ended in October with 119 per cent.
All six stations as a table: precipitation, temperature and sunshine
Precipitation as a percentage of the 1991 to 2020 average
| Station | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct |
|---|---|---|---|---|---|---|---|---|
| Zurich-Fluntern | 71 % | 18 % | 41 % | 82 % | 32 % | 71 % | 14 % | 84 % |
| Bern | 75 % | 17 % | 67 % | 34 % | 26 % | 100 % | 18 % | 92 % |
| Davos | 59 % | 31 % | 70 % | 100 % | 43 % | 61 % | 42 % | 91 % |
| Zermatt | 101 % | 22 % | 44 % | 54 % | 64 % | 88 % | 20 % | 61 % |
| Engelberg | 101 % | 31 % | 74 % | 64 % | 60 % | 66 % | 26 % | 76 % |
| Säntis | 61 % | 15 % | 65 % | 87 % | 30 % | 76 % | 29 % | 119 % |
Temperature: deviation from the average in degrees
| Station | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct |
|---|---|---|---|---|---|---|---|---|
| Zurich-Fluntern | +0.6 | +2.0 | +1.7 | +4.0 | +3.3 | +3.9 | +3.2 | +1.3 |
| Bern | +0.3 | +2.2 | +1.2 | +3.7 | +3.9 | +4.1 | +2.9 | +1.4 |
| Davos | +0.8 | +2.3 | +1.5 | +2.5 | +2.4 | +3.4 | +2.9 | +1.3 |
| Zermatt | +0.2 | +3.1 | +1.6 | +3.4 | +3.2 | +3.1 | +3.5 | +1.4 |
| Engelberg | +0.7 | +2.2 | +1.4 | +3.5 | +3.3 | +3.3 | +3.0 | +1.4 |
| Säntis | +0.0 | +2.6 | +1.2 | +3.1 | +2.3 | +3.3 | +3.0 | +1.0 |
Sunshine duration as a percentage of the 1991 to 2020 average
| Station | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct |
|---|---|---|---|---|---|---|---|---|
| Zurich-Fluntern | 131 % | 141 % | 137 % | 128 % | 139 % | 126 % | 159 % | 106 % |
| Bern | 131 % | 142 % | 122 % | 116 % | 135 % | 120 % | 149 % | 89 % |
| Davos | 116 % | 140 % | 129 % | 111 % | 118 % | 111 % | 142 % | 107 % |
| Zermatt | 109 % | 137 % | 114 % | 114 % | 120 % | 98 % | 126 % | 98 % |
| Engelberg | 126 % | 134 % | 136 % | 123 % | 135 % | 117 % | 151 % | 87 % |
| Säntis | 123 % | 136 % | 133 % | 124 % | 138 % | 128 % | 143 % | 107 % |
Does a wet month have to come now?
The roulette picture does not quite hold, for two reasons.
First, “drier than the average” is not a coin with 50 per cent. In Zurich 56 per cent of all months since 1864 lie below the precipitation average, in Bern 58, in Davos 63, because a few very wet months pull the average upwards. At 56 per cent per month, eight dry months in a row is an event of about one per cent (0.56 to the power of 8), not 0.4 per cent as with a coin. And that one per cent applies to each of the nearly 1,950 months in which such a run could begin.
Second, precipitation in consecutive months is hardly connected. The correlation between a month and the month before is 0.06 in Zurich, 0.10 in Bern and −0.01 in Davos. Zero would mean no connection. The precipitation of one month says almost nothing about the next. So there is no debt that has to be paid off, and nothing that would make the next month wet. Conversely it does not mean the run must end soon: each new month is roughly another throw with a 56 to 63 per cent chance of “dry”.
This fits how often such runs actually occur. In each measurement series we counted all runs of at least eight months below the average and compared them with what would be expected if every month were dry independently, with its measured probability:
In Zurich, Davos and Zermatt the number matches chance almost exactly (9 against 8.2, 17 against 17.1, 7 against 6.7). Bern is somewhat above with 14 against 11, Engelberg below with 2 against 6.3. The weather is not entirely without memory: the drought from February 1920 to August 1921 lasted 19 months in Zurich, and that would be extremely unlikely by pure chance. It is also striking that three of the nine Zurich runs lie since the year 2000 (2003, 2018 and the current one). That may be chance or a sign of more frequent dry spells. With nine cases it cannot be decided.
The runs of all stations as a table
| Station | Months below the mean | Runs of 8+ months, observed | expected by chance | Longest run | Current run |
|---|---|---|---|---|---|
| Zurich-Fluntern | 56 % | 9 | 8.2 | 19 months (Feb 1920 to Aug 1921) | 8 months |
| Bern | 58 % | 14 | 11.0 | 12 months (Oct 1893 to Sep 1894) | 8 months |
| Davos | 63 % | 17 | 17.1 | 13 months (Aug 1884 to Aug 1885) | 8 months |
| Zermatt | 56 % | 7 | 6.7 | 13 months (Aug 1997 to Aug 1998) | 7 months |
| Engelberg | 54 % | 2 | 6.3 | 8 months (Oct 1908 to May 1909) | 7 months |
We leave out the Säntis here: its series contains noticeably more long runs than chance would predict (25 of eight months or more against about 15), and we have not established why.
And why so long warm?
With temperature the roulette picture fits even worse. The 1991 to 2020 average sits in the middle of a period, and it has warmed further since. In the years 2011 to 2025, about 68 per cent of all months in Zurich and Bern were warmer than this average, in Davos and Zermatt 71. At around 70 per cent per month, nine warm months in a row are no outlier any more (0.7 to the power of 9 is a good four per cent).
The current run is not a record either. The longest was 16 months in Zurich and Bern (May 2023 to August 2024), 14 in Zermatt (November 2021 to December 2022), 12 in Davos (June 2019 to May 2020). The current one has been running for nine months in Zurich, Bern, Davos and Zermatt, and for twelve in Engelberg.
The water has not gone
Precipitation is a quantity at one place. If it rains less here, the water has not disappeared: it rains elsewhere or not at all, and the air has not “kept” it. Where the missing rain has fallen instead, we have not examined.
Will November be drier because the summer was dry?
The thought behind it is a good one: dry soils evaporate less, the air would pick up less moisture, and autumn would turn drier. We checked it against the measurement series. For each year we set the precipitation from March to October against the precipitation in November, both as a percentage of the average.
The result is weak. In the Central Plateau and in the Valais there is no connection: the correlation is 0.03 in Zurich, 0.02 in Zermatt and 0.00 in Engelberg. In Bern (0.08), Davos (0.12) and on the Säntis (0.22) a weak one can be seen. Even the largest value (Säntis) means: the summer explains at most five per cent of the variation in November.
Calculated differently: in the years with the driest 25 per cent of the summer halves of the year, November precipitation in Bern averaged 87 per cent of the normal amount, in the other years 99. In Davos it was 75 against 96, on the Säntis 77 against 97. In Zurich it was 96 against 98, in Zermatt 97 against 87 (that is, the other way round), in Engelberg 103 against 100. Summer 2026, at 52 to 64 per cent of the usual amount, lies clearly in the driest quarter at all six stations.
All six stations as a table
| Station | Years | Mar to Oct 2026 in % of mean | Correlation | November after the driest quarter of summers | November, other years | November below mean, driest quarter / others |
|---|---|---|---|---|---|---|
| Zurich-Fluntern | 162 | 52 % | 0.03 | 96 % | 98 % | 25 of 41 / 69 of 121 |
| Bern | 162 | 54 % | 0.08 | 87 % | 99 % | 30 of 41 / 70 of 121 |
| Davos | 156 | 64 % | 0.12 | 75 % | 96 % | 32 of 40 / 67 of 116 |
| Zermatt | 133 | 56 % | 0.02 | 97 % | 87 % | 21 of 34 / 61 of 99 |
| Engelberg | 162 | 62 % | 0.00 | 103 % | 100 % | 24 of 41 / 69 of 121 |
| Säntis | 143 | 61 % | 0.22 | 77 % | 97 % | 29 of 36 / 67 of 107 |
So the answer to the question is: perhaps a little in Bern and in the mountains, not in the Central Plateau and the Valais. Whether November will be dry, the summer does not tell reliably. That is what the forecast is for, and for the first days it is clear: after the MeteoSwiss run of 4 November it stays mostly dry until 12 November: at most 2 millimetres a day at the Titlis, about 5 millimetres on 9 November on the Säntis and in Zurich. Only in the Engadine on 11 and 12 November the Piz Corvatsch gets almost 10 and a good 8 millimetres, as snow at temperatures around minus 4 degrees. That is the weather of these days, not the summer.
And the ski season?
Nothing can be concluded from all this about the winter. In our piece on the warmest September in 36 years we recalculated 35 winters: the September temperature explains four per cent of the differences between winters. We have not calculated the precipitation of the summer half of the year against the snow amount in winter here, so a connection is not proven in these data, but not ruled out either.
What a warm year does do is shift the start: without a wet-bulb temperature below minus two degrees no snow cannon runs, and that is the real risk factor for early starts. What we know about the winter as a whole is on the winter forecast, and that is, frankly, little.
Read on: The warmest September in 36 years · Winter forecast · Opening dates · All open resorts
Sources and limits
- Measurements: MeteoSwiss, SwissMetNet (ogd-smn), monthly values (“tre200m0”, “rre150m0”, “sre000m0”) of the stations Zurich-Fluntern, Bern, Davos, Zermatt, Engelberg and Säntis, up to September 2026. October 2026 is missing in the monthly file and is summed or averaged from the daily values up to 31 October. Retrieved on 4 November 2026.
- 1991 to 2020 average: calculated by ourselves from the monthly values. It can differ slightly from MeteoSwiss’s official, homogenised climate normals. For the questions about runs and ratios that is enough, not for the last tenth of a degree.
- Wettingen and Zurich: The app shows values for your own town. We use the nearest long-term station. The figures for Wettingen may differ slightly.
- Runs and chance: Every month below the average of its calendar month is counted. The expectation under independence is an approximation.
- Säntis: Left out of the runs (reason above). It is included in the November calculation.
- Forecast: MeteoSwiss, local point forecast (ogd-local-forecasting), run of 4 November 2026, 20:00 UTC.
- Checking the figures:
scripts/klima-serien.mjsandscripts/wochenende-schnee.mjsin the repository. Running them again gives the state of the day. - What is missing: Where the missing rain has fallen. A calculation on the snow amount in winter. And an explanation of why the connections in Bern and in the mountains are stronger than on the Central Plateau.
