Heavy Rain and Global Warming
Wettest day of the year at 9,760 weather stations, 1950 to 2025: +6.5 percent per degree C of global warming (interval +5.1 to +8.3), against the 7 percent of the Clausius-Clapeyron relation. Wettest five days: +6.3 percent. By hemisphere, latitude band, continent and country, with the yearly total, the number of wet days and robustness checks.
Data as of . Last revised (2026-11-17).
Key Facts
- From 1950 to 2025 the wettest day of the year rose by 6.5 percent per degree C of global warming at 9,760 stations in 430 grid cells (95 percent interval +5.1 to +8.3 percent). The interval lies above zero, so an increase is clear. The interval holds 7 percent, so the rate cannot be told apart from the Clausius-Clapeyron rate.
- From 1950 to 2025 the wettest five days of the year rose by 6.3 percent per degree C of global warming at 9,760 stations in 430 grid cells (95 percent interval +4.6 to +7.4 percent). The interval lies above zero, so an increase is clear. The interval holds 7 percent, so the rate cannot be told apart from the Clausius-Clapeyron rate.
- At 63.3 percent of the 9,760 stations the station's own rate for the wettest day of the year, 1950 to 2025, is above zero; 50 percent is expected with no change.
- At the 4,810 stations within 40 degrees of the equator and below 1,000 m (3,281 ft), where little of the year's precipitation falls as snow, the yearly total rose by 4.1 percent per degree C of global warming from 1950 to 2025 (95 percent interval +1.4 to +6.2 percent). At the other 4,950 stations, further from the equator or higher up, it rose by 9.6 percent per degree C of global warming (+7.0 to +10.7 percent). The two intervals do not overlap: the yearly total is clearly higher where snow falls. A gauge catches snow far less well than rain, so stations there also report more when winters bring less snow and more rain, or when gauges change, whether or not more falls. So the yearly total at those stations is not reliable evidence of more precipitation.
- Station by station at the low-snow stations, the wettest day of the year grew 3.9 percent per degree faster than the yearly total from 1950 to 2025 (95 percent interval +1.0 to +5.7 percent, 4,810 stations in 137 grid cells). The interval lies above zero, so the difference is clear.
- For 1950 to 2025, none of these checks changes the sign of the result or moves it to half or twice the published rate.
- Of the 9,760 stations, 4,306 (44.1 percent) are in the United States, 2,507 (25.7 percent) in Australia and 531 (5.4 percent) in Russia; the stations are spread over 52 countries. The network of 1950 to 2025 is densest there, so every region is weighted by area and not by its number of stations.
The Rate per Degree
1950 to 2025Warmer air can hold more water vapour: the amount it holds when saturated rises by about 7 percent for every degree C (3.8 percent per degree F), a law of physics called the Clausius-Clapeyron relation. If the heaviest rain of a year draws on all the vapour the air can hold and nothing else changes, the wettest day should grow at about that rate as the world warms.
This page measures the rate at weather stations. For every station with at least 50 years of daily readings since 1950, the wettest day of each year is set against the global mean temperature of that year. The station rates are then pooled by area, so that countries with many stations do not decide the result.
| Measure | Stations | Grid cells | Percent per degree C | 95% interval | Against no change | Against 7 percent |
|---|---|---|---|---|---|---|
| Wettest day of the year | 9,760 | 430 | +6.5% | +5.1 to +8.3 | above zero | holds 7 |
| Wettest five days of the year | 9,760 | 430 | +6.3% | +4.6 to +7.4 | above zero | holds 7 |
From 1950 to 2025 the wettest day of the year rose by 6.5 percent per degree C of global warming at 9,760 stations in 430 grid cells (95 percent interval +5.1 to +8.3 percent). The interval lies above zero, so an increase is clear. The interval holds 7 percent, so the rate cannot be told apart from the Clausius-Clapeyron rate.
From 1950 to 2025 the wettest five days of the year rose by 6.3 percent per degree C of global warming at 9,760 stations in 430 grid cells (95 percent interval +4.6 to +7.4 percent). The interval lies above zero, so an increase is clear. The interval holds 7 percent, so the rate cannot be told apart from the Clausius-Clapeyron rate.
The global temperature yardstick rose by 1.10 C (1.98 F) between 1951 to 1980 and 2025, so a rate of +6.5 percent per degree amounts to about +7.2 percent on the wettest day of the year over that warming.
Year by Year Against the Global Temperature
1950 to 2025| Year | Wettest day, pooled stations | 7 percent per degree | Measured rate, +6.5 percent per degree |
|---|---|---|---|
| 1950 | +0.3% | -0.4% | -0.4% |
| 1951 | -1.5% | -0.4% | -0.4% |
| 1952 | -1.9% | -0.3% | -0.3% |
| 1953 | +2.2% | -0.1% | 0.0% |
| 1954 | +4.5% | 0.0% | 0.0% |
| 1955 | +3.5% | -0.2% | -0.1% |
| 1956 | -4.9% | -0.5% | -0.5% |
| 1957 | -1.2% | -0.6% | -0.5% |
| 1958 | +0.7% | -0.3% | -0.3% |
| 1959 | +0.3% | 0.0% | 0.0% |
| 1960 | +1.5% | +0.3% | +0.3% |
| 1961 | -1.3% | +0.3% | +0.3% |
| 1962 | -4.2% | +0.3% | +0.2% |
| 1963 | -0.6% | +0.2% | +0.2% |
| 1964 | +1.1% | -0.1% | -0.1% |
| 1965 | -4.7% | -0.3% | -0.3% |
| 1966 | +0.6% | -0.5% | -0.4% |
| 1967 | +0.3% | -0.6% | -0.5% |
| 1968 | +0.4% | -0.4% | -0.3% |
| 1969 | -1.4% | -0.1% | -0.1% |
| 1970 | -2.6% | +0.1% | +0.1% |
| 1971 | +1.2% | -0.1% | -0.1% |
| 1972 | -0.7% | +0.1% | +0.1% |
| 1973 | +4.2% | +0.2% | +0.2% |
| 1974 | +1.5% | +0.1% | +0.1% |
| 1975 | +1.1% | +0.2% | +0.2% |
| 1976 | -0.7% | 0.0% | 0.0% |
| 1977 | +0.3% | +0.1% | +0.1% |
| 1978 | -0.9% | +0.4% | +0.3% |
| 1979 | +4.2% | +0.7% | +0.6% |
| 1980 | -0.1% | +1.3% | +1.2% |
| 1981 | +3.0% | +1.5% | +1.4% |
| 1982 | +3.9% | +1.6% | +1.5% |
| 1983 | +1.1% | +1.8% | +1.7% |
| 1984 | +0.7% | +1.6% | +1.5% |
| 1985 | -0.7% | +1.3% | +1.2% |
| 1986 | -2.1% | +1.4% | +1.3% |
| 1987 | +1.7% | +1.4% | +1.3% |
| 1988 | +3.8% | +1.8% | +1.6% |
| 1989 | +3.4% | +2.0% | +1.8% |
| 1990 | +2.6% | +2.3% | +2.1% |
| 1991 | +3.7% | +2.4% | +2.3% |
| 1992 | +1.1% | +2.2% | +2.0% |
| 1993 | +3.0% | +2.2% | +2.0% |
| 1994 | +0.4% | +2.0% | +1.9% |
| 1995 | +3.4% | +2.1% | +2.0% |
| 1996 | +5.4% | +2.3% | +2.1% |
| 1997 | +3.2% | +2.7% | +2.5% |
| 1998 | +3.6% | +3.2% | +3.0% |
| 1999 | +1.8% | +3.1% | +2.9% |
| 2000 | +8.7% | +3.2% | +3.0% |
| 2001 | +1.2% | +3.3% | +3.0% |
| 2002 | +0.3% | +3.2% | +3.0% |
| 2003 | +3.0% | +3.6% | +3.3% |
| 2004 | +4.7% | +3.8% | +3.6% |
| 2005 | +5.9% | +4.1% | +3.8% |
| 2006 | +3.0% | +4.1% | +3.8% |
| 2007 | +5.6% | +4.2% | +3.9% |
| 2008 | +3.3% | +4.2% | +3.9% |
| 2009 | +1.5% | +4.1% | +3.8% |
| 2010 | +8.2% | +4.3% | +4.0% |
| 2011 | +6.6% | +4.2% | +3.9% |
| 2012 | +6.3% | +4.4% | +4.1% |
| 2013 | +3.4% | +4.4% | +4.1% |
| 2014 | +7.1% | +4.5% | +4.2% |
| 2015 | +4.7% | +5.0% | +4.6% |
| 2016 | +3.8% | +5.6% | +5.2% |
| 2017 | +6.4% | +6.1% | +5.7% |
| 2018 | +3.1% | +6.3% | +5.9% |
| 2019 | -0.2% | +6.5% | +6.1% |
| 2020 | +4.3% | +6.6% | +6.1% |
| 2021 | +8.7% | +6.4% | +6.0% |
| 2022 | +5.8% | +6.5% | +6.0% |
| 2023 | +0.4% | +6.8% | +6.3% |
| 2024 | +6.9% | +7.3% | +6.7% |
| 2025 | n/a | +7.7% | +7.2% |
In the ten years 2015 to 2024 the pooled wettest day averaged +4.4 percent against its 1951 to 1980 mean, while the yardstick stood 0.90 C (1.63 F) above its 1951 to 1980 level. Single years scatter widely around both lines, because the wettest day of a year depends on a few storms.
A straight line through the 75 yearly values of 1950 to 2025 against the global temperature gives +5.9 percent per degree C (95 percent interval +4.1 to +7.7 percent). That interval comes from the scatter of the years, the one in the table above from the scatter between grid cells; the two are separate ways to the same question.
A year is drawn when its cells hold at least 55 percent of the area weight of all cells. Reports from outside the United States arrive late, so the last years can have fewer stations: the lowest coverage of the last five years is 48 percent in 2025, with 5,515 stations.
Where More, Where Less
1950 to 2025| Cell centre | Stations | Wettest day, percent per degree C |
|---|---|---|
| 27.5 S, 152.5 E | 309 | -6.6% |
| 52.5 N, 7.5 E | 264 | +12.0% |
| 32.5 S, 117.5 E | 259 | -5.2% |
| 37.5 S, 142.5 E | 232 | -2.7% |
| 32.5 S, 137.5 E | 219 | +6.6% |
| 32.5 S, 152.5 E | 217 | -8.1% |
| 37.5 S, 147.5 E | 215 | -1.2% |
| 37.5 N, 97.5 W | 215 | +8.7% |
| 37.5 N, 82.5 W | 201 | +6.4% |
| 42.5 N, 92.5 W | 190 | +12.3% |
| 32.5 S, 147.5 E | 170 | -9.7% |
| 42.5 N, 97.5 W | 170 | +13.2% |
| 37.5 N, 87.5 W | 160 | +10.0% |
| 37.5 N, 92.5 W | 159 | +15.9% |
| 32.5 N, 97.5 W | 157 | +10.1% |
| 32.5 N, 87.5 W | 154 | +9.1% |
| 37.5 N, 77.5 W | 153 | +10.2% |
| 42.5 N, 72.5 W | 151 | +18.7% |
| 32.5 N, 92.5 W | 149 | +10.0% |
| 42.5 N, 87.5 W | 149 | +10.5% |
| 42.5 N, 77.5 W | 135 | +11.8% |
| 32.5 N, 82.5 W | 132 | +5.1% |
| 42.5 N, 82.5 W | 125 | +12.3% |
| 52.5 N, 12.5 E | 116 | +3.5% |
| 37.5 N, 102.5 W | 115 | +6.5% |
| 47.5 N, 122.5 W | 113 | +9.5% |
| 42.5 S, 147.5 E | 109 | -1.1% |
| 52.5 N, 2.5 E | 108 | +16.5% |
| 37.5 N, 122.5 W | 106 | +3.4% |
| 57.5 N, 12.5 E | 105 | +14.7% |
| 42.5 N, 102.5 W | 99 | +3.2% |
| 37.5 N, 107.5 W | 97 | +2.2% |
| 32.5 N, 102.5 W | 91 | +3.1% |
| 42.5 N, 122.5 W | 91 | -12.1% |
| 47.5 N, 7.5 E | 91 | +1.1% |
| 42.5 N, 112.5 W | 89 | +5.3% |
| 22.5 N, 102.5 W | 87 | +12.2% |
| 47.5 N, 102.5 W | 84 | +5.8% |
| 27.5 N, 97.5 W | 82 | +7.2% |
| 22.5 S, 147.5 E | 81 | -4.8% |
Of the 430 grid cells, 306 have a median rate above zero and 217 above 7 percent per degree for 1950 to 2025; of the 276 cells with at least 3 stations, 210 and 143. A single cell can lie far from the pooled rate by chance alone, so the tables below pool the cells of a region and say which regions differ from 7 percent after allowing for how many regions are compared.
| Latitude band | Wettest day, percent per degree C | Lower end of the interval | Upper end of the interval | 7 percent per degree |
|---|---|---|---|---|
| 30 to 40 S | -1.2% | -5.2% | +4.1% | +7.0% |
| 20 to 30 S | -2.5% | -8.1% | +3.5% | +7.0% |
| 10 to 20 S | +3.8% | +1.4% | +11.6% | +7.0% |
| 20 to 30 N | +10.1% | +4.1% | +14.4% | +7.0% |
| 30 to 40 N | +6.4% | +3.4% | +10.0% | +7.0% |
| 40 to 50 N | +8.3% | +6.1% | +10.9% | +7.0% |
| 50 to 60 N | +7.3% | +4.3% | +10.6% | +7.0% |
| 60 to 70 N | +10.7% | +7.7% | +14.7% | +7.0% |
Of the 22 regions in the tables below, the wettest day's rate for 1950 to 2025 lies above 7 percent per degree in 0, below it in 4 (Southern hemisphere, Oceania, 30 to 40 S and Australia) and cannot be told apart from it in 18, after allowing for the number of regions compared.
| Hemisphere | Stations | Grid cells | Wettest day, % per degree C | 95% interval | Wettest day against the two lines | Wettest five days, % per degree C | 95% interval |
|---|---|---|---|---|---|---|---|
| Northern hemisphere | 7,245 | 382 | +8.3% | +6.1 to +9.7 | above zero, not clear of 7 percent | +6.7% | +5.2 to +7.9 |
| Southern hemisphere | 2,515 | 48 | -1.0% | -4.8 to +3.1 | below 7 percent, not clear of zero | -2.4% | -7.6 to +7.2 |
| Continent | Stations | Grid cells | Wettest day, % per degree C | 95% interval | Wettest day against the two lines | Wettest five days, % per degree C | 95% interval |
|---|---|---|---|---|---|---|---|
| Asia | 319 | 67 | +10.1% | +5.1 to +13.7 | above zero, not clear of 7 percent | +6.9% | +4.7 to +12.7 |
| Europe | 2,000 | 196 | +7.3% | +4.1 to +8.8 | above zero, not clear of 7 percent | +8.6% | +6.3 to +9.6 |
| North America | 4,917 | 121 | +9.1% | +6.4 to +10.6 | above zero, not clear of 7 percent | +5.1% | +2.9 to +7.6 |
| Oceania | 2,514 | 48 | -1.1% | -5.0 to +1.9 | below 7 percent, not clear of zero | -2.5% | -6.2 to +7.2 |
| Latitude band | Stations | Grid cells | Wettest day, % per degree C | 95% interval | Wettest day against the two lines | Wettest five days, % per degree C | 95% interval |
|---|---|---|---|---|---|---|---|
| 60 to 70 N | 522 | 87 | +10.7% | +7.7 to +14.7 | above zero, not clear of 7 percent | +10.3% | +7.5 to +12.7 |
| 50 to 60 N | 1,201 | 91 | +7.3% | +4.3 to +10.6 | above zero, not clear of 7 percent | +6.7% | +4.6 to +10.6 |
| 40 to 50 N | 2,453 | 76 | +8.3% | +6.1 to +10.9 | above zero, not clear of 7 percent | +6.9% | +3.2 to +10.6 |
| 30 to 40 N | 2,432 | 56 | +6.4% | +3.4 to +10.0 | above zero, not clear of 7 percent | +5.2% | +2.9 to +7.4 |
| 20 to 30 N | 449 | 31 | +10.1% | +4.1 to +14.4 | above zero, not clear of 7 percent | +4.7% | +0.9 to +11.6 |
| 10 to 20 S | 135 | 11 | +3.8% | +1.4 to +11.6 | above zero, not clear of 7 percent | +10.3% | +1.2 to +16.3 |
| 20 to 30 S | 769 | 20 | -2.5% | -8.1 to +3.5 | not clear of zero or of 7 percent | -5.5% | -9.3 to +7.7 |
| 30 to 40 S | 1,497 | 14 | -1.2% | -5.2 to +4.1 | below 7 percent, not clear of zero | -2.5% | -11.4 to +0.9 |
| Country | Stations | Grid cells | Wettest day, % per degree C | 95% interval | Wettest day against the two lines | Wettest five days, % per degree C | 95% interval |
|---|---|---|---|---|---|---|---|
| United States | 4,306 | 67 | +6.5% | +5.1 to +10.0 | above zero, not clear of 7 percent | +4.6% | +2.8 to +6.7 |
| Australia | 2,507 | 41 | -0.6% | -4.8 to +3.1 | below 7 percent, not clear of zero | -0.3% | -7.6 to +7.7 |
| Russia | 531 | 141 | +7.9% | +5.6 to +9.5 | above zero, not clear of 7 percent | +9.2% | +6.4 to +10.3 |
| Mexico | 345 | 19 | +16.1% | +9.1 to +34.0 | above zero, not clear of 7 percent | +9.5% | +4.7 to +30.8 |
| Canada | 235 | 47 | +8.5% | +4.5 to +11.5 | above zero, not clear of 7 percent | +6.3% | +0.9 to +14.2 |
| China | 182 | 48 | +12.2% | +4.0 to +14.0 | above zero, not clear of 7 percent | +6.4% | +2.9 to +12.7 |
| Norway | 181 | 15 | +14.0% | +2.3 to +17.8 | above zero, not clear of 7 percent | +15.8% | +11.7 to +18.6 |
| Japan | 124 | 11 | +13.1% | +3.8 to +14.5 | above zero, not clear of 7 percent | +9.6% | +4.2 to +18.5 |
A region is listed with at least 10 grid cells and 30 stations. "Above" or "below" a line means the region passed a control of the false discovery rate at 5 percent over all regions and its interval excludes the line; otherwise the region is "not clear" of it, even where its own interval misses the line. With fewer than 30 cells the interval is somewhat too narrow: in a simulation with 10 cells it held the true rate in about 92 of 100 cases, not 95.
The Wettest Day Against the Yearly Total
1950 to 2025The Clausius-Clapeyron rate is about the heaviest rain. The rain of a whole year, taken over the globe, is limited by how fast water evaporates, which grows more slowly with temperature. The table sets the wettest day and the wettest five days beside the yearly total and the number of wet days (days with 1 mm or more), each measured the same way at the same stations.
| Measure | Stations | Grid cells | Percent per degree C | 95% interval | Stations above zero | Clear increase | Clear decrease |
|---|---|---|---|---|---|---|---|
| Wettest day of the year | 9,760 | 430 | +6.5% | +5.1 to +8.3 | 63.3% | 9.5% | 2.6% |
| Wettest five days of the year | 9,760 | 430 | +6.3% | +4.6 to +7.4 | 59.4% | 10.0% | 4.1% |
| Yearly total | 9,760 | 430 | +7.9% | +5.9 to +9.9 | 63.0% | 23.3% | 7.0% |
| Number of wet days | 9,760 | 430 | +1.6% | +0.6 to +2.8 | 52.4% | 18.7% | 15.8% |
At the 4,810 stations within 40 degrees of the equator and below 1,000 m (3,281 ft), where little of the year's precipitation falls as snow, the yearly total rose by 4.1 percent per degree C of global warming from 1950 to 2025 (95 percent interval +1.4 to +6.2 percent). At the other 4,950 stations, further from the equator or higher up, it rose by 9.6 percent per degree C of global warming (+7.0 to +10.7 percent). The two intervals do not overlap: the yearly total is clearly higher where snow falls. Some of that can be real, since a warmer atmosphere is expected to bring more precipitation to high latitudes. But a gauge catches snow far less well than rain, so stations there also report more when winters bring less snow and more rain, or when gauges change, whether or not more falls. These data cannot separate the two, so the yearly total at those stations is not reliable evidence of more precipitation, and this section reads the yearly total from the low-snow stations.
| Measure | Low-snow stations | % per degree C | 95% interval | Other stations | % per degree C | 95% interval |
|---|---|---|---|---|---|---|
| Wettest day of the year | 4,810 | +5.3% | +3.1 to +8.5 | 4,950 | +7.7% | +5.6 to +8.9 |
| Wettest five days of the year | 4,810 | +4.3% | +1.2 to +7.4 | 4,950 | +6.2% | +4.5 to +7.6 |
| Yearly total | 4,810 | +4.1% | +1.4 to +6.2 | 4,950 | +9.6% | +7.0 to +10.7 |
| Wettest day minus yearly total, station by station | 4,810 | +3.9% | +1.0 to +5.7 | 4,950 | -0.9% | -2.6 to +0.3 |
| Wettest five days minus yearly total, station by station | 4,810 | +1.1% | -0.8 to +2.5 | 4,950 | -2.1% | -3.0 to -1.3 |
Station by station at the low-snow stations, the wettest day of the year grew 3.9 percent per degree faster than the yearly total from 1950 to 2025 (95 percent interval +1.0 to +5.7 percent, 4,810 stations in 137 grid cells). The interval lies above zero, so the difference is clear.
Over all stations the same difference is -0.1 percent per degree (95 percent interval -1.2 to +1.3 percent, 9,760 stations in 430 grid cells), but that figure rests on the yearly totals of stations where snow falls and says little about rain.
Station by station at the low-snow stations, the wettest five days of the year grew 1.1 percent per degree faster than the yearly total from 1950 to 2025 (95 percent interval -0.8 to +2.5 percent, 4,810 stations in 137 grid cells). The interval holds zero, so the two rates cannot be told apart.
Over all stations the same difference is -1.0 percent per degree (95 percent interval -2.1 to +0.2 percent, 9,760 stations in 430 grid cells), but that figure rests on the yearly totals of stations where snow falls and says little about rain.
At 63.3 percent of the 9,760 stations the station's own rate for the wettest day of the year, 1950 to 2025, is above zero; 50 percent is expected with no change. 9.5 percent of the stations show an increase that is individually clear at the 5 percent level and 2.6 percent a decrease, where chance alone gives about 2.5 percent each. After controlling the false discovery rate at 5 percent over all stations, 43 increases remain and 11 decreases. One station alone says little: the median standard error of a single station's rate is 13.7 percent points per degree.
Does the Result Hold?
1950 to 2025| Check | Stations | Grid cells | Wettest day, % per degree C | 95% interval | Wettest five days, % per degree C | 95% interval | Yearly total, % per degree C | 95% interval |
|---|---|---|---|---|---|---|---|---|
| As published | 9,760 | 430 | +6.5% | +5.1 to +8.3 | +6.3% | +4.6 to +7.4 | +7.9% | +5.9 to +9.9 |
| From 1960 instead of 1950 | 8,909 | 425 | +7.8% | +5.3 to +9.2 | +5.6% | +3.2 to +7.2 | n/a | n/a |
| Only stations with 70 or more years | 4,212 | 280 | +5.3% | +3.5 to +7.0 | +4.6% | +2.8 to +6.7 | +7.3% | +4.4 to +9.8 |
| Without the United States | 5,454 | 381 | +6.6% | +4.3 to +8.5 | +6.9% | +4.8 to +8.6 | +8.0% | +6.2 to +10.0 |
| Only years with 350 or more days of readings | 7,503 | 408 | +5.9% | +4.3 to +7.4 | +6.7% | +5.9 to +7.9 | n/a | n/a |
| Every value rounded to a whole millimetre | 9,760 | 430 | +6.5% | +5.2 to +8.5 | +6.3% | +4.6 to +7.5 | n/a | n/a |
| Weighted mean of the cells instead of their median | 9,760 | 430 | +6.4% | +3.9 to +8.7 | +6.0% | +3.3 to +8.4 | +7.6% | +4.9 to +10.3 |
| Only stations within 40 degrees of the equator and below 1,000 m (3,281 ft) | 4,810 | 137 | +5.3% | +3.1 to +8.5 | +4.3% | +1.2 to +7.4 | +4.1% | +1.4 to +6.2 |
| One regression of the pooled yearly series | 9,760 | 430 | +5.9% | +4.1 to +7.7 | +5.2% | +3.3 to +7.2 | +7.0% | +4.2 to +9.9 |
For 1950 to 2025, none of these checks changes the sign of the result or moves it to half or twice the published rate. Over the published value and all checks the wettest day's rate runs from +5.3 to +7.8 percent per degree C. The yearly total is shown for the checks that need no new station fits.
Rain in the United States is reported in hundredths of an inch and elsewhere in tenths of a millimetre or coarser steps. At the median station the wettest day of a typical year brings 49.5 mm (1.95 in) of rain or melted snow, so half a millimetre, the most that rounding to a whole millimetre can shift a value, is 1.0 percent of it. 13.1 percent of the yearly wettest days of 1950 to 1979 are whole millimetres and 20.3 percent of those of 1996 to 2025; reporting to a tenth of a millimetre or a hundredth of an inch gives about 10 percent by chance. The share changed, so part of the network changed its reporting step; the check with every value rounded to a whole millimetre shows how much that can matter.
How the Rate Is Measured
The data are daily precipitation totals (rain and melted snow) of GHCN-Daily stations. A year of a station counts from 1950 to 2025 when it has at least 300 days with a reading and its wettest day reached 1 mm. A station is used with at least 50 such years, at least 10 of them in the first 25 and in the last 25 years of the period, so that no station covers only one end of the warming. A missing value stays missing and is never read as zero rain. 102 station years (0.02 percent) were skipped because the wettest day stayed under 1 mm; 653,415 station years are used.
The global temperature of a year is NOAA's global mean of land and ocean, averaged over that year and the 3 years before it, so that single warm years such as El Nino years do not drive the result. For each station the natural logarithm of the yearly value is fitted against that temperature by least squares. A slope b in logarithms means the value is multiplied by exp(b) per degree C, which is 100 (exp(b) - 1) percent per degree; the page reports that percent.
The wettest five days are the largest total of five consecutive days with all five reported, and never less than the wettest day. The yearly total and the number of wet days are divided by the number of days with a reading, so that a year with fewer readings does not look drier.
One station is noisy, so the rates are pooled. The stations of a 5 degree grid cell give the cell's median rate. The cells are then pooled by their median, weighted by the cosine of the latitude, which is proportional to the cell's area. The 95 percent interval comes from drawing the cells again at random with replacement 2,000 times and pooling each draw; the random draws use a fixed starting number, so the same data give the same interval. The interval treats the cells as independent.
For the hemispheres, continents, latitude bands and countries many comparisons are made at once, so some intervals would miss a line by chance. The Benjamini-Hochberg procedure keeps the expected share of false calls among the regions called "above" or "below" at 5 percent. The same procedure is applied to the stations one by one.
What the Numbers Cover and Where They Stop
Of the 9,760 stations, 4,306 (44.1 percent) are in the United States, 2,507 (25.7 percent) in Australia and 531 (5.4 percent) in Russia; the stations are spread over 52 countries. Pooling by grid cell and area keeps those countries from deciding the result, but where there are no long records nothing is measured, and a cell with one or two stations weighs as much as a cell with many. Africa and South America have too few grid cells or stations with a record of 50 years for a row of their own.
The yardstick is the global mean temperature. Land warms faster than the global mean and each region at its own pace, so a rate per degree of global warming is higher than the same change counted per degree of local warming, and the 7 percent of the Clausius-Clapeyron relation applies to the air where the rain forms. Observed rates can also differ from 7 percent because storms change: stronger updraughts can lift the rate above it, a shift of storm tracks can move it either way, and over land far from the sea the supply of moisture can lag behind the warming.
Over 1950 to 2025 the global temperature yardstick rose with the calendar year (the correlation of the two is 0.96), so this analysis cannot separate warming from anything else that changed gradually over the same decades, at the stations or in the atmosphere. The page says how the wettest day moved with the global temperature. It does not show that the warming caused the change.
A day is the station's observation day, read at a fixed hour, not the wettest 24 hours, and a year is a calendar year, which cuts the wet season in two where it falls around New Year. Rain gauges catch less than falls, most of all in wind and snow, and a new gauge, a new site or a new reading hour can shift a station's record; the records are used as measured and are not adjusted. A change of that kind at many stations of a country at once would appear here as a change of rain.
Snow needs a caution of its own. A rain gauge catches a smaller part of falling snow than of rain, least of all in wind. Where winters bring less snow and more rain than they did, or where a network changed its gauges or the way it corrects their readings, stations report more precipitation without more falling. A warmer atmosphere is also expected to bring more precipitation to high latitudes, and these data cannot tell the two apart. The yearly total and the number of wet days are the most exposed to this; the wettest day is the least exposed where the heaviest falls of the year are rain. The page therefore gives the yearly total separately for the 4,810 stations within 40 degrees of the equator and below 1,000 m (3,281 ft) and for the other 4,950. The rule is rough: snow falls at some stations of the first group and hardly ever at some of the second. 1 station has no known elevation and is counted with the second group.
The rates describe 1950 to 2025 and the stations used. They are not a forecast, and a single station's rate, with a standard error near 14 percent points per degree, says little about that place. The yardstick's level of 1951 to 1980 is the starting point for the warming quoted above.
Download the Data
Source: NOAA National Centers for Environmental Information, Global Historical Climatology Network Daily (GHCN-Daily). Records are computed by World Extreme Weather Monitor and are not official NOAA products. (US Government open data (NOAA NCEI)) Source: NOAA National Centers for Environmental Information, NOAA Global Surface Temperature Dataset (NOAAGlobalTemp), Version 6.1, annual land and ocean time series (Huang, B., X. Yin, M.J. Menne and R.S. Vose, 2026, doi:10.25921/vvaa-wq11). Used as the measure of global warming; the fits to station data are computed by World Extreme Weather Monitor and are not NOAA products. (US Government open data (NOAA NCEI Open Data Policy, citation requested)) Data as of 2026-11-17. Station regressions, grid cell medians, pooled rates, bootstrap intervals and the yearly series computed by this site from GHCN-Daily precipitation and the NOAAGlobalTemp global mean; not an official NOAA product.