Minnesota keeps more lake ice records than any other state, and it is the only one that tracks the freeze as carefully as the thaw. The state climatology office has been collecting ice-in and ice-out reports since the middle of the nineteenth century.
Dates from past winters. For today’s ice, ask the agency that manages the water.
The DNR maps ice-out reports for hundreds of Minnesota lakes as they come in each spring, and keeps a separate ice-in map for the freeze. Their medians are computed from 1950 onward for lake-to-lake consistency, so they will not always match the whole-record figures on this page.
40 of 40 of the most recent recorded years had already seen ice out on Detroit by this date.
The median across the whole record is April 20, which is 172 days ago on this year’s calendar. The record itself runs to 2020, so this is a comparison against history and not a statement about the current season.
This counts recorded years, not this year; the current season comes from the agency linked above.
Ordered by how much record there is behind them. The median is across every recorded year; the second figure is across the most recent 30 recorded years, which is the fairer thing to compare against a lake you know today.
Names are reproduced exactly as the source database records them, which is often a shortened form: a lake filed as “Detroit” or “George” is Detroit Lake or Lake George to everyone who lives on it. The names stay as recorded, because United States lakes split unpredictably between “Lake X” and “X Lake”.
Lake in Minnesota · 128 recorded years, 1893 to 2020
Across the whole record, ice out has moved barely at all across the record.
Freeze over has moved 11 days later per century across the record, median Nov 22.
Mean date per decade. Only decades with four or more observations are shown, and the scale starts near the shortest bar rather than at January 1, so small differences stay visible.
Lake in Minnesota · 86 recorded years, 1897 to 2016
Across the whole record, ice out has moved 9 days earlier per century across the record.
Freeze over has moved 3 days later per century across the record, median Dec 2.
Mean date per decade. Only decades with four or more observations are shown, and the scale starts near the shortest bar rather than at January 1, so small differences stay visible.
Lake in Minnesota · 150 recorded years, 1867 to 2020
Across the whole record, ice out has moved 6 days earlier per century across the record.
Mean date per decade. Only decades with four or more observations are shown, and the scale starts near the shortest bar rather than at January 1, so small differences stay visible.
Lake in Minnesota · 144 recorded years, 1855 to 2020
Across the whole record, ice out has moved 4 days earlier per century across the record.
Mean date per decade. Only decades with four or more observations are shown, and the scale starts near the shortest bar rather than at January 1, so small differences stay visible.
| Lake | Record | Median ice out | Last 30 years | Earliest | Latest | Change per century |
|---|---|---|---|---|---|---|
| Detroit | 1893 to 2020 (128) | Apr 20 | Apr 18 | Mar 23 1910 | May 17 1950 | -0.9 days |
| Green | 1897 to 2016 (86) | Apr 16 | Apr 12 | Mar 21 2012 | May 8 2013 | -8.5 days |
| Osakis | 1867 to 2020 (150) | Apr 20 | Apr 14 | Mar 13 1878 | May 14 1950 | -5.5 days |
| Minnetonka | 1855 to 2020 (144) | Apr 14 | Apr 10 | Mar 11 1878 | May 9 1856 | -3.8 days |
| Waconia | 1940 to 2017 (69) | Apr 12 | Apr 8 | Mar 7 2017 | May 1 1941 | record too short |
| Pepin | 1844 to 2006 (73) | Apr 9 | Apr 4 | Feb 1 2000 | May 1 1857 | -11.3 days |
| Minnewaska | 1906 to 2017 (112) | Apr 14 | Apr 10 | Mar 22 2012 | May 11 2013 | -0.6 days |
| Clear | 1894 to 2008 (109) | Apr 3 | Apr 1 | Mar 6 1987 | Apr 27 1951 | +0.6 days |
| Fountain | 1912 to 2017 (97) | Apr 6 | Apr 5 | Mar 15 1931 | Apr 30 1953 | +1.6 days |
| White Bear | 1928 to 2017 (90) | Apr 12 | Apr 8 | Mar 15 2002 | May 4 1950 | -8.7 days |
| Vermilion | 1893 to 2006 (89) | Apr 30 | Apr 29 | Apr 10 1945 | May 23 1950 | -2.3 days |
| Rainy | 1930 to 2016 (87) | May 5 | May 2 | Apr 7 2012 | May 22 1950 | -7.8 days |
| Itasca | 1929 to 2017 (81) | Apr 23 | Apr 20 | Mar 27 2012 | May 18 1950 | -6.7 days |
| Bemidji | 1932 to 2017 (80) | Apr 26 | Apr 25 | Apr 3 2012 | May 22 1950 | -6.6 days |
| Diamond(mail) | 1928 to 2017 (75) | Apr 12 | Apr 13 | Mar 8 2017 | May 7 2013 | -0.9 days |
| Leech | 1936 to 2005 (70) | Apr 27 | Apr 25 | Apr 9 1945 | May 23 1950 | record too short |
| Fall | 1935 to 2017 (67) | Apr 28 | Apr 25 | Mar 28 2012 | May 19 1950 | -11.2 days |
| Galpin | 1946 to 2014 (66) | Apr 7 | Apr 5 | Mar 9 2000 | Apr 27 1965 | record too short |
| Kabetogama | 1952 to 2017 (66) | Apr 30 | Apr 29 | Apr 2 2012 | May 19 1996 | record too short |
| Big Stone | 1951 to 2017 (65) | Apr 10 | Apr 3 | Mar 17 2016 | May 8 2013 | record too short |
| Bone | 1942 to 2014 (65) | Apr 12 | Apr 9 | Mar 16 2000 | Apr 30 1968 | record too short |
| Big Sandy | 1930 to 2017 (63) | Apr 21 | Apr 19 | Mar 27 2012 | May 12 2013 | -7.6 days |
| Sisseton | 1935 to 2017 (61) | Apr 3 | Apr 2 | Mar 13 1990 | Apr 26 1951 | -4.7 days |
| Gull | 1950 to 2017 (55) | Apr 21 | Apr 19 | Mar 27 2012 | May 18 1950 | record too short |
| Mille Lacs | 1956 to 2017 (55) | Apr 24 | Apr 24 | Mar 27 2012 | May 16 2013 | record too short |
| Shagawa | 1965 to 2017 (51) | Apr 27 | Apr 26 | Mar 28 2012 | May 17 1996 | record too short |
| Buffalo | 1953 to 2017 (50) | Apr 11 | Apr 9 | Mar 7 2017 | May 2 2013 | record too short |
| Shetek | 1960 to 2016 (49) | Apr 3 | Apr 2 | Mar 9 2000 | Apr 24 1962 | record too short |
| Lizzie | 1950 to 1995 (46) | Apr 21 | Apr 19 | Apr 4 1981 | May 14 1950 | record too short |
| Winsted | 1887 to 2016 (44) | Apr 12 | Apr 10 | Mar 9 1987 | Apr 29 2013 | -6.4 days |
| Gunflint | 1940 to 2016 (41) | May 6 | May 5 | Mar 24 2012 | May 26 1966 | record too short |
Dates are day of year, rendered on a common non leap calendar so the same stored value always reads the same way. “Change per century” is a least squares line through the observations, shown only where a record spans 80 years or more; a century figure extrapolated from a 40 year window describes the decades that happened to be sampled rather than the lake. Where a record ends before 2010 the card says so; a great many of these series stopped when a particular person stopped keeping them.
Of the 18 Minnesota lakes with a record long enough to state a century trend, 16 show ice out moving earlier. The median shift across those lakes is -5.5 days per century. That is a description of a line fitted to observations, not a forecast, and any single lake can and does swing three or four weeks between consecutive years.
4 of these Minnesota water bodies also carry a freeze over record. Freeze up is the harder half to observe and there is far less of it nationally, which is a shame, because in this dataset it moves more than ice out does.
Ice-out is a statewide countdown. The DNR asks the public to report ice-out on their own lake, which is why Minnesota has more reporting lakes than anywhere else in the country. In a warm spring the reports come in from the south and sweep north over about three weeks.
Benson, B., Magnuson, J. and Sharma, S. (2000). Global Lake and River Ice Phenology Database. (G01377, Version 1). Boulder, Colorado USA. National Snow and Ice Data Center. https://doi.org/10.7265/N5W66HP8 Accessed 2026-08-24.
Osakis has been reporting ice-out since 1867 and Minnetonka since 1855, which puts both among the oldest continuously observed lakes in the United States.
Ice out and ice in dates are historical observations compiled by the National Snow and Ice Data Center and, for the New England lakes, by the U.S. Geological Survey. They describe past years only, and a past year says nothing about the ice on a lake today. For current ice and for anything involving going out on it, contact Minnesota DNR ice safety. National picture: every state with a lake ice record →
Hero image: “Shagawa Lake - Frozen in Winter - Ely, Minnesota (44605661414)” by Tony Webster from Minneapolis, Minnesota, United States, CC BY 2.0