Skip to Main Content

Lake Ice

C. Duguay, L. Brown, K.-K. Kang, H. Kheyrollah Pour

Interdisciplinary Centre on Climate Change and
Department of Geography & Environmental Management
University of Waterloo, Canada

November 14, 2011


  • Ice cover duration in 2010-2011 was comparable to that of 2009-2010 for North America and much of eastern Siberia, but longer by 2-3 weeks for most parts of northern Europe.
  • Ice cover duration was shorter by as much as 4-5 weeks in 2010-2011 compared to the 1997-2010 average for the eastern Canadian Arctic.

Lake ice is a sensitive indicator of climate variability and change. Lake ice phenology, which encompasses freeze-up (ice-on) and break-up (ice-off) dates, and ice cover duration, is largely influenced by air temperature changes and is, therefore, a robust indicator of regional climate conditions (e.g. Duguay et al., 2006; Kouraev et al., 2007; Latifovic and Pouliot, 2007).

The analysis of ground-based observational records has provided evidence of later freeze-up (ice-on) and earlier break-up (ice-off) dates over the Northern Hemisphere, particularly during the second half of the 20th century (e.g., Brown and Duguay, 2010). In the last 20 years, however, ground-based lake ice networks have been eroded to the point where they can no longer provide the quality of observations necessary for climate monitoring. Satellite remote sensing is the most logical means for establishing an Arctic-wide lake ice observational network.

In this first report on the state of Arctic lake ice cover, ice phenology dates (freeze-up/ice-on and break-up/ice-off dates) and ice cover duration derived from two satellite-based products are documented for the 2010-2011 ice season and compared to average conditions for the length of the available satellite historical records. The NOAA Interactive Multisensor Snow and Ice Mapping System (IMS) 4 km resolution grid daily product was used to analyze Arctic-wide ice conditions from the largest lakes since its availability in 2004 (Fig. HTC26). The IMS incorporates a wide variety of satellite imagery (AVHRR, GOES, SSMI, etc.) as well as derived mapped products (USAF Snow/Ice Analysis, AMSU, etc.) and surface observations (see Helfrich et al., 2007 for details). Ice-on and ice-off dates as well as ice duration could be derived at the pixel level from this product. Freeze-up (break-up) dates and ice-on (ice-off) dates have the same meaning in this report.

Fig. 26 -- Location of lakes and CIS data sets

Fig. HTC26. Location of lakes from IMS 4-km (in blue) and CIS data sets (numbered in red, 1-26). IMS (Interactive Multisensor Snow and Ice Mapping System) and CIS (Canadian Ice Service) data go back as far as 2004 and 1997, respectively.

Weekly ice-cover observations from the Canadian Ice Service (CIS) for 26 of the largest lakes across northern Canada and Alaska were also analyzed, starting with the first reported ice season (1997-1998) (Fig. HTC26). Ice analysts at the CIS determine a single lake-wide ice fraction value in tenths (ranging from 0 [open water] to 10 [complete ice cover]) once per week week from the visual interpretation of NOAA AVHRR and Radarsat ScanSAR images. Complete freeze over (CFO) and water clear of ice (WCI) dates can be derived from this product with about a one-week accuracy. CFO is determined as the date when the ice fraction changes from 9 to 10 and remains at this value for the entire winter period, while WCI is determined as the date when the lake-ice fraction passes from 1 to 0. Lake-wide ice cover duration corresponds to the number of days between CFO and WCI within an ice season.

Freeze-up period

Analysis of IMS data shows that freeze-up in ice season 2010-2011 was close to the 2004-2010 mean, with a few lakes showing either later or earlier dates (±10-20 days) in much of North America and eastern Siberia (Fig. HTC27). There are no regions showing a clear temporal coherence, except for northern Quebec (10-20 days later), Baffin Island (10-20 days earlier) and northern/eastern Europe (20-40 days earlier). Similar differences are observed when ice seasons 2010-2011 and 2009-2010 are compared. There are relatively no differences in freeze-up dates for most regions of the Arctic, with the exception of the Hudson Bay region and Baffin Island in North America (up to 4 weeks later) and northern/eastern Europe (up to 4 weeks earlier). The penetration of cold air from the Arctic into Europe (see the essay on Temperature and Clouds) is the most plausible explanation for the earlier freeze-up of the 2010-2011 ice season.

Fig. 27 -- Difference in number of days between freeze-up dates and mean

Fig. HTC27. Difference in number of days between freeze-up dates of 2010 and mean (2004-2009). Source IMS 4-km data.

The longer time series of weekly CIS observations over Canada/Alaska reveals that, in contrast to the 1997-2010/1998-2010 mean conditions, CFO of ice season 2010-2011 was later by 1 week on average (± 2 weeks; 18 lakes) but with 3 lakes in the central sector of the Canadian Arctic (Great Bear Lake, Great Slave Lake and Hjalmar Lake) displaying earlier CFO by as much as 3 weeks (Fig. HTC28). A combination of factors, including lake size, break-up dates of the previous spring, and colder open water season temperatures (reduced heat storage) may explain the earlier CFO in 2010-2011.

Fig. 28 -- Dates of complete freeze over and water clear of ice

Fig. HTC28. Dates of complete freeze over (CFO) and water clear of ice (WCI) derived from CIS database for mean periods (1997-2010, 1998-2010, 2003-2010) compared to ice season 2010-2011 (light blue). The length of the horizontal bars indicates ice cover duration (ICD).

Break-up period

Break-up dates of ice season 2010-2011 derived from IMS are on average close to those of the 2004-2010 mean period but with some regional differences (Fig. HTC29), being especially highly variable (low coherence) in central Canada (±10-20 days). A large number of lakes in northern Europe (northern portion) and eastern Siberia experienced earlier break-up (10-20 days), while for much of eastern Europe and the southern portion of northern Europe break-up dates were 10-30 days later. The colder climate conditions of winter 2010-2011 in this sector, which led to earlier freeze-up, likely promoted thicker ice conditions (not verified) and, as a consequence, in combination with colder early spring temperatures, delayed ice break-up (see the essay on Temperature and Clouds). Break-up dates are generally comparable between ice seasons 2010-2011 and 2009-2010 for all regions, excluding eastern Europe and eastern Siberia, where they occurred 2-4 weeks later and earlier, respectively. The impact of earlier/later break-up dates on July lake surface temperatures (LSTs) derived from MODIS using the method described in Kheyrollah Pour et al. (in review) is illustrated in Fig. HTC30. For eastern Siberia, mean monthly LSTs in July 2011 were 3-4°C warmer than in 2010 (Fig. HTC30a), while they were as cold by the same amount in eastern Europe (Fig. HTC30b). The earlier break-up of 2011 in eastern Siberia is consistent with the record low June snow cover extent over Eurasia (see the essay on Snow).

Fig. 29 -- Difference in number of days between break-up dates and mean

Fig. HTC29. Difference in number of days between break-up dates of 2011 and mean (2004-2010). Source: IMS 4-km data.

Fig. 30a -- Monthly lake surface temperature eastern Siberia
Fig. 30b -- Monthly lake surface temperature northern and eastern Europe

Fig. HTC30. Monthly lake surface temperature derived from combined MODIS Aqua/Terra data (Kheyrollah Pour et al., in review) for July 2010 and 2011 centered on (a) eastern Siberia and (b) northern/eastern Europe.

The CIS time series reveals earlier WCI in 2011 compared to 1997-2010/1998-2010 mean conditions for all 22 lakes analyzed over Canada/Alaska (Fig. HTC28). WCI occurred 12 days earlier (range 1-24 days) on average. For lakes with a shorter historical series (starting in 2003), WCI was still earlier in 2011 by the same number of days on average compared to the 2003-2010 mean. These results suggest a continued trend in earlier break-up dates across the Canadian Arctic first documented by Duguay et al. (2006) with ground-based observations (1966-1995) and extended from the analysis of NOAA AVHRR satellite observations (1985-2004) by Latifovic and Pouliot (2007).

Ice cover duration

When comparing the 2010-2011 ice season against the mean for 2004-2010 from IMS, the net effect of earlier/later freeze-up and earlier/later break-up results in ice cover duration (ICD) being generally shorter by about 2 weeks in central and eastern Arctic Canada, with the exception of lakes on Baffin Island and western Canada/Alaska (longer by 2-3 weeks). ICD was also shorter in eastern Siberia by 2-3 weeks and longer in northern/eastern Europe by as much as 4-6 weeks (Fig. HTC 31). The combination of earlier freeze-up and later break-up due to colder winter/early spring climate conditions explains the longer ICD over Europe. The last two ice seasons (2010-2011 and 2009-2010), however, experienced comparable ICD length for Arctic North America and most of eastern Siberia, and somewhat longer in 2010-2011 for most parts of Europe (2-3 weeks).

Fig. 31 -- Difference in number of days between ice cover duration and the mean

Fig. HTC31. Difference in number of days between ice cover duration of 2010-2011 and the mean for 2004-2010. Source IMS 4-km data.

Comparing the CIS mean (1997-2010/1998-2010) against 2010-2011 reveals a shorter ICD by 2-3 weeks for the majority of lakes with some of the strongest changes (about 30 fewer days) observed on Baffin Island and western Hudson Bay (Fig. HTC28). Only 2 lakes of the full 14-year series, Great Bear Lake and Great Slave Lake in the central region, show a slightly longer ICD (1 week).


Brown L. C. and C.R. Duguay. 2010. The response and role of ice cover in lake-climate interactions. Progress in Physical Geography, 34(5): 671-704.

Duguay, C.R., T.D. Prowse, B.R. Bonsal, R.D. Brown, M.P. Lacroix, and P. Ménard. 2006. Recent trends in Canadian lake ice cover. Hydrological Processes. 20: 781-801.

Helfrich, S. R., D. McNamara, B.H. Ramsay, T. Baldwin, and T. Kasheta. 2007. Enhancements to, and forthcoming developments in the Interactive Multisensor Snow and Ice Mapping System (IMS). Hydrological Processes. 21: 1576-1586.

Kheyrollah Pour, H., C.R. Duguay, A. Martynov, and L.C. Brown. In review. Simulation of surface temperature and ice cover of large northern lakes with 1-D models: A comparison with MODIS satellite data and in situ measurements. Tellus Series A: Dynamic Meteorology and Oceanography.

Kouraev, A. V., S.V. Semovski, M.N. Shimaraev, N.M. Mognard, B. Légresy, and F. Remy. 2007. Observations of Lake Baikal ice from satellite altimetry and radiometry. Remote Sensing of Environment. 108: 240-253.

Latifovic, R. and D. Pouliot. 2007. Analysis of climate change impacts on lake ice phenology in Canada using the historical satellite data record. Remote Sensing of Environment. 106: 492-508, 2007.