More than 90% of the world's lakes are located in the Northern Hemisphere (north of 30°N), and many experience seasonal ice cover. Lake ice acts as an insulator, shielding the water from atmospheric forcing and affecting albedo. Ice phenology (timing of ice formation and loss) regulates seasonal surface energy and thermal regimes. However, climate change significantly impacts lake ice, leading to later ice formation and earlier break-up, thus extending the ice-free season. This change in ice phenology affects the seasonal and inter-annual variability of lake surface water temperature, with cascading effects on lake physical and biogeochemical processes. Historical trends in ice phenology influence lake warming rates during ice-off and ice-on months. Earlier ice break-up increases incoming shortwave radiation due to reduced albedo, resulting in increased net surface heating and warmer water temperatures. This can lead to earlier phytoplankton growth, altered spring blooms, earlier thermal stratification, and warmer surface water temperatures during the open-water season. Conversely, later ice-on dates can lead to increased radiation input but also greater heat loss through evaporation. The magnitude of lake surface temperature change during ice-off and ice-on, compared to summer warming, remains uncertain. While some studies have indicated excess warming during these transitional periods, a comprehensive analysis across larger geographic regions is lacking. This study addresses this knowledge gap by investigating excess lake warming during ice-off and ice-on months across Northern Hemisphere lakes using satellite observations and modeled lake surface temperatures.

Previous research has shown the vulnerability of lake ice to climate change, with many lakes experiencing substantial shifts in ice phenology. Studies using in situ observations and model simulations have suggested that earlier ice-off and/or later ice-on can lead to excess lake surface warming, but this phenomenon has not been explored across large geographic regions. The impacts of changing ice phenology on lake temperatures are multifaceted. Earlier ice break-up can lead to a cascade of ecological effects, such as earlier phytoplankton blooms and changes in species composition. It can also alter thermal stratification patterns and affect the timing of other physical and biogeochemical processes within the lake. The relative contribution of air temperature, solar radiation, and wind speed to lake warming rates has been studied extensively, but the magnitude of changes during ice-off and ice-on remains uncertain.

This study employed two datasets to analyze monthly trends in lake surface water temperature (LSWT): (i) satellite observations from 963 globally distributed lakes (1995–2012) and (ii) modeled LSWT from 109,405 representative lakes (1979–2020) derived from the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 reanalysis product. The ERA5 data, simulated using the FLake model, provided hourly lake surface temperature, along with other climate variables (surface air temperature, downward shortwave and longwave radiation). Satellite data from the Along Track Scanning Radiometer (ATSR) Reprocessing for Climate: Lake Surface Water Temperature and Ice Cover (ARC-Lake) dataset were used for validation. Lake ice phenology (ice-on and ice-off dates) was determined using daily LSWT from ERA5 and ARC-Lake. Ice-cover periods were identified when LSWT was <1°C, with ice-on defined as the first day with 10 consecutive days of ice cover and ice-off as the first day with 10 consecutive ice-free days. The analysis included calculating monthly LSWT trends, ratios of trends in ice-off/ice-on months to open-water season trends (R_IOFF, R_ION), and sensitivities of LSWT to changes in ice phenology (S_IOFF, S_ION). To explore potential drivers of excess warming, correlations between LSWT trends and climate variables (air temperature, radiation), geographic variables (latitude, elevation), and lake depth were investigated. Lake stratification dates were also calculated to understand the interaction between warming and stratification. Data from the European Space Agency’s Climate Change Initiative (CCI) Lakes project were used for additional validation.

The analysis revealed widespread excess lake warming during ice-off months across the Northern Hemisphere. The average ratio of LSWT trends during the ice-off month (LSWT_IOFF) to open-water trends (R_IOFF) was 1.4, indicating significantly greater warming during ice-off. This excess warming was strongly correlated with an earlier ice break-up, with 79% of lakes showing a significant advancement in ice-off dates. The spatial pattern of LSWT_IOFF trends closely mirrored the trends in ice-off dates. While surface air temperature and radiation explained only a small portion of the spatial variation in R_IOFF, the trend in ice-off date explained a substantial 37%. Lake depth also played a role, mediating the magnitude of excess warming; deeper lakes showed less sensitivity to changes in ice-off dates. The sensitivity of LSWT to a 1-day advancement in ice-off (S_IOFF) averaged 0.14°C day⁻¹, meaning an 8.1-day advancement (observed from 1979–2020) contributed to 1.1°C of excess warming during ice-off. In contrast, the impact of later ice-on dates (S_ION) was smaller (0.06°C day⁻¹). Analysis showed that both S_IOFF and S_ION were influenced by incoming radiation and lake depth; shallower lakes and higher radiation levels resulted in greater sensitivity. Satellite observations corroborated these findings, showing excess warming during ice-off in a substantial proportion of lakes.

The study's findings clearly demonstrate that earlier ice loss significantly accelerates lake warming in the Northern Hemisphere. The strong correlation between earlier ice-off dates and excess warming during ice-off highlights the crucial role of lake ice phenology in driving lake thermal dynamics. The observed excess warming is primarily driven by increased incoming solar radiation due to reduced albedo and an extended period of open water. The influence of lake depth underscores the importance of considering lake morphology when assessing climate change impacts on lake thermal regimes. The findings are significant because they quantify the magnitude of excess warming attributable to changing ice phenology, which has important implications for lake ecosystems.

This study provides robust evidence for the amplified warming of Northern Hemisphere lakes due to earlier ice-off dates. The significant contribution of ice phenology to excess warming necessitates its inclusion in future climate change impact assessments on lake ecosystems. Further research should focus on the long-term consequences of excess warming on lake biodiversity, water quality, and ecosystem services, particularly in light of projected future alterations in ice phenology and the increasing frequency of ice-free winters.

The study's reliance on modeled and satellite data introduces some limitations. Satellite data represent skin-surface temperature, which may differ slightly from bulk water temperature. The use of a 1D lake model simplifies lake bathymetry, potentially introducing uncertainties in projections. The model doesn't incorporate temporal changes in water clarity, which can affect both surface and bottom water temperatures. Additionally, the constant lake depth and surface area in the model may be a source of uncertainty, especially in shallow lakes with fluctuating water levels. Despite these limitations, the study provides a valuable contribution to understanding lake thermal changes in a warming world.