Widespread loss of safe lake ice access in response to a warming climate

and ice-off (ice phenology) is changing around the Northern Hemisphere. Lake ice duration has been a central concern, where long-term records have shown an overall increase in open- water days of 0.62 days per decade between 1931–2005, with a nonlinear increase in the rate of open-water days in the last 30 years of the time series [2]. Lake ice phenology through long- term records extending over a century (1846–2019) show ice formation to be 11 days later and ice breakup to be 6.8 days earlier, which was a near doubling of ice loss compared to the same lake ice records between 1846–1995 [3]. Moreover, lake ice is thinning at a rate of 0.033 m per C increase in air temperature [1]. Therefore, long-term lake ice records show an overall loss of ice cover duration, timing, and thickness directly related to warming air temperatures [1,4]. However, climate change and warming winters have farther reaching effects by also degrading lake ice quality conditions [5,6]. Ice quality depends in large part on winter air temperatures, where warm temperatures induce diel melt-freeze cycles and/or precipitation to fall as both snow or rain. These temperature and precipitation conditions lead to white ice formation from freezing slush that introduces many air bubbles and small ice crystals that reflect light [7,8]. White ice is also notable for being structurally weaker than black ice [9]. The closer to 0˚C the air temperature is, the weaker the ice becomes [10]. For example, white ice at -0.5˚C is 51% weaker than black ice

at the same temperature, owing to the increased strength of ice with increased density [10]. White ice formation is generally expected toward the end of the season when temperatures warm and snow atop the ice column leads to a larger fraction of white ice as the ice column evolves through- out the season toward breakup [5]. However, white ice conditions are possible at the beginning of the season as well when ice formation and snow conditions coincide to form larger proportions of white ice [11]. Therefore, warming conditions expected throughout the 21st Century may lead to a more consistent presence of white ice throughout the season. Deteriorating ice conditions pose a significant safety hazard for human activities on frozen lakes. Freshwater ice plays a critical role in human winter mobility across the Northern Hemi- sphere [12] spanning millions of lakes [13]. Warmer air temperatures, however, are linked to increased search-and-rescue events as well as non-fatal and fatal ice-related drownings [14,15]. While the majority of drownings occur during seasonal transitions when ice is weakest (begin- ning and end of the ice season) [15], even the peak ice season can pose dangers if ice quality is compromised. This was evident in the February 2021 tragedy in Sweden, where fatalities occurred despite seemingly sufficient ice thickness due to the absence of the stronger “black ice” layer and the sole presence of weaker “white ice” [5]. In this study, we build on earlier global lake ice research [1,16–18] by investigating the

influence of climate warming and ice quality on the duration of the transition period between unsafe and safe ice conditions on Northern Hemisphere lakes. Safe ice conditions are defined by a combination of thickness and ice quality, where 10 cm of black ice or 20 cm of white ice are required for a person to safely walk on the ice [5]. By incorporating the period of unsafe ice cover in the analysis, this work adds vital context to ice phenology studies and directly addresses the impact of changing climate on ecosystem services, specifically pertaining to safe ice use. We aim to address three key questions: 1) how will the timing of safe ice availability change during various climate warming scenarios (1˚C, 2˚C, and 4˚C), 2) how will different ice quality conditions (100% black ice to 100% white ice) interact with warming climates, and 3) what are the potential consequences of changing ice phenology on safe ice use? We antici- pate a lengthening unsafe transition period for both ice formation and melting due to warming air temperatures impacting the overall ice cover season, ice thickness, and ice quality. We posit that a slower ice development at the beginning of winter will delay the period of safe ice avail- ability, while rapid spring warming will hasten the breakdown of the ice cover below safe thick- ness levels. Furthermore, diminished ice quality is expected to extend the unsafe period due to the requirement of greater ice thickness for safe weight support. PLOS ONE Changing ice conditions inhibit safe use

PLOS ONE | https://doi.org/10.1371/journal.pone.0313994 December 11, 2024 2 / 19 html. Derived CESM2-LE data for this analysis are available at https://doi.org/10.6084/m9.figshare. 26882467.v1 Code to reproduce the figures is available at https://github.com/jculpepper7/ice_ safety_2024. Funding: We would also like to thank the Natural Sciences and Engineering Research Council Discovery Grant, the York University Research Chair programme, and ArcticNet, a Network for Centres of Excellence Canada to SS for providing funding to support this research. RIW was supported by a UKRI Natural Environment Research Council (NERC) Independent Research Fellowship [grant number NE/T011246/1]. Lei Huang was supported by the National Natural Science Foundation of China (No. 42201049) and the Second Tibetan Plateau Scientific Expedition and Research (STEP) (grant number 2019QZKK0202). The CESM2-LE simulations presented here have been conducted through a partnership between the Institute for Basic Sciences (IBS) Center for Climate Physics (ICCP) in South Korea and the Community Earth System Model (CESM) group at the National Center for Atmospheric Research (NCAR) in the United States, representing a broad collaborative effort between scientists from both centers. The simulations were conducted on the IBS/ICCP supercomputer “Aleph.” The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist.