Declining lake ice in response to climate change can impact spending for local communities

financial values associated with lake ice. Articles were chosen if they explicitly mentioned lake ice and had a monetary value attached to them. We searched using these terms in both Google and Google Scholar through over 20 pages until no other documents could be found in English. We included all articles that had economic data within them (n = 31). Most of these articles were newspaper articles or government documents and none were peer-reviewed articles. In the second search, we used the above search terms along with the names of several spe- cific lakes that we identified in the first search as having some form of monetary revenue. We acknowledge that our search used only English search terms, potentially neglecting articles written in the languages of the target countries. Thus, we supplemented our searches by asking limnologists across North America, Europe, and Asia from the Global Lakes Ecological Obser- vatory Network (GLEON) for additional examples. The monetary values identified in each article were presented as a number or in a table. We extracted these values into a spreadsheet and converted all foreign currency into United States Dollar (USD) using the January 2022 exchange rate. To project the revenue changes in lake ice activities, we maintained the reported monetary value constant, factoring in only the variations in lake ice duration. This involved multiplying the monetary value by the percentage change in ice duration for each year until 2099, specific to the corresponding year and geographic area outlined in the respective study.

We obtained country-level information for countries with lake ice in the Northern Hemi- sphere from the World Bank Indicator Database (https://data.worldbank.org/indicator). For all countries with lake ice, we obtained information on Gross Domestic Product in 2019 (USD), total annual freshwater withdrawals in 2017 (billion cubic meters), and greenhouse gas emissions in 2016 (metric tons per capita). For the political borders of each country, we used the administrative boundaries provided by the Database of Global Administrative Areas (https://gadm.org/). In total, 50 countries were identified to have lake ice in the Northern Hemisphere. For each country, we calculated the total area of lake ice across all pixels found within the political boundaries. The area of lake ice for each country was average across all lake models and climate scenarios but separated for each year and each of the three RCP scenarios. We determined the area of ice lost for each country by subtracting the mean area observed between 1970–1999 from 2070–2099. Statistical analysis To determine if changes in the surface area of lake ice were associated with the socio-economic characteristics of countries, we fit linear models with the difference in area of lake ice (between the two timeframes) as the response variable. As predictors, we crossed the RCPs with each of the four socio-economic variables: GDP, annual freshwater withdrawal, and greenhouse gas emissions. Both the change in the area of lake ice and each socio-economic variable was log-

transformed to meet assumptions of normality and account for the order of magnitude differ- ences in the variables of some countries. All analyses were conducted in R Version 4.1.2 [38] using packages raster [39], ncdf4 [40], zoo [41], and tidyverse [42]. All code used for the estima- tion of lake ice patterns, economic calculations, and statistical analyses are available at https:// afilazzola.github.io/IceVariability/. Results Climate change effects on lake ice Under climate change, lakes are transitioning to shorter ice durations or not freezing entirely. Lakes can display four types of ice phenologies including permanently frozen, seasonally PLOS ONE Financial loss of declining lake ice PLOS ONE | https://doi.org/10.1371/journal.pone.0299937 July 5, 2024 4 / 15