Source Data for: Accelerating lake ice loss threatens the viability of the winter economy in Northeast China

This dataset contains the underlying source data for Figures 2, 3, 4, and 5 of the manuscript "Accelerating lake ice loss threatens the viability of the winter economy in Northeast China". The dataset presents century-scale reconstructed lake ice phenology (Freeze-up, Break-up, and Ice Duration) for 32 lakes in Northeast China, spanning from 1901 to 2100. Data were generated using a physics-informed hybrid model driven by CMIP6 meteorological forcing. Contents: Historical Period (1901–2014): Simulations driven by CMIP6 historical experiments. Future Period (2015–2100): Projections driven by CMIP6 scenarios (SSP126, SSP370, and SSP585). Methodology Note: The hybrid model couples a thermodynamic lake ice core with a machine learning residual correction module to reduce simulation errors in ice phenology. Data Structure: The data is provided in a single Excel file (.xlsx), where each worksheet corresponds to a specific figure panel in the manuscript.本数据集为论文《加速湖泊冰盖消退威胁中国东北地区冬季经济可持续性》（Accelerating lake ice loss threatens the viability of the winter economy in Northeast China）的图2、图3、图4及图5提供原始基础数据。 本数据集涵盖中国东北地区32个湖泊1901年至2100年的百年尺度湖泊冰物候（封冻、解冻与冰期时长）重建结果，其数据由依托第六次耦合模式比较计划（CMIP6）气象强迫场驱动的物理信息混合模型生成。 数据内容： 历史时段（1901–2014）：采用CMIP6历史试验驱动的模拟结果。 未来时段（2015–2100）：采用CMIP6情景（共享社会经济路径126（SSP126）、共享社会经济路径370（SSP370）及共享社会经济路径585（SSP585））驱动的预测结果。 方法说明：该混合模型耦合湖泊热力学冰芯模型与机器学习残差校正模块，以降低冰物候模拟误差。 数据结构：本数据集以单个Excel文件（.xlsx）形式提供，每个工作表对应论文中的特定图面板。