Benchmarking lake ice thickness for the Northern Hemisphere and its response to future climate

Abstract:

Lake ice features (e.g., the ice thickness) are important to the lake ecosystem, regional hydro-climate system, and human activities on the ice. However, information on the lake ice thickness (LIT) is extremely limited due to the difficulty in deriving in situ measurements, which is further exacerbated by the absence of the specialized remote sensing platform. Despite intensive development and applications of the lake ice models driven by the general circulation model outputs, the evaluation of the global LIT was mostly based on hypothesized lakes in each grid cell of the climate forcing data. Therefore, a benchmark for authentic global LIT is urgently needed. In this study, we used satellite altimetry (i.e., Jason-1/2/3) to retrieve ice thickness (altimetric ice thickness) in 11 global large lakes including Lake Baikal, Great Slave Lake, and Great Bear Lake. The altimetry-derived ice thickness showed good consistency with in situ measurements in Baker Lake and Great Slave Lake with the correlation coefficient over 0.8 and the root-mean-square error (RMSE) ~0.2 m. In addition, a 1-D lake ice model driven by remotely sensed products (i.e., MODIS surface temperature and albedo) and reanalysis forcing data was developed and validated using the altimetric LIT, showing satisfactory performances in the 11 large lakes. Subsequently, we derived the global LIT (annual maximum ice thickness) for 1260 lakes larger than 50 km² during 2003–2018 and projected changes of the LIT in 2091–2099 under RCP 2.6 and RCP 6.0. Results show a potential decrease of ~0.3 m in global mean annual maximum LIT under the high concentration pathway, thereby providing important insights into how the global LIT evolves under a changing environment.