Search USGS⌕ Search
Home/Archive/USGS

USGS · 70270412

Modeling daily ice cover in northern hemisphere lakes with a long short‐term memory neural network

Xinchen He·Konstantinos M. Andreadis·Allison H. Roy·Theodore Langhorst·Abhishek Kumar·Caitlyn S. Butler
View original source ↗Export metadata ↓

Abstract

Quantifying lake ice loss is crucial for understanding the impact of climate change on lake ecosystems. In this study, we trained a deep learning model (Long-Short Term Memory with Landsat observations, 1984–2012) to simulate Northern Hemisphere lake ice changes at a fine spatial scale (> 0.1 km 2 ) from 1980 to 2022. The model achieved good performance overall during the test period (2013–2022), and the derived ice-on and ice-off matched well with two independent ice phenology data sets. Results reveal a 76.8% increase in intermittently ice-covered lakes from the 1980s to the 2010s, alongside a 10.7-day shorter ice duration and a 3.9 percentage-points reduction in annual mean ice cover fractions. The model can track daily partial ice cover changes, providing a novel contribution to understanding shifts in lake ice cover with climate change. These findings can provide valuable insights for future limnology studies, such as improving estimates of greenhouse gas emissions from lakes.

Explore related subjects

Geophysical Research LettersArticle

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Xinchen He, Konstantinos M. Andreadis, Allison H. Roy, Theodore Langhorst, Abhishek Kumar, Caitlyn S. Butler. 2025-06-17. Modeling daily ice cover in northern hemisphere lakes with a long short‐term memory neural network. https://doi.org/10.1029/2024gl113544

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Evolution of summit magma pressure during the 2018 caldera collapse at Kılauea volcano

We utilized geodetic data from the 16 May to 4 August 2018 caldera collapse sequence to refine the geometries and pressure evolution of two summit reservoirs at Kīlauea volcano: the shallower Halema‘uma‘u (HMM) and deeper South Caldera (SC) reservoirs. We modeled magma reservoirs as pressurized spheroidal cavities and constrained reservoir geometries with Global Positioning System (GPS) and Interferometric Synthetic Aperture Radar displacements. Inverting high-rate GPS and borehole tilt time series recovered time-dependent reservoir pressure changes and caldera ring fault shear stress. During the collapse sequence, the HMM reservoir accommodated episodic, co-collapse pressure increases of 1.9 MPa (median) and a net 5.8 MPa (median) decrease. Simultaneously, the SC reservoir underwent monotonic depressurization of 6.7 MPa (median). Negligible co-collapse pressure changes in the SC reservoir imply it was not exclusively supplying magma to the eruption vent. These results provide a coherent view of magma reservoir pressure evolution during the 2018 eruption.

Hawaii↗
USGS2026

EMIT reveals elevated calcite signatures in shorelines of Lake Powell and Lake Mead exposed by drought and Colorado River management

Nearly half of Earth’s seasonal freshwater storage variability occurs in human-managed reservoirs, resulting in large areas of reservoir-bed that are episodically inundated or subaerially exposed. Here we use hyperspectral data from NASA’s Earth Surface Mineral Dust Source Investigation (EMIT) mission to map calcite mineral assemblages across exposed surfaces in two large U.S. reservoirs, Lake Mead and Lake Powell. Results show distinct patterns in the occurrence of carbonate, clay, and mica minerals. Reservoir-deposited sediments exhibited significantly higher calcite proportions than adjacent, lithologically similar, non-inundated terrestrial surfaces. These differences likely reflect hydrologically-driven calcite precipitation, wherein warm, clear, productive reservoir waters drive calcium carbonate out of solution. We also detect reductions in calcite proportions in areas of the shoreline disturbed by human recreation. EMIT's imaging capabilities enable rapid, high-resolution mineralogical inventories across vast areas where field sampling is infeasible. We find EMIT is a scalable tool for monitoring geochemical changes associated with large water level fluctuations.

Arizona, Nevada, Utah↗
USGS2026

First application of a coastal-dune model (AeoLiS) to simulate fluvial-aeolian morphodynamics

Quantifying aeolian processes is important to decision-making in river management and for optimizing strategies for ecosystem conservation, natural resource management, and recreation. We applied the AeoLiS aeolian transport model, originally developed for coastal environments, to a dryland river corridor to evaluate its ability to simulate sediment dynamics in a fluvial-aeolian system. The model reproduced spatial and temporal patterns of geomorphic change at a Grand Canyon sandbar, including aeolian dune formation, observed in field data. The model effectively captured trends in observed sediment movement in the spring and summer months. Thus, despite assuming a single grain-size, static vegetation, and not directly accounting for variable sediment supply from fluvial sources, the model performs well at capturing aeolian processes in a complex fluvial-aeolian system. AeoLiS demonstrates strong potential for dryland river applications and management, due to its flexibility and compatibility with coupled modeling frameworks.

Arizona↗