Cold-Region Rivers

River Ice Identification and Spectral Classification

We study relative band differences and machine-learning methods for identifying river ice on the Tibetan Plateau. Work published in 2025 addressed river ice identification in complex terrain; a 2026 study combined Landsat 5–9 and Sentinel-2 imagery with explainable machine learning for fine-grained spectral classification of river ice.

The Tibetan Plateau river ice spectral classification samples cover winter imagery from 2000–2024 and are available on request from the authors.

Long-Term Distribution and Phenology

We used Landsat imagery from 1999–2018 to construct a river ice distribution time series for the Babao River basin and analyse the relationships between river ice occurrence, terrain and air temperature. The results were published in Remote Sensing of Environment (2020).

Overlapping-orbit imagery from Sentinel-2 and Landsat 8 is combined to identify freeze-up, stable ice cover and melt phases. The methodology was published in Journal of Hydrology (2023). The data catalogue also includes monthly river ice data for the Yellow River basin for 1987–2022.

Ice Thickness Retrieval and Ground Validation

We estimate plateau river ice thickness using Sentinel-1 dual-polarisation C-band SAR, with validation data from RTK elevation surveys, drilling and repeated profile measurements. This work examines relationships between ice-surface elevation, ice thickness and radar signals. The results were published in Remote Sensing (2022).

Remote Sensing of River Discharge

We use channel radiometric information from near-infrared imagery to monitor discharge changes in small rivers in high-altitude mountain regions and develop remote sensing methods for estimating discharge in data-scarce catchments. The results were published in Water Resources Research (2019). A book chapter published in 2026 reviews river ice remote sensing methods and research advances in the context of climate change.