Hongxing Li

River ice is closely related to hydrological processes, water resource utilization, river hydraulic projects, and navigation systems, and the spatial distribution extent of river ice serves as critical foundational data for studying river ice. The Three Rivers Source Region (TRSR) is a representative high and cold mountain; in this region, river ice is well developed and significantly impacts the river runoff process. However, river ice data in this region are scarce, seriously limiting related research. In this study, a 30-meter spatial resolution river ice extent dataset (1990-2023) for the TRSR was successfully developed by employing the Google Earth Engine (GEE) and Landsat imagery, achieving an overall accuracy of 97.96% and a kappa coefficient of 0.86. The Operational Land Imager (OLI) sensor images resulted in the highest mapping accuracy, followed by the Thematic Mapper (TM) images and Enhanced Thematic Mapper (ETM) images. Notably, in areas with different elevations, slopes, and aspects, the mapping accuracy did not significantly differ. This dataset reveals, for the first time, that the average area of river ice distributed in the TRSR was 2073.0 km2, with a coverage rate of 0.64%, indicating a noticeable downward trend. Additionally, the river ice extent became more fragmented, and river ice was distributed more significantly in high-altitude regions and areas with gentle slopes than in other areas. However, no obvious characteristics regarding the slope aspect of the river ice distribution were observed. This research provides a comprehensive technical framework for large-scale river ice extraction using the GEE platform, and it offers an understanding of the distribution features and changes in river ice in the TRSR and even the broader Tibetan Plateau.

Net ecosystem exchange (NEE) is a central metric for assessing carbon cycling, and its accurate quantification is critical for understanding terrestrial-atmosphere carbon exchange dynamics. However, in complex alpine regions, high-resolution NEE estimation remains challenging due to limited observations and heterogeneous surface processes. To address this, we developed a multimodal feature fusion model (Multimodal-CNN-Attention-RF, MMCA-RF) that integrates convolutional neural networks (CNN) and random forest (RF) for NEE estimation in the Babao River Basin on the northeastern Tibetan Plateau. The model incorporates a cross-modal attention mechanism to dynamically optimize feature interactions, thereby better capturing the spatially heterogeneous responses of vegetation to environmental drivers. Results demonstrate that MMCA-RF exhibits strong stability and generalization, with R² values of 0.89 (training) and 0.85 (testing). Based on model outputs, the Babao River Basin acted as a carbon sink during 2017–2023, with a mean annual NEE of −100.86 gC m⁻² yr⁻¹. Spatially, NEE showed pronounced heterogeneity, while seasonal variation followed a unimodal pattern. Among vegetation types, grasslands contributed the largest total carbon sink, whereas open woodlands showed the highest sequestration efficiency per unit area. Driver analysis identified temperature as the dominant control on NEE spatial variation, with interactions between temperature, precipitation, and topography further enhancing heterogeneity. This study provides a high-accuracy modeling approach for monitoring carbon cycling in alpine ecosystems and offers insights into the stability of regional carbon pools under climate change.

The launch of Sentinel-1 satellite introduces a novel approach for synthetic aperture radar (SAR) interferometry (InSAR). However, its capabilities for topographic mapping are reportedly limited. It is very challenging to create high-quality digital elevation models (DEMs) from InSAR data in mountain area. The main goal of the study was to generate a new high-quality DEM by combining Sentinel-1A and Sentinel-1B data with a short temporal baseline and multiple perpendicular baselines from 43 to 143 m. Five DEMs were produced from five interferometric pairs. The five DEMs were fused to generate a more reliable fused DEM. The performance of the new DEMs was evaluated against ICESat-2/ATL06 product and global DEM products (NASADEM, advanced spaceborne thermal emission and reflection radiometer (ASTER) global digital elevation model (GDEM), and advanced land observing satellite ALOS World 3D-30 m). The results showed that all Sentinel-1 DEMs performed better than ASTER GDEM, four of DEMs had higher accuracies than NASADEM. The fused DEM with the vertical error of 9.08 m for steep terrain (slope>20°) revealed higher accuracy than three global DEMs. The accuracy of DEM was related to terrain slope and land cover type. The accuracies of DEMs decreased as slope increased. The DEMs in glacier area revealed higher errors than those in bare rock. Besides, there was no clear relationship between the perpendicular baseline and vertical accuracy of DEM. The interferometric pair with the shortest baseline (43 m) produced the worst quality DEM, while the interferometric pair with slightly shorter baseline (69 m) produced the highest quality DEM. The study revealed the outstanding accuracy of the new DEM product, which is very valuable data for local glacier research.

Lake ice phenology serves as a sensitive indicator of climate change in the lake-rich Northeast China. In this study, the freeze-up date (FUD), break-up date (BUD), and ice cover duration (ICD) of 31 lakes were extracted from a time series of the land water surface temperature (LWST) derived from the combined MOD11A1 and MYD11A1 products for the hydrological years 2001 to 2021. Our analysis showed a high correlation between the ice phenology measures derived by our study and those provided by hydrological records (R² of 0.89) and public datasets (R² > 0.7). There was a notable coherence in lake ice phenology in Northeast China, with a trend in later freeze-up (0.21 days/year) and earlier break-up (0.19 days/year) dates, resulting in shorter ice cover duration (0.50 days/year). The lake ice phenology of freshwater lakes exhibited a faster rate of change compared to saltwater lakes during the period from HY2001 to HY2020. We used redundancy analysis and correlation analysis to study the relationships between the LWST and lake ice phenology with various influencing factors, including lake properties, local climate factors, and atmospheric circulation. Solar radiation, latitude, and air temperature were found to be the primary factors. The FUD was more closely related to lake characteristics, while the BUD was linked to local climate factors. The large-scale oscillations were found to influence the changes in lake ice phenology via the coupled influence of air temperature and precipitation. The Antarctic Oscillation and North Atlantic Oscillation correlate more with LWST in winter, and the Arctic Oscillation correlates more with the ICD.

Deep learning (DL) has been widely applied to enhance the accuracy and reconstruction of snowpack parameters due to its powerful nonlinear mapping ability. Here, we propose a broadband snow albedo (BA) estimation method that infuses the asymptotic radiative transfer (ART) model into the DL model and thus imposes constraints on the input variables of the DL model and compares it to the MOD10A1 BA product and the BA simulated using the ART model alone. The results show that coupling a physical mechanism-based snow radiative transfer model with a DL model can successfully improve the estimation accuracy of BA. Compared to the MOD10A1 BA product and the BA simulated using the ART model alone, the accuracy of the BA simulated by the convolutional neural network model under the constraints of the ART model is improved by 12% and 3%, respectively, and the accuracy of the BA simulated by the long short-term memory model under the constraints of the ART model is improved by 14% and 5%, respectively. The coupling paradigm that integrates the snow radiative transfer model into the DL model effectively balances the advantages of both models. This coupling paradigm creates a complementary association between the snow radiative transfer model and the DL model. This research effectively enhances the simulation and predictive capabilities of snow albedo and is expected to promote the in-depth application of remote sensing big data in combination with physical modeling in the field of cryosphere parameter estimation.

Dust deposited on the surface of snow and glaciers can significantly reduce the snow and ice albedo and accelerate melting. Manual observations of the dust mass concentration (DMC) on snow and glacier surfaces are routinely performed at many locations worldwide. However, snow and ice surface DMC monitoring methods based on remote sensing data still face challenges. This study presents a new retrieval scheme for estimating dust load on snow-covered surfaces from MODIS and VIIRS in Northeast and Northwest China that utilizes a classical snow radiative transfer model. Our results indicate that the coefficient of variation (CV) of DMC retrieved from in situ measurements of snow spectral reflectance is 4%, which is within a +4% difference compared to DMC observed in snow and ice samples. Estimating atmospheric dust pollutants content deposited on snow surfaces based on satellite remote sensing observations is feasible. In Northwest China, the root mean square error (RMSE) of the DMC values retrieved from VNP09GA data is 9.78 ppm, while that of the DMC values retrieved from MOD09GA data is 13.74 ppm. In Northeast China, the RMSE of the DMC values retrieved from VNP09GA data is 73.98 ppm, while that of the DMC values retrieved from MOD09GA data is 184.32 ppm. The research results can realize continuous monitoring of the atmospheric dust pollutants deposited on snow surfaces, which is of great practical significance to understanding and studying the pollution process of atmospheric dust on snow.

In this study, we classified the variability in snow cover persistence across China by using a novel method; continuous snow cover days and variability of snow cover were used as the evaluation indicators based on a long-term Advanced Very High Resolution Radiometer (AVHRR) snow cover extent (SCE) product. The product has been generated by the snow research team in the Northwest Institute of Eco-Environment and Resources (NIEER), Chinese Academy of Sciences. There were obvious differences in snow cover classification in three snow cover areas (northern Xinjiang, northeast China, and the Tibetan Plateau): northern Xinjiang was dominated by persistent snow cover, most regions of northeast China were covered by persistent and periodic variable snow cover. There was the most abundant snow cover classification in the Tibetan Plateau. The extents of persistent and periodic variable snow cover were gradually shrinking due to rising temperatures and decreasing snowfall during 1981–2019. In contrast, non-periodic variable snow cover areas increased significantly. This method takes into account the stability, continuity, and variability of snow cover, and better captures the characteristics and changes of snow cover across China. Based on our research, we found that snow disasters in ephemeral-type (belong to non-periodic variable snow cover) regions cannot be well prevented because of the unfixed snow cover timing. Therefore, we recommend that monitoring and forecasting of snow cover in these snow cover regions should be strengthened.

Nanxiaohegou watershed is one of the experimental basins used to study soil and water conservation in the gully regions of Loess Plateau by Xifeng Soil and Water Conservation Scientific Experimental Station of the Yellow River Conservancy Commission (hereinafter referred to as Xifeng Station). The establishment of Xifeng Station is approved in 1951. And Nanxiaohegou was selected as a comprehensive treatment experimental base in 1952. Runoff fields and precipitation stations have been set up since 1954. Long-term runoff and sediment yields from different runoff fields designed with different micro-topography, slopes, land cover types and treatment measures have been obtained. This dataset is a collection of observation data from 154 runoff fields in Nanxiaohegou Watershed from 1954 to 2018. The observation data includes precipitation, runoff, sediment volume, and soil moisture content. This dataset will provide scientific, reliable, and long-term observation data for studying the law of soil and water loss in gully regions of Loess Plateau, building a model for predicting soil erosion, and developing soil and water conservation measures.