Glacier Mapping Research Articles

Application of Artificial Intelligence in Glacier Studies: A State-of-the-Art Review

Assessing glaciers using recent and historical data and predicting the future impacts on them due to climate change are crucial for understanding global glacier mass balance, regional water resources, and downstream hydrology. Computational methods are crucial for analyzing current conditions and forecasting glacier changes using remote sensing and other data sources. Due to the complexity and large data volumes, there is a strong demand for accelerated computing. AI-based approaches are increasingly being adopted for their efficiency and accuracy in these tasks. Thus, in the current state-of-the-art review work, available research results on the application of AI methods for glacier studies are addressed. Using selected search terms, AI-based publications are collected from research databases. They are further classified in terms of their geographical locations and glacier-related research purposes. It was found that the majority of AI-based glacier studies focused on inventorying and mapping glaciers worldwide. AI techniques like U-Net, Random forest, CNN, and DeepLab are mostly utilized in glacier mapping, demonstrating their adaptability and scalability. Other AI-based glacier studies such as glacier evolution, snow/ice differentiation, and ice dynamic modeling are reviewed and classified, Overall, AI methods are predominantly based on supervised learning and deep learning approaches, and these methods have been used almost evenly in glacier publications over the years since the beginning of this research area. Thus, the integration of AI in glacier research is advancing, promising to enhance our comprehension of glaciers amid climate change and aiding environmental conservation and resource management.

Annual Minimum Snow/Ice Extent Variations over Greenland since 2000: Ice Sheet, Peripheral Areas, and Relation to Ice Mass Balance

Abstract A novel satellite image processing technique was utilized to produce an annual time series of the minimum snow/ice (MSI) extent over the entire Greenland landmass for the period 2000–22. The information was derived from the Moderate Resolution Imaging Spectroradiometer 10-day clear-sky composites over the April–September period. The data products were generated from 250-m swath imagery. The annual aggregates were downscaled to a 150-m grid for consistency with data on margins of Greenland available from the Geological Survey of Denmark and Greenland (GEUS) and the Greenland Ice Mapping Project. Interannual variations in the MSI extent were derived and analyzed for each of the seven major glacier basins in Greenland split into the main ice sheet, represented by a static map, and the peripheral areas from which all variations originated. Four of the seven regions demonstrated statistically significant negative trends in the MSI extent. The entire Greenland area also showed a declining snow/ice extent although this was not statistically significant. The region-wide and peripheral snow/ice extent varied from a minimum of 1.807 × 106 km2 (1.449 × 105 km2 for peripheral areas) observed in 2012 to a maximum of 1.860 × 106 km2 (1.977 × 105 km2) observed in 2006 with an average value of 1.829 × 106 km2 (1.664 × 105 km2). The derived MSI variations showed a statistically significant correlation with the near-surface 2-m air temperature from the ERA5-Land reanalysis and Greenland ice mass balance from GEUS for all catchments, with correlation coefficients for the entire area equal to −0.74 and 0.53, respectively. The mapping of many peripheral glaciers and ice shelves included in the glaciology databases and utilized for the IPCC reporting is not always consistent with our results and requires improvement, especially in coastal areas.

Comprehensive assessment of rock glaciers in the Himachal Himalayas: Updated inventory and labelling

Rock glaciers are geomorphological features often used as a visible expression of mountain permafrost. These are creeping ice-debris with distinct ridge and furrow structures on the surface with a steep frontal slope. Rock glaciers, being the valuable past permafrost indicators, also have utmost hydrological significance in near future. Therefore, mapping of rock glaciers is an important step in order to understand permafrost regimes better. The Himalayas have large occurrences of these features and this study in Himachal Himalayas complied 789 rock glaciers, covering an area of about 336.2 km2. Different labels based on genesis, location, shape, form, surface relief and activity revealed rock glaciers were mainly derived from talus slopes (239) and exhibited tongue shape (377), primarily found in cirques (531). Most of them were classified as simple units (603) with well-developed surface relief (387), and they were found to be predominantly intact (760). The topographical parameters suggest majority of the rock glaciers are located between 4000 and 4800 m with mean elevation to be 4635 m. These rock glaciers are present at gentle to steep slope gradient (0 to 45°) with curvature ranging between −3.5 and 4.5, and majority showing convex curvature. The slope aspect conducive for formation of rock glaciers in Himachal is northerly (N, NW, NE). Principal geology for these rock glaciers belongs to slate, phyllite, quartzarenite, limestone and meta basics. The climatic parameters and indices also affect the rock glaciers occurrence significantly. The mean land surface temperature (LST) for majority rock glaciers lies between 0 and −15°C. While, the mean NDSI of all the rock glaciers varies from 0.04 to 0.68 and mean NDVI varies from −0.06 to 0.08. Overall, the inventory along with labels is a valuable database for understanding the distribution and characteristics of rock glaciers in the Himachal Himalayas.

Rock glaciers as proxy for machine learning based debris‐covered glacier mapping of Kinnaur District, Himachal Pradesh

AbstractThis research introduces an innovative approach by utilising rock glaciers (RGs) as a proxy for mapping debris‐covered glaciers (DCGs). This approach focuses on the interconnected nature of glaciers, DCGs and RGs in a continuum where DCGs can transform into RGs over time due to various processes. This study utilises six machine learning models—logistic regression (LR), support vector machine (SVM), K‐nearest neighbour (KNN), Naïve Bayes (NB), decision tree (DT) and random forest (RF)—combined with multispectral satellite data (Sentinel‐2 and Landsat 8) and topographical data derived from ALOS PALSAR DEM. Performance metrics such as accuracy, area under the curve (AUC) score, precision, recall and F1‐score were evaluated to assess model performance. This detailed mapping provides a precise estimation of the extent of DCGs in the Kinnaur district. The estimated DCG areas revealed intriguing variation across models, with RF (9.71%), KNN (9.67%) and NB (9.41%) yielding similar predictions. SVM (11.61%) projected a slightly larger DCG area, whereas DT (5.54%) and LR (25.55%) provided contrasting results. Validation against high‐resolution satellite images, Google Earth images and glacier inventories confirmed the accuracy and reliability of our approach. Based on our findings for our specific study, the most effective method for mapping DCGs is RF, followed by KNN, NB, DT and SVM. The combination of machine learning models and RG data presents a novel and promising approach to remote sensing‐based DCG mapping, with potential applications for other regions and broader environmental studies.

Glacial geomorphology of the Vanch River basin, the western Pamirs, Tajikistan

ABSTRACT Geomorphological mapping of glaciers and glacial landforms helps describe the current environment and investigate past glacial changes. Here we present the first detailed 1:80,000 glacial geomorphological map of the Vanch River basin (about 2100 km2), the western Pamirs, Tajikistan [38.6667°N, 72.0000°E]. Mapping was performed using DEM (30 m), Landsat 8 imagery (15 m), and Google Earth images (∼1 m). The resultant map provides the spatial distribution of glaciers and associated erosional and depositional landforms such as cirques, glacial valleys, rock glaciers, and moraines. Morphology of the lower section of the main valley outlines the limits of Vanch Glacier in the Middle Pleistocene. Comparison with former glacier inventories indicates that the glaciers have generally been stable, but the proportion of debris-covered areas has increased since 2007. The identified 184 intact rock glaciers represent hidden water storage, providing a new basis for estimating water resources in the basin.

Surface facies analysis of the Gangotri and neighbouring glaciers, central Himalaya

Glaciers are primarily monitored using medium-to-high resolution satellite data, undermining the potential of coarse-resolution data. In pursuance of this, high resolution 10 m super-resolved glacier maps derived from 56 m coarse-resolution AWiFS data are applied here to assess the facies, firn-line altitude, and frontal variations of the Gangotri and neighbouring glaciers, central Himalaya between 2005 and 2017. The wet and warming trends estimated over the study area appear to have caused excess firn (56.53 ± 6.22%) and ice (27.50 ± 3.03%) melting, contributing to the significant progression in fresh and slightly metamorphosed snow (12.09 ± 1.33%), wet-snow (21.79 ± 2.40%), ice-mixed debris (9.24 ± 1.02%) and supraglacial debris (2.49 ± 0.27%) during 2005–2016. Mean firn-line of the study glaciers has ascended from 5327 ± 23 m to 5376 ± 24 m at an average rate of 3.44 ± 0.45 m a−1 during 2005–2016. Mean firn-line altitude ascent is the highest for the sparsely debris-covered (<10% debris) Arwa glacier followed by the extensively debris-covered (≥35% debris) Gangotri, Bhagirathi-Kharak and Satopanth glaciers. Contrastively, the moderately debris-covered (17–29% debris) Raktvarn and Chaturangi glaciers show slight variations in their mean firn-line altitudes. These firn-line variations are governed by the rising average annual temperature, glacier size and predominant glacier facie. All the glaciers show an overall tendency of termini retreat at variable rates during 2005–2017. The highest retreat rate is estimated for the Gangotri glacier (12.01 ± standard deviation: 8.16 m a−1) followed by Chaturangi (7.97 ± 5.79 m a−1), Bhagirathi-Kharak (5.99 ± 9.26 m a−1), Raktvarn (3.28 ± 2.28 m a−1), Satopanth (1.89 ± 2.87 m a−1), and Arwa (0.85 ± 1.90 m a−1) glaciers. These retreat rates vary significantly with the exclusion of static points in the retreat estimation, revealing its subjective nature. The temporal facies maps obtained here have the potential for the hydrological modelling of meltwater production of the study glaciers.

MONITORING OF GLACIERS ON HORSESHOE ISLAND, ANTARCTICA BASED ON A DEEP LEARNING APPROACH FROM HIGH-RESOLUTION ORTHOPHOTOS (TAE-6 &amp; TAE-7)

Abstract. Global climate change is a phenomenon that seriously affects the balance of a wide variety of ecosystems and is the intense focus of climate scientists and environmental researchers. In this context, periodic monitoring of glacier areas in terms of a better understanding of atmosphere-ocean interactions; thus, predicting the effects of climate change and planning against future threats by evaluating environmental impacts is an important research area. Especially the polar regions, where the melting of glaciers and the rise of sea levels are visibly observed, are important for climate scientists in providing crucial observations to understand and predict global climate change. In this study, within the scope of the international bilateral cooperation project carried out in cooperation with Istanbul Technical University (ITU) and the Bulgarian Academy of Sciences (BAS) (Project No: 121N033), the spatial changes in snow/glacier areas obtained from UAV Photogrammetry products generated during the 6th and 7th Antarctic National Science Expeditions. Snow/glacier areas were segmented with the K-Net deep learning approach which has been previously tested for accuracy and provides glacier mapping with accuracy metrics over 99%, on the high spatial resolution orthophotos produced during the two periods. The snow/glacier areas difference between the two periods were calculated and compared and water bodies which are critical areas, were specifically examined. The result of this comparison shows that the glacier area decreased by approximately 11% in just 1 year. However, to better understand these changes in snow/glacier areas, the region needs to be observed closely for longer time periods. It is thought that future studies will contribute to efforts to manage global environmental impacts and cope with climate change by focusing on monitoring and better understanding changes in these critical regions.

Updating glacier inventories on the periphery of Antarctica and Greenland using multi-source data

Abstract Melting and calving of glaciers and ice caps in Antarctica and Greenland could potentially contribute significantly to global sea level rise. Updates to existing outlines that provide critical glacier baseline information in both regions could help in the analysis of particular changes in glacier parameters such as area and volume from time-series inventories. Here we synthesize previously established techniques and apply new multi-source datasets to update glacier outlines in selected test areas of Antarctica and Greenland, as well as to reduce uncertainties and errors during the mapping process. The workflow includes mapping glacier boundaries, subdividing glaciers by watersheds and assigning glacier attributes. Complicated glacier scenarios and updating challenges in polar regions are discussed and demonstrated by representative case studies. For the first time in Antarctica, we analyze the effect of terminus types on mapped glacier areas, and in Greenland we compare the differences with glacier mapping results using Landsat OLI and ETM+. With new data sources, the methods described in this study might help to create glacier outlines on a larger scale in Antarctica and Greenland. Although data sources can be substituted, the enormous amount of manual labor required to update glacier inventories remains a significant challenge.

Temporal variation in glacier surface area and glacial lakes in glaciated river basins of Arunachal Pradesh

Abstract Climate change-induced glacier recession has sparked a dynamic transformation of glaciers in high-mountain areas worldwide, resulting in genesis, expansion, and dissipation of glacial lakes that pose a potential threat to downstream communities, underscoring the need for regular monitoring. This study incorporates the automated glacier extraction index to improve debris-free glacier mapping accuracy by reducing water and shadow area categorization errors using multi-temporal Landsat-5 Thermal Mapper, Landsat-8 Operational Land Imager, and digital elevation model (Shuttle Radar Topography Mission) to obtain precise glacier extent. Waterbodies were delineated using thresholding techniques applied to water indices: normalized difference water index, modified normalized difference water index, normalized difference turbidity index, and slope information. The study attempted to map and analyze the temporal variations of glaciers and glacier lakes in the glaciated sub-basins of Arunachal Pradesh: Dibang Basin, Mago Basin, and Subansiri Basin. Glaciers receded at a shrinkage rate of 1.08% per year (1988–2017) in Mago Basin, 1.01% per year (2013–2022) in Subansiri Basin, and 1.42% per year (1995–2021) in Dibang Basin. Glacier lakes showed an increasing trend in number and area due to glacier melt, expanding at rates of 3.31% in the Mago Basin, 0.48% in the Subansiri Basin, and 0.68% in the Dibang Basin per year. Findings suggest that global climate change is likely the primary factor driving glacier changes in these basins.

A Comparative Analysis of Machine Learning Techniques for National Glacier Mapping: Evaluating Performance through Spatial Cross-Validation in Perú

Accurate glacier mapping is crucial for assessing future water security in Andean ecosystems. Traditional accuracy assessment may be biased due to overlooking spatial autocorrelation during map validation. In recent years, spatial cross-validation (CV) strategies have been proposed in environmental and ecological modeling to reduce bias in predictive accuracy. In this study, we demonstrate the influence of spatial autocorrelation on the accuracy assessment of glacier surface predictive models. This is achieved by comparing the performance of several widely used machine learning algorithms including the gradient-boosting machines (GBM), k-nearest neighbors (KNN), random forest (RF), and logistic regression (LR) for mapping nine main Peruvian glacier regions. Spatial and non-spatial cross-validation methods were used to evaluate the model’s classification errors in terms of the Matthews correlation coefficient. Performance differences of up to 18% were found between bias-reduced (spatial) and overoptimistic (non-spatial) cross-validation results. Regarding only spatial CV, the k-nearest neighbors were the overall best model across Huallanca (0.90), Huayhuasha (0.78), Huaytapallana (0.96), Raura (0.93), Urubamba (0.96), Vilcabamba (0.93), and Vilcanota (0.92) regions, consistently demonstrating the highest performance followed by logistic regression at Blanca (0.95) and Central (0.97) regions. Our validation approach, accounting for spatial characteristics, provides valuable insights for glacier mapping studies and future efforts on glacier retreat monitoring. Incorporating this approach improves the reliability of glacier mapping, guiding future national-level initiatives.

Mapping Glacier Basal Sliding Applying Machine Learning

AbstractDuring the RESOLVE project (“High‐resolution imaging in subsurface geophysics: development of a multi‐instrument platform for interdisciplinary research”), continuous surface displacement and seismic array observations were obtained on Glacier d’Argentière in the French Alps for 35 days in May 2018. The data set is used to perform a detailed study of targeted processes within the highly dynamic cryospheric environment. In particular, the physical processes controlling glacial basal motion are poorly understood and remain challenging to observe directly. Especially in the Alpine region for temperate based glaciers where the ice rapidly responds to changing climatic conditions and thus, processes are strongly intermittent in time and heterogeneous in space. Spatially dense seismic and Global Positioning System (GPS) measurements are analyzed applying machine learning to gain insight into the processes controlling glacial motions of Glacier d’Argentière. Using multiple bandpass‐filtered copies of the continuous seismic waveforms, we compute energy‐based features, develop a matched field beamforming catalog and include meteorological observations. Features describing the data are analyzed with a gradient boosting decision tree model to directly estimate the GPS displacements from the seismic noise. We posit that features of the seismic noise provide direct access to the dominant parameters that drive displacement on the highly variable and unsteady surface of the glacier. The machine learning model infers daily fluctuations and longer term trends. The results show on‐ice displacement rates are strongly modulated by activity at the base of the glacier. The techniques presented provide a new approach to study glacial basal sliding and discover its full complexity.

Automatic Extraction of the Calving Front of Pine Island Glacier Based on Neural Network

Calving front location plays a crucial role in studying ice–ocean interaction, mapping glacier area change, and constraining ice dynamic models. However, relying solely on visual interpretation to extract annual changes in the calving front of ice shelves is a time-consuming process. In this study, a comparative analysis was conducted on the segmentation obtained from fully convolutional networks (FCN), U-Net, and U2-Net models, revealing that U2-Net exhibited the most effective classification. Notably, U2-Net outperformed the other two models by more than 30 percent in terms of the F1 parameter. Therefore, this paper introduces an automated approach that utilizes the U2-Net model to extract the calving front of ice shelves based on a Landsat image, achieving an extraction accuracy of 58 m. To assess the model’s performance on additional ice shelves in the polar region, the calving front of the Totten and Filchner ice shelves were also extracted for the past decade. The findings demonstrated that the ice velocity of the Filchner ice shelf exceeded that of the Totten ice shelf. Between February 2014 and March 2015, the majority of the calving fronts along the Filchner Ice Shelf showed an advancing trend, with the fastest-moving front measuring 3532 ± 58 m/yr.

94 GHz Radar Backscatter Characteristics of Alpine Glacier Ice

AbstractMeasuring the radar backscatter characteristics of glacier ice at different frequencies and incidence angles is fundamental to predicting the glacier mapping performance of a sensor. However, such measurements at 94 GHz do not exist. To address this knowledge gap, we collected 94 GHz radar backscatter data from the surface of Rhônegletscher in Switzerland using the All‐Weather Volcano Topography Imaging Sensor (AVTIS2) real‐aperture Frequency Modulated Continuous Wave radar. We determine the mean normalized radar cross section to be −9.9 dB. The distribution closely follows a log‐normal distribution with a high goodness of fit (R2 = 0.99) which suggests that radar backscatter is diffuse and driven by surface roughness. Further, we quantified the uncertainty of AVTIS2 3D point clouds to be 1.30–3.72 m, which is smaller than other ground‐based glacier surface mapping radars. These results demonstrate that glacier surfaces are an efficient scattering target at 94 GHz, hence demonstrating the suitability of millimeter‐wave radar for glacier monitoring.

The superiority of the Adjusted Normalized Difference Snow Index (ANDSI) for mapping glaciers using Sentinel-2 multispectral satellite imagery

ABSTRACT Accurate monitoring of glaciers’ extents and their dynamics is essential for improving our understanding of the impacts of climate and environmental changes in cold regions. The satellite-based Normalized Difference Snow Index (NDSI) has been widely used for mapping snow cover and glaciers around the globe. However, mapping glaciers in snow-covered areas using existing indices remains a challenging task due to their incapabilities in separating snow, glaciers, and water. This study aimed to evaluate a new satellite-based index and apply machine learning algorithms to improve the accuracy of mapping glaciers. A new index based on satellite data from Sentinel-2 was tested, which we call the Adjusted Normalized Difference Snow Index (ANDSI). ANDSI (besides NDSI) was used with five different machine learning algorithms, namely Artificial Neural Network, C5.0 Decision Tree Algorithm, Naive Bayes classifier, Support Vector Machine, and Extreme Gradient Boosting, to map glaciers, and their performance was evaluated against ground reference data. Four glacierized regions in different countries (Canada, China, Sweden, and Switzerland-Italy) were selected as study sites to evaluate the performance of the proposed ANDSI. Results showed that the proposed ANDSI outperformed the original NDSI, and the C5.0 classifier showed the best overall accuracy and Kappa among the selected five machine learning classifiers in the majority of cases. The original NDSI yielded results with an average overall accuracy of (around) 91% and the proposed ANDSI with (around) 95% for glacier mapping across all models and study regions. This study demonstrates that the proposed ANDSI serves as a superior and improved method for accurately mapping glaciers in cold regions.

Glacier changes and their impacts on the melt runoff in high and moderate elevation ranges of Himalayan Upper Ganges Basin

AbstractThe rise in temperatures over the Himalayan regions due to climate change have shown profound effects on the seasonality and long‐term availability of snow melt runoff and glacier melt runoff. As per the sixth assessment report by the Intergovernmental Panel on Climate Change, the Himalayan glaciers are retreating and the fraction of precipitation falling as snow is altering. This study analyses the effect of snow‐glacier changes on the melt runoff (1986–2020) over the Upper Ganga river basin (up to Rishikesh) using a fully distributed model namely, Spatial Process in Hydrology (SPHY), which governs by Temperature Index Model approach with varying degree‐day factors. This study performs the two‐step calibration approach for model parameterization and to evaluate the snow and glacier melt runoff utilizing the observed discharge and real time satellite datasets. The SPHY derived snow cover was compared with the Moderate Resolution Imaging Spectroradiometer (MODIS) derived snow covers. Results show that the SPHY modelled fractional snow covers found comparable to MODIS derived snow covers (~70%–90% correlation). The R2 based comparison of SPHY modelled runoff (Q) and observed Q at two gauges namely, Maneri (~0.85) and Rishikesh (~0.80) found reasonable. For the assessment of glacier changes and corresponding melt runoff, temporal glacier maps (i.e., for years 1990, 2000, 2010, 2020) were prepared utilizing the Landsat imageries. Results showed a noticeable reduction in the glacier areas (~5% mass reduction in 30 years) and an increase in glacier melt runoff at different stations (e.g., at Bhojwasa ~3%–4% and at Rishikesh ~1%–2%). Similarly, a decrease in snow melt runoff (~3%–4%) and an increase in rainfall runoff (~3%) were computed. A month wise percent of change analysis has shown significant fluctuations in the magnitude and seasonality of all runoff components (i.e., snow Q, glacier Q, base Q and rainfall induced Q) in the Himalayan Upper Ganga basin.

Brief communication: Glacier mapping and change estimation using very high-resolution declassified Hexagon KH-9 panoramic stereo imagery (1971–1984)

Abstract. The panoramic cameras (PCs) on board Hexagon KH-9 (KH-9PC) satellite missions from 1971–1984 captured very high-resolution stereo imagery with up to 60 cm spatial resolution. This study explores the potential of this imagery for glacier mapping and change estimation. We assess KH-9PC imagery using data from the KH-9 mapping camera (KH-9MC), KH-4PC, and SPOT and Pléiades satellite imagery. The high resolution of KH-9PC leads to higher-quality DEMs, which better resolve the accumulation region of the glaciers in comparison to the KH-9MC. On stable terrain, KH-9PC DEMs achieve an elevation accuracy of &lt;4 m with respect to SPOT and Pléiades DEMs. While the estimated geodetic mass balances using PC and MC data are similar after outlier filtering, the elevation change data show superior spatial coverage and considerably less noise when using KH-9PC data.

A Bibliometric and Visualized Analysis of Remote Sensing Methods for Glacier Mass Balance Research

In recent decades, climate change has led to global warming, glacier melting, glacial lake outbursts, sea level rising, and more extreme weather, and has seriously affected human life. Remote sensing technology has advanced quickly, and it offers effective observation techniques for studying and monitoring glaciers. In order to clarify the stage of research development, research hotspots, research frontiers, and limitations and challenges in glacier mass balance based on remote sensing technology, we used the tools of bibliometrics and data visualization to analyze 4817 works of literature related to glacier mass balance based on remote sensing technology from 1990 to 2021 in the Web of Science database. The results showed that (1) China and the United States are the major countries in the study of glacier mass balance based on remote sensing technology. (2) The Chinese Academy of Sciences is the most productive research institution. (3) Current research hotspots focus on “Climate change”, “Inventory”, “Dynamics”, “Model”, “Retreat”, “Glacier mass balance”, “Sea level”, “Radar”, “Volume change”, “Surface velocity”, “Glacier mapping”, “Hazard”, and other keywords. (4) The current research frontiers include water storage change, artificial intelligence, High Mountain Asia (HMA), photogrammetry, debris cover, geodetic method, area change, glacier volume, classification, satellite gravimetry, grounding line retreat, risk assessment, lake outburst flood, glacier elevation change, digital elevation model, geodetic mass balance, (DEM) generation, etc. According to the results of the visual analysis of the literature, we introduced the three commonly used methods of glacier mass balance based on remote sensing observation and summarized the research status and shortcomings of different methods in glacier mass balance. We considered that the future research trend is to improve the spatial and temporal resolution of data and combine a variety of methods and data to achieve high precision and long-term monitoring of glacier mass changes and improve the consistency of results. This research summarizes the study of glacier mass balance using remote sensing, which will provide valuable information for future research across this field.

Half-a-century (1971–2020) of glacier shrinkage and climatic variability in the Bhaga basin, western Himalaya

Glacier shrinkage is a globally occurring phenomena. High-resolution change detection based on frequent mapping and monitoring of high-altitude glaciers is necessary to precisely evaluate future water availability and to understand glacier evolution under different climatic scenarios in the Hindukush-Karakoram-Himalayan (HKH) region. This also holds true for the Bhaga basin of the western Himalaya. This study investigates glacier and glacier lake changes in the Bhaga basin, over the last five decades based on satellite imagery including Corona KH4 (1971), Landsat 7 Enhanced Thematic Mapper Plus (ETM+; 2000), Linear Imaging Self-Scanning Sensor (LISS IV; 2013), and Sentinel 2 (2020). Regional temperature and precipitation trends were evaluated from gridded climatic datasets (1900–2020). In the Bhaga basin 306 glaciers (>0.2 km2) were mapped with a total area of 360.3 ± 4.0 km2, of which 55.7 ± 0.6 km2 was covered with debris in 2013. The total glacier covered area decreased by ∼8.2 ± 1.5 % (0.16 ± 0.03 % yr−1) during the entire observation period 1971–2020, with noticeable heterogeneity between tributary watersheds. In the past two decades (2000–2020), the deglaciation rate has increased significantly (0.25 % yr−1) compared to the previous decades (1971–2000; 0.12 % yr−1). Glacier lake area increased by 0.6 ± 0.1 km2 (0.012 km2 yr−1) between 1971 and 2020. The NCEP/NCAR climatic data reveals an increase of 0.63°C in temperature and a decrease of 6.39 mm in precipitation for the period 1948–2018. In comparison, APHRODITE data shows an increasing trend in temperature of 1.14°C between 1961 and 2015 and decreasing trend in precipitation of 31 mm between 1951 and 2007. Both NCEP/NCAR and APHRODITE data reveal significant temperature increase and precipitation decrease since the 1990s, which have probably augmented ice loss in the Bhaga basin during the early 21st century.

Boundary Aware U-Net for Glacier Segmentation

Large-scale study of glaciers improves our understanding of global glacier change and is imperative for monitoring the ecological environment, preventing disasters, and studying the effects of global climate change. Glaciers in the Hindu Kush Himalaya (HKH) are particularly interesting as the HKH is one of the world's most sensitive regions for climate change. In this work, we: (1) propose a modified version of the U-Net for large-scale, spatially non-overlapping, clean glacial ice, and debris-covered glacial ice segmentation; (2) introduce a novel self-learning boundary-aware loss to improve debris-covered glacial ice segmentation performance; and (3) propose a feature-wise saliency score to understand the contribution of each feature in the multispectral Landsat 7 imagery for glacier mapping. Our results show that the debris-covered glacial ice segmentation model trained using self-learning boundary-aware loss outperformed the model trained using dice loss. Furthermore, we conclude that red, shortwave infrared, and near-infrared bands have the highest contribution toward debris-covered glacial ice segmentation from Landsat 7 images.

Mapping of Glaciers on Horseshoe Island, Antarctic Peninsula, with Deep Learning Based on High-Resolution Orthophoto

Antarctica plays a key role in the hydrological cycle of the Earth’s climate system, with an ice sheet that is the largest block of ice that reserves Earth’s 90% of total ice volume and 70% of fresh water. Furthermore, the sustainability of the region is an important concern due to the challenges posed by melting glaciers that preserve the Earth’s heat balance by interacting with the Southern Ocean. Therefore, the monitoring of glaciers based on advanced deep learning approaches offers vital outcomes that are of great importance in revealing the effects of global warming. In this study, recent deep learning approaches were investigated in terms of their accuracy for the segmentation of glacier landforms in the Antarctic Peninsula. For this purpose, high-resolution orthophotos were generated based on UAV photogrammetry within the Sixth Turkish Antarctic Expedition in 2022. Segformer, DeepLabv3+ and K-Net deep learning methods were comparatively analyzed in terms of their accuracy. The results showed that K-Net provided efficient results with 99.62% accuracy, 99.58% intersection over union, 99.82% precision, 99.76% recall and 99.79% F1-score. Visual inspections also revealed that K-Net was able to preserve the fine details around the edges of the glaciers. Our proposed deep-learning-based method provides an accurate and sustainable solution for automatic glacier segmentation and monitoring.