Mapping Radar Glacier Zones and Dry Snow Line in the Antarctic Peninsula Using Sentinel-1 Images

Short-term glacier area changes, glacier geometry dependence, and regional climatic variations forcing, King George Island, Antarctica

This study investigates the transient snowline (TSL) altitude for summer 2020, as well as glacial area loss in King George Island Icefields since 1988 using Sentinel-1 and 2 and Landsat Thematic Mapper (TM) imagery. Trends and anomalies in atmospheric temperature, U-wind, and V-wind were examined using ERA5 solutions. Results show the wet-snow zone corresponds to values of ≤ -13dB, and 44.3% of the glacial area is located above the TSL (≥ 300 m). Glacial area for 2020 is 999.95 km², and losses in the period represent 104.9 km² (error <1%) - a retreat of 3.17 km² / year. Glaciers in Keller Peninsula and Bellingshausen Dome lost the most area (28% and 17%, respectively) and did not have a TSL in 2020; followed by Warszawa (15%), Kraków (13%), and Eastern (10%), where the TSL was verified. Percentage area loss values increased with decreases in dimensions, area above TSL, and maximum elevation. Calving glaciers with ice-flow toward deeper and steeper submarine sectors (Bransfield Strait) exhibited greater glacier variations. The trend in warming atmospheric temperature was greater in the Bransfield Strait than in the Drake Passage. TSL and retreat difference between glaciers were influenced by climatic and ocean input, as well as multiple environmental factors.

Classification of glacier with supervised approaches using PolSAR data

Glacier comprises distinct features (snow, ice, and debris cover) and their identification and classification using satellite imagery is still a challenging task. Classification of different glacier features (zones) using remote sensing data is useful for numerous environmental and societal applications. The purpose of this study is to develop the fully polarimetric SAR (PolSAR) deep neural networks classification approach for the extraction of different features of the alpine glaciers. The developed approach was tested and classification results were compared with the support vector machines-based classification over the part of two glaciers: Siachen glacier and Bara Shigri glacier. The overall accuracy (OA) of GF-DNN classification is relatively high (91.17% for Siachen and 89% for Bara Shigri) with a good kappa coefficient (0.88 for Siachen and 0.85 for Bara Shigri) as compared to SVM for both the selected glaciers. An improvement of more than 10% is achieved in the OA of GF-DNN classification as compared to SVM for both the glaciers. The obtained classified results and accuracy demonstrates the potential of deep neural networks-based glacier features classification approach for glaciated terrain features.

On the Detection of Snow Cover Changes over the Australian Snowy Mountains Using a Dynamic OBIA Approach

This study detected the spatial changes in Snow Cover Area (SCA) over the Snowy Mountains in New South Wales, Australia. We applied a combination of Object-Based Image Analysis (OBIA) algorithms by segmentation, classification, and thresholding rules to extract the snow, water, vegetation, and non-vegetation land covers. For validation, the Maximum Snow Depths (MSDs) were collected at three local snow observation sites (namely Three Mile Dam, Spencer Creek, and Deep Creek) from 1984 to 2020. Multiple Landsat 5, 7, and 8 imageries extracted daily MSDs. The process was followed by applying an Estimation Scale Parameter (ESP) tool to build the local variance (LV) of object heterogeneity for each satellite scene. By matching the required segmentation parameters, the optimal separation step of the image objects was weighted for each of the image bands and the Digital Elevation Model (DEM). In the classification stage, a few land cover classes were initially assigned, and three different indices—Normalized Differential Vegetation Index (NDVI), Surface Water Index (SWI), and a Normalized Differential Snow Index (NDSI)—were created. These indices were used to adjust a few classification thresholds and ruleset functions. The resulting MSDs in all snow observation sites proves noticeable reduction trends during the study period. The SCA classified maps, with an overall accuracy of nearly 0.96, reveal non-significant trends, although with considerable fluctuations over the past 37 years. The variations concentrate in the north and south-east directions, to some extent with a similar pattern each year. Although the long-term changes in SCA are not significant, since 2006, the pattern of maximum values has decreased, with fewer fluctuations in wet and dry episodes. A preliminary analysis of climate drivers’ influences on MSD and SCA variability has also been performed. A dynamic indexing OBIA indicated that continuous processing of satellite images is an effective method of obtaining accurate spatial–temporal SCA information, which is critical for managing water resources and other geo-environmental investigations.

Land Cover-Specific Local Incidence Angle Correction: A Method for Time-Series Analysis of Forest Ecosystems

This study deals with a local incidence angle correction method, i.e., the land cover-specific local incidence angle correction (LC-SLIAC), based on the linear relationship between the backscatter values and the local incidence angle (LIA) for a given land cover type in the monitored area. Using the combination of CORINE Land Cover and Hansen et al.’s Global Forest Change databases, a wide range of different LIAs for a specific forest type can be generated for each scene. The algorithm was developed and tested in the cloud-based platform Google Earth Engine (GEE) using Sentinel-1 open access data, Shuttle Radar Topography Mission (SRTM) digital elevation model, and CORINE Land Cover and Hansen et al.’s Global Forest Change databases. The developed method was created primarily for time-series analyses of forests in mountainous areas. LC-SLIAC was tested in 16 study areas over several protected areas in Central Europe. The results after correction by LC-SLIAC showed a reduction of variance and range of backscatter values. Statistically significant reduction in variance (of more than 40%) was achieved in areas with LIA range >50° and LIA interquartile range (IQR) >12°, while in areas with low LIA range and LIA IQR, the decrease in variance was very low and statistically not significant. Six case studies with different LIA ranges were further analyzed in pre- and post-correction time series. Time-series after the correction showed a reduced fluctuation of backscatter values caused by different LIAs in each acquisition path. This reduction was statistically significant (with up to 95% reduction of variance) in areas with a difference in LIA greater than or equal to 27°. LC-SLIAC is freely available on GitHub and GEE, making the method accessible to the wide remote sensing community.

Delineation of Radar Glacier Zones in the Antarctic Peninsula Using Polarimetric SAR

Climate change is a cause of the expansion of snowmelt phenomena in the Antarctic, and shifts in position of wet and dry snow lines have been considered as good indicators of climate changes. The impacts of climate change are observable by the delineation of significant position change of glacier zones. The principal limitation of current glacier zone classification methods by synthetic aperture radar (SAR) image is that it is difficult to discriminate dry-snow and wet-snow zones using only single-polarimetric radar backscattering intensity. This study tried to solve the problem using polarimetric SAR (PolSAR). Analysis indicates that polarimetric decomposition elements could be efficient characteristics to delineate radar glacier zones by recognition of principal backscatter patterns. Further, two radar glacier zone classification processes for polarimetric SAR are proposed: a supervised support vector machine (SVM) classification process and a simple decision-tree classification method. These methods enable reliable delineation of radar glacier zones in the Antarctic Peninsula. Polarimetric SAR, which provides more information about the scattering processes and target structure, proves to be an efficient tool for delineating radar glacier zones and snowmelt detection.

Freezing and thawing of lakes on the Nelson and King George Islands, Antarctic, using Sentinel 1A synthetic aperture radar images

This article aims to analyze the dynamics of freezing and thawing of Antarctic lakes located in ice-free areas on Nelson Island and Fildes Peninsula, where response to changes in air temperature and precipitation rates occur rapidly, during the period from July 2016 to December 2018. In these places, which are difficult to access, remote sensing is an important alternative, especially considering the use of active remote sensors such as the Synthetic Aperture Radar (SAR), which has less restriction regarding the presence of clouds over the study area. Three backscatter thresholds were defined (σ) for the identification of the physical state of the water of the lakes of the study region, applied in Sentinel 1A SAR (S1A) images under Horizontal Horizontal (HH) polarization and Interferometric Wide (IW) imaging mode. These images, along with the air temperature data obtained by the Interim Re-Analysis (ERA-Interim) atmospheric reanalysis model, provided the evidence for the interpretation of the freezing and thawing periods of the lakes. The thresholds applied for the definition of the physical state of the lake water were greater than - 14 dB for frozen water, between - 14 and - 17 dB for the surface, with up to 60% of their frozen area, and less than - 17 dB for open water. The temporal analysis revealed that the lakes start to thaw in October, become completely thawed in February, and freeze again in March. Nevertheless, it can be said that the S1A satellite allows a satisfactory identification of the liquid and solid phases of the water in the lakes of the study region.

Satellite Remote Sensing of the Greenland Ice Sheet Ablation Zone: A Review

The Greenland Ice Sheet is now the largest land ice contributor to global sea level rise, largely driven by increased surface meltwater runoff from the ablation zone, i.e., areas of the ice sheet where annual mass losses exceed gains. This small but critically important area of the ice sheet has expanded in size by ~50% since the early 1960s, and satellite remote sensing is a powerful tool for monitoring the physical processes that influence its surface mass balance. This review synthesizes key remote sensing methods and scientific findings from satellite remote sensing of the Greenland Ice Sheet ablation zone, covering progress in (1) radar altimetry, (2) laser (lidar) altimetry, (3) gravimetry, (4) multispectral optical imagery, and (5) microwave and thermal imagery. Physical characteristics and quantities examined include surface elevation change, gravimetric mass balance, reflectance, albedo, and mapping of surface melt extent and glaciological facies and zones. The review concludes that future progress will benefit most from methods that combine multi-sensor, multi-wavelength, and cross-platform datasets designed to discriminate the widely varying surface processes in the ablation zone. Specific examples include fusing laser altimetry, radar altimetry, and optical stereophotogrammetry to enhance spatial measurement density, cross-validate surface elevation change, and diagnose radar elevation bias; employing dual-frequency radar, microwave scatterometry, or combining radar and laser altimetry to map seasonal snow depth; fusing optical imagery, radar imagery, and microwave scatterometry to discriminate between snow, liquid water, refrozen meltwater, and bare ice near the equilibrium line altitude; combining optical reflectance with laser altimetry to map supraglacial lake, stream, and crevasse bathymetry; and monitoring the inland migration of snowlines, surface melt extent, and supraglacial hydrologic features.

A Combination of PROBA-V/MODIS-based Products with Sentinel-1 SAR Data for Detecting Wet and Dry Snow Cover in Mountainous Areas

In the present study, we explore the value of employing both vegetation indexes as well as land surface temperature derived from Project for On-Board Autonomy – Vegetation (PROBA-V) and Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, respectively, to support the detection of total (wet + dry) snow cover extent (SCE) based on a simple tuning machine learning approach and provide reliability maps for further analysis. We utilize Sentinel-1-based synthetic aperture radar (SAR) observations, including backscatter coefficient, interferometric coherence, and polarimetric parameters, and four topographical factors as well as vegetation and temperature information to detect the total SCE with a land cover-dependent random forest-based approach. Our results show that the overall accuracy and F-measure are over 90% with an ’Area Under the receiver operating characteristic Curve (ROC)’ (AUC) score of approximately 80% over five study areas located in different mountain ranges, continents, and hemispheres. These accuracies are also confirmed by a comprehensive validation approach with different data sources, attesting the robustness and global transferability. Additionally, based on the reliability maps, we find an inversely proportional relationship between classification reliability and vegetation density. In conclusion, comparing to a previous study only utilizing SAR-based observations, the method proposed in the present study provides a complementary approach to achieve a higher total SCE mapping accuracy while maintaining global applicability with reliable accuracy and corresponding uncertainty information.

Remote Sensing of Snow Cover Using Spaceborne SAR: A Review

The importance of snow cover extent (SCE) has been proven to strongly link with various natural phenomenon and human activities; consequently, monitoring snow cover is one the most critical topics in studying and understanding the cryosphere. As snow cover can vary significantly within short time spans and often extends over vast areas, spaceborne remote sensing constitutes an efficient observation technique to track it continuously. However, as optical imagery is limited by cloud cover and polar darkness, synthetic aperture radar (SAR) attracted more attention for its ability to sense day-and-night under any cloud and weather condition. In addition to widely applied backscattering-based method, thanks to the advancements of spaceborne SAR sensors and image processing techniques, many new approaches based on interferometric SAR (InSAR) and polarimetric SAR (PolSAR) have been developed since the launch of ERS-1 in 1991 to monitor snow cover under both dry and wet snow conditions. Critical auxiliary data including DEM, land cover information, and local meteorological data have also been explored to aid the snow cover analysis. This review presents an overview of existing studies and discusses the advantages, constraints, and trajectories of the current developments.

Wet and Dry Snow Detection Using Sentinel-1 SAR Data for Mountainous Areas with a Machine Learning Technique

Traditional studies on mapping wet snow cover extent (SCE) often feature limitations, especially in vegetated and mountainous areas. The aim of this study is to propose a new total and wet SCE mapping strategy based on freely accessible spaceborne synthetic aperture radar (SAR) data. The approach is transferable on a global scale as well as for different land cover types (including densely vegetated forest and agricultural regions), and is based on the use of backscattering coefficient, interferometric SAR coherence, and polarimetric parameters. Furthermore, four topographical factors were included in the simple tuning of random forest-based land cover type-dependent classification strategy. Results showed the classification accuracy was above 0.75, with an F-measure higher than 0.70, in all five selected regions of interest located around globally distributed mountain ranges. Whilst excluding forest-type land cover classes, the accuracy and F-measure increases to 0.80 and 0.75. In cross-location model set, the accuracy can also be maintained at 0.80 with non-forest accuracy up to 0.85. It has been found that the elevation and polarimetric parameters are the most critical factors, and that the quality of land cover information would also affect the subsequent mapping reliability. In conclusion, through comprehensive validation using optical satellite and in-situ data, our land cover-dependent total SCE mapping approach has been confirmed to be robustly applicable, and the holistic SCE map for different months were eventually derived.

Multi-Feature Classification of Multi-Sensor Satellite Imagery Based on Dual-Polarimetric Sentinel-1A, Landsat-8 OLI, and Hyperion Images for Urban Land-Cover Classification

This paper focuses on evaluating the ability and contribution of using backscatter intensity, texture, coherence, and color features extracted from Sentinel-1A data for urban land cover classification and comparing different multi-sensor land cover mapping methods to improve classification accuracy. Both Landsat-8 OLI and Hyperion images were also acquired, in combination with Sentinel-1A data, to explore the potential of different multi-sensor urban land cover mapping methods to improve classification accuracy. The classification was performed using a random forest (RF) method. The results showed that the optimal window size of the combination of all texture features was 9 × 9, and the optimal window size was different for each individual texture feature. For the four different feature types, the texture features contributed the most to the classification, followed by the coherence and backscatter intensity features; and the color features had the least impact on the urban land cover classification. Satisfactory classification results can be obtained using only the combination of texture and coherence features, with an overall accuracy up to 91.55% and a kappa coefficient up to 0.8935, respectively. Among all combinations of Sentinel-1A-derived features, the combination of the four features had the best classification result. Multi-sensor urban land cover mapping obtained higher classification accuracy. The combination of Sentinel-1A and Hyperion data achieved higher classification accuracy compared to the combination of Sentinel-1A and Landsat-8 OLI images, with an overall accuracy of up to 99.12% and a kappa coefficient up to 0.9889. When Sentinel-1A data was added to Hyperion images, the overall accuracy and kappa coefficient were increased by 4.01% and 0.0519, respectively.