@article{BaumhoerDietzKneiseletal.2019,
  author    = {Baumhoer, Celia A. and Dietz, Andreas J. and Kneisel, C. and Kuenzer, C.},
  title     = {Automated Extraction of Antarctic Glacier and Ice Shelf Fronts from Sentinel-1 Imagery Using Deep Learning},
  series = {Remote Sensing},
  volume    = {11},
  journal   = {Remote Sensing},
  number    = {21},
  issn      = {2072-4292},
  doi       = {10.3390/rs11212529},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-193150},
  pages     = {2529},
  year      = {2019},
  abstract  = {Sea level rise contribution from the Antarctic ice sheet is influenced by changes in glacier and ice shelf front position. Still, little is known about seasonal glacier and ice shelf front fluctuations as the manual delineation of calving fronts from remote sensing imagery is very time-consuming. The major challenge of automatic calving front extraction is the low contrast between floating glacier and ice shelf fronts and the surrounding sea ice. Additionally, in previous decades, remote sensing imagery over the often cloud-covered Antarctic coastline was limited. Nowadays, an abundance of Sentinel-1 imagery over the Antarctic coastline exists and could be used for tracking glacier and ice shelf front movement. To exploit the available Sentinel-1 data, we developed a processing chain allowing automatic extraction of the Antarctic coastline from Seninel-1 imagery and the creation of dense time series to assess calving front change. The core of the proposed workflow is a modified version of the deep learning architecture U-Net. This convolutional neural network (CNN) performs a semantic segmentation on dual-pol Sentinel-1 data and the Antarctic TanDEM-X digital elevation model (DEM). The proposed method is tested for four training and test areas along the Antarctic coastline. The automatically extracted fronts deviate on average 78 m in training and 108 m test areas. Spatial and temporal transferability is demonstrated on an automatically extracted 15-month time series along the Getz Ice Shelf. Between May 2017 and July 2018, the fronts along the Getz Ice Shelf show mostly an advancing tendency with the fastest moving front of DeVicq Glacier with 726 ± 20 m/yr.},
  language  = {en}
}
@article{DirscherlDietzKneiseletal.2020,
  author    = {Dirscherl, Mariel and Dietz, Andreas J. and Kneisel, Christof and Kuenzer, Claudia},
  title     = {Automated mapping of Antarctic supraglacial lakes using a Machine Learning approach},
  series = {Remote Sensing},
  volume    = {12},
  journal   = {Remote Sensing},
  number    = {7},
  issn      = {2072-4292},
  doi       = {10.3390/rs12071203},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-203735},
  year      = {2020},
  abstract  = {Supraglacial lakes can have considerable impact on ice sheet mass balance and global sea-level-rise through ice shelf fracturing and subsequent glacier speedup. In Antarctica, the distribution and temporal development of supraglacial lakes as well as their potential contribution to increased ice mass loss remains largely unknown, requiring a detailed mapping of the Antarctic surface hydrological network. In this study, we employ a Machine Learning algorithm trained on Sentinel-2 and auxiliary TanDEM-X topographic data for automated mapping of Antarctic supraglacial lakes. To ensure the spatio-temporal transferability of our method, a Random Forest was trained on 14 training regions and applied over eight spatially independent test regions distributed across the whole Antarctic continent. In addition, we employed our workflow for large-scale application over Amery Ice Shelf where we calculated interannual supraglacial lake dynamics between 2017 and 2020 at full ice shelf coverage. To validate our supraglacial lake detection algorithm, we randomly created point samples over our classification results and compared them to Sentinel-2 imagery. The point comparisons were evaluated using a confusion matrix for calculation of selected accuracy metrics. Our analysis revealed wide-spread supraglacial lake occurrence in all three Antarctic regions. For the first time, we identified supraglacial meltwater features on Abbott, Hull and Cosgrove Ice Shelves in West Antarctica as well as for the entire Amery Ice Shelf for years 2017-2020. Over Amery Ice Shelf, maximum lake extent varied strongly between the years with the 2019 melt season characterized by the largest areal coverage of supraglacial lakes (~763 km\(^2\)). The accuracy assessment over the test regions revealed an average Kappa coefficient of 0.86 where the largest value of Kappa reached 0.98 over George VI Ice Shelf. Future developments will involve the generation of circum-Antarctic supraglacial lake mapping products as well as their use for further methodological developments using Sentinel-1 SAR data in order to characterize intraannual supraglacial meltwater dynamics also during polar night and independent of meteorological conditions. In summary, the implementation of the Random Forest classifier enabled the development of the first automated mapping method applied to Sentinel-2 data distributed across all three Antarctic regions.},
  language  = {en}
}
@phdthesis{Baumhoer2021,
  author    = {Baumhoer, Celia Am{\´e}lie},
  title     = {Glacier Front Dynamics of Antarctica - Analysing Changes in Glacier and Ice Shelf Front Position based on SAR Time Series},
  doi       = {10.25972/OPUS-24581},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-245814},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2021},
  abstract  = {The Antarctic Ice Sheet stores ~91\% of the global ice volume which is equivalent to a sea-level rise of 58.3 meters. Recent disintegration events of ice shelves and retreating glaciers along the Antarctic Peninsula and West Antarctica indicate the current vulnerable state of the Antarctic Ice Sheet. Glacier tongues and ice shelves create a safety band around Antarctica with buttressing effects on ice discharge. Current decreases in glacier and ice shelf extent reduce the effective buttressing forces and increase ice discharge of grounded ice. The consequence is a higher contribution to sea-level rise from the Antarctic Ice Sheet. So far, it is unresolved which proportion of Antarctic glacier retreat can be attributed to climate change and which part to the natural cycle of growth and decay in the lifetime of a glacier. The quantitative assessment of the magnitude, spatial extent, distribution, and dynamics of circum-Antarctic glacier and ice shelf retreat is of utmost importance to monitor Antarctica's weakening safety band. In remote areas like Antarctica, earth observation provides optimal properties for large-scale mapping and monitoring of glaciers and ice shelves. Nowadays, the variety of available satellite sensors, technical advancements regarding spatial resolution and revisit times, as well as open satellite data archives create an ideal basis for monitoring calving front change. A systematic review conducted within this thesis revealed major gaps in the availability of glacier and ice shelf front position measurements despite the improved satellite data availability. The previously limited availability of satellite imagery and the time-consuming manual delineation of calving fronts did neither allow a circum-Antarctic assessment of glacier retreat nor the assessment of intra-annual changes in glacier front position. To advance the understanding of Antarctic glacier front change, this thesis presents a novel automated approach for calving front extraction and explores drivers of glacier retreat. A comprehensive review of existing methods for glacier front extraction ascertained the lack of a fully automatic approach for large-scale monitoring of Antarctic calving fronts using radar imagery. Similar backscatter characteristics of different ice types, seasonally changing backscatter values, multi-year sea ice, and m{\´e}lange made it challenging to implement an automated approach with traditional image processing techniques. Therefore, the present abundance of satellite data is best exploited by integrating recent developments in big data and artificial intelligence (AI) research to derive circum-Antarctic calving front dynamics. In the context of this thesis, the novel AI-based framework "AntarcticLINES" (Antarctic Glacier and Ice Shelf Front Time Series) was created which provides a fully automated processing chain for calving front extraction from Sentinel-1 imagery. Open access Sentinel-1 radar imagery is an ideal data source for monitoring current and future changes in the Antarctic coastline with revisit times of less than six days and all-weather imaging capabilities. The developed processing chain includes the pre-processing of dual-polarized Sentinel-1 imagery for machine learning applications. 38 Sentinel-1 scenes were used to train the deep learning architecture U-Net for image segmentation. The trained weights of the neural network can be used to segment Sentinel-1 scenes into land ice and ocean. Additional post-processing ensures even more accurate results by including morphological filtering before extracting the final coastline. A comprehensive accuracy assessment has proven the correct extraction of the coastline. On average, the automatically extracted coastline deviates by 2-3 pixels (93 m) from a manual delineation. This accuracy is in range with deviations between manually delineated coastlines from different experts. For the first time, the fully automated framework AntarcticLINES enabled the extraction of intra-annual glacier front fluctuations to assess seasonal variations in calving front change. Thereby, for example, an increased calving frequency of Pine Island Glacier and a beginning disintegration of Glenzer Glacier were revealed. Besides, the extraction of the entire Antarctic coastline for 2018 highlighted the large-scale applicability of the developed approach. Accurate results for entire Antarctica were derived except for the Western Antarctic Peninsula where training imagery was not sufficient and should be included in future studies. Furthermore, this dissertation presents an unprecedented record of circum-Antarctic calving front change over the last two decades. The newly extracted coastline for 2018 was compared to previous coastline products from 2009 and 1997. This revealed that the Antarctic Ice Sheet shrank 29,618±1193 km2 in extent between 1997-2008 and gained an area of 7,108±1029 km2 between 2009-2018. Glacier retreat concentrated along the Antarctic Peninsula and West Antarctica. The only East Antarctic coastal sector primarily experiencing calving front retreat was Wilkes Land in 2009-2018. Finally, potential drivers of circum-Antarctic glacier retreat were identified by combining data on glacier front change with changes in climate variables. It was found that strengthening westerlies, snowmelt, rising sea surface temperatures, and decreasing sea ice cover forced glacier retreat over the last two decades. Relative changes in mean air temperature could not be identified as a driver for glacier retreat and further investigations on extreme events in air temperature are necessary to assess the effect of atmospheric forcing on frontal retreat. The strengthening of all identified drivers was closely connected to positive phases of the Southern Annular Mode (SAM). With increasing greenhouse gases and ozone depletion, positive phases of SAM will occur more often and force glacier retreat even further in the future. Within this thesis, a comprehensive review on existing Antarctic glacier and ice shelf front studies was conducted revealing major gaps in Antarctic calving front records. Therefore, a fully automated processing chain for glacier and ice shelf front extraction was implemented to track circum-Antarctic calving front fluctuations on an intra-annual basis. The large-scale applicability was certified by presenting two decades of circum-Antarctic calving front change. In combination with climate variables, drivers of recent glacier retreat were identified. In the future, the presented framework AntarcticLINES will greatly contribute to the constant monitoring of the Antarctic coastline under the pressure of a changing climate.},
  language  = {en}
}
@article{DirscherlDietzKneiseletal.2021,
  author    = {Dirscherl, Mariel and Dietz, Andreas J. and Kneisel, Christof and Kuenzer, Claudia},
  title     = {A novel method for automated supraglacial lake mapping in Antarctica using Sentinel-1 SAR imagery and deep learning},
  series = {Remote Sensing},
  volume    = {13},
  journal   = {Remote Sensing},
  number    = {2},
  issn      = {2072-4292},
  doi       = {10.3390/rs13020197},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-222998},
  year      = {2021},
  abstract  = {Supraglacial meltwater accumulation on ice sheets can be a main driver for accelerated ice discharge, mass loss, and global sea-level-rise. With further increasing surface air temperatures, meltwater-induced hydrofracturing, basal sliding, or surface thinning will cumulate and most likely trigger unprecedented ice mass loss on the Greenland and Antarctic ice sheets. While the Greenland surface hydrological network as well as its impacts on ice dynamics and mass balance has been studied in much detail, Antarctic supraglacial lakes remain understudied with a circum-Antarctic record of their spatio-temporal development entirely lacking. This study provides the first automated supraglacial lake extent mapping method using Sentinel-1 synthetic aperture radar (SAR) imagery over Antarctica and complements the developed optical Sentinel-2 supraglacial lake detection algorithm presented in our companion paper. In detail, we propose the use of a modified U-Net for semantic segmentation of supraglacial lakes in single-polarized Sentinel-1 imagery. The convolutional neural network (CNN) is implemented with residual connections for optimized performance as well as an Atrous Spatial Pyramid Pooling (ASPP) module for multiscale feature extraction. The algorithm is trained on 21,200 Sentinel-1 image patches and evaluated in ten spatially or temporally independent test acquisitions. In addition, George VI Ice Shelf is analyzed for intra-annual lake dynamics throughout austral summer 2019/2020 and a decision-level fused Sentinel-1 and Sentinel-2 maximum lake extent mapping product is presented for January 2020 revealing a more complete supraglacial lake coverage (~770 km\(^2\)) than the individual single-sensor products. Classification results confirm the reliability of the proposed workflow with an average Kappa coefficient of 0.925 and a F\(_1\)-score of 93.0\% for the supraglacial water class across all test regions. Furthermore, the algorithm is applied in an additional test region covering supraglacial lakes on the Greenland ice sheet which further highlights the potential for spatio-temporal transferability. Future work involves the integration of more training data as well as intra-annual analyses of supraglacial lake occurrence across the whole continent and with focus on supraglacial lake development throughout a summer melt season and into Antarctic winter.},
  language  = {en}
}