Deep learning-based coastline detection algorithms have begun to outshine traditional statistical methods in recent years. However, they are usually trained only as single-purpose models to either segment land and water or delineate the coastline. In contrast to this, a human annotator will usually keep a mental map of both segmentation and delineation when performing manual coastline detection. To take into account this task duality, we, therefore, devise a new model to unite these two approaches in a deep learning model. By taking inspiration from the main building blocks of a semantic segmentation framework (UNet) and an edge detection framework (HED), both tasks are combined in a natural way. Training is made efficient by employing deep supervision on side predictions at multiple resolutions. Finally, a hierarchical attention mechanism is introduced to adaptively merge these multiscale predictions into the final model output. The advantages of this approach over other traditional and deep learning-based methods for coastline detection are demonstrated on a data set of Sentinel-1 imagery covering parts of the Antarctic coast, where coastline detection is notoriously difficult. An implementation of our method is available at https://github.com/khdlr/HED-UNet .

In a warming Arctic, permafrost-related disturbances, such as retrogressive thaw slumps (RTS), are becoming more abundant and dynamic, with serious implications for permafrost stability and bio-geochemical cycles on local to regional scales. Despite recent advances in the field of earth observation, many of these have remained undetected as RTS are highly dynamic, small, and scattered across the remote permafrost region. Here, we assessed the potential strengths and limitations of using deep learning for the automatic segmentation of RTS using PlanetScope satellite imagery, ArcticDEM and auxiliary datasets. We analyzed the transferability and potential for pan-Arctic upscaling and regional cross-validation, with independent training and validation regions, in six different thaw slump-affected regions in Canada and Russia. We further tested state-of-the-art model architectures (UNet, UNet++, DeepLabv3) and encoder networks to find optimal model configurations for potential upscaling to continental scales. The best deep learning models achieved mixed results from good to very good agreement in four of the six regions (maxIoU: 0.39 to 0.58; Lena River, Horton Delta, Herschel Island, Kolguev Island), while they failed in two regions (Banks Island, Tuktoyaktuk). Of the tested architectures, UNet++ performed the best. The large variance in regional performance highlights the requirement for a sufficient quantity, quality and spatial variability in the training data used for segmenting RTS across diverse permafrost landscapes, in varying environmental conditions. With our highly automated and configurable workflow, we see great potential for the transfer to active RTS clusters (e.g., Peel Plateau) and upscaling to much larger regions.

An accurate parameterization of glacier calving is essential for understanding glacier dynamics and constraining ice-sheet models. The increasing availability and quality of remote sensing imagery open the prospect of a continuous and precise mapping of relevant parameters, such as calving front locations. However, it also calls for automated and scalable analysis strategies. Deep neural networks provide powerful tools for processing large quantities of remote sensing data. In this contribution, we assess the benefit of diverse input data for calving front extraction. In particular, we focus on Landsat-8 imagery supplementing single-band inputs with multispectral data, topography, and textural information. We assess the benefit of these three datasets using a dropped-variable approach. The associated reference dataset comprises 728 manually delineated calving front positions of 23 Greenland and two Antarctic outlet glaciers from 2013 to 2021. Resulting feature importance emphasizes both the potential integrating additional input information as well as the significance of their thoughtful selection. We advocate utilizing multispectral features as their integration leads generally to more accurate predictions compared with conventional single-band inputs. This is especially prevalent for challenging ice mélange and illumination conditions. In contrast, the application of both textural and topographic inputs cannot be recommended without reservation, since they may lead to model overfitting. The results of this assessment are not only relevant for advancing automated calving front extraction but also for a wider range of glaciology-related land surface classification tasks using deep neural networks.

A major uncertainty in predicting the behaviour of marine-terminating glaciers is ice dynamics driven by non-linear calving front retreat, which is poorly understood and modelled. Using 124919 calving front positions for 149 marine-terminating glaciers in Svalbard from 1985 to 2023, generated with deep learning, we identify pervasive calving front retreats for non-surging glaciers over the past 38 years. We observe widespread seasonal cycles in calving front position for over half of the glaciers. At the seasonal timescale, peak retreat rates exhibit a several-month phase lag, with changes on the west coast occurring before those on the east coast, coincident with regional ocean warming. This spatial variability in seasonal patterns is linked to different timings of warm ocean water inflow from the West Spitsbergen Current, demonstrating the dominant role of ice-ocean interaction in seasonal front changes. The interannual variability of calving front retreat shows a strong sensitivity to both atmospheric and oceanic warming, with immediate responses to large air and ocean temperature anomalies in 2016 and 2019, likely driven by atmospheric blocking that can influence extreme temperature variability. With more frequent blocking occurring and continued regional warming, future calving front retreats will likely intensify, leading to more significant glacier mass loss.

The frontal position of an ice shelf is an important parameter for ice dynamic modelling, the computation of mass fluxes, mapping glacier area change, calculating iceberg production rates and the estimation of ice discharge to the ocean. Until now, continuous and up-to-date information on Antarctic calving front locations is scarce due to the time-consuming manual delineation of fronts and the previously limited amount of suitable earth observation data. Here, we present IceLines, a novel data set on Antarctic ice shelf front positions to assess calving front change at an unprecedented temporal and spatial resolution. More than 19,400 calving front positions were automatically extracted creating dense inter- and intra-annual time series of calving front change for the era of Sentinel-1 (2014-today). The calving front time series can be accessed via the EOC GeoService hosted by DLR and is updated on a monthly basis. For the first time, the presented IceLines data set provides the possibility to easily include calving front dynamics in scientific studies and modelling to improve our understanding about ice sheet dynamics.

Aerial scene recognition is a fundamental visual task and has attracted an increasing research interest in the last few years. Most of current researches mainly deploy efforts to categorize an aerial image into one scene-level label, while in real-world scenarios, there often exist multiple scenes in a single image. Therefore, in this paper, we propose to take a step forward to a more practical and challenging task, namely multi-scene recognition in single images. Moreover, we note that manually yielding annotations for such a task is extraordinarily time- and labor-consuming. To address this, we propose a prototype-based memory network to recognize multiple scenes in a single image by leveraging massive well-annotated single-scene images. The proposed network consists of three key components: 1) a prototype learning module, 2) a prototype-inhabiting external memory, and 3) a multi-head attention-based memory retrieval module. To be more specific, we first learn the prototype representation of each aerial scene from single-scene aerial image datasets and store it in an external memory. Afterwards, a multi-head attention-based memory retrieval module is devised to retrieve scene prototypes relevant to query multi-scene images for final predictions. Notably, only a limited number of annotated multi-scene images are needed in the training phase. To facilitate the progress of aerial scene recognition, we produce a new multi-scene aerial image (MAI) dataset. Experimental results on variant dataset configurations demonstrate the effectiveness of our network. Our dataset and codes are publicly available1.