- 1Paris, France (mylizpat@gmail.com)
- 2Scott Polar Research Institute, University of Cambridge, Cambridge, England (wgr2@cam.ac.uk)
- 3Institute of Geophysics, Polish Academy of Sciences, Warsaw, Poland (zswirad@igf.edu.pl)
- 4Scott Polar Research Institute, University of Cambridge, Cambridge, England (ovt20@cam.ac.uk)
Understanding the distribution and variations of sea and glacier ice coverage is critically important for assessing both the impacts and drivers of climate change, particularly in the Arctic. Sea ice responds dynamically to both ocean and atmospheric motion, implying variability on very short timescales which is challenging to monitor. In this study we assess the ability of PlanetScope satellite imagery, offering both high spatial and temporal resolutions, to analyse temporal variability. The study area is Isbjørnhamna-Hansbukta area in north-western Hornsund, Svalbard - a fjord characterised by both in situ formed sea ice (fast ice and drift ice broken from the fast ice e.g. by waves), pack ice drifting into the fjord from south-west with Sørkapp Current that brings cold water masses from Barents Sea, and glacier ice from calving Hansbeen.
We selected cloud-free images over a 4.5 ✕ 4.9 km AOI, large enough to depict the spread of sea ice to ensure accuracy in the analysis. From ten images captured in 2023, we collected sample reflectance data for three categories: thin ice, thick ice, and water. Thin ice in the AOI is typically grey and grey-white sea ice (10-30 cm) as well as brash ice and growlers, while thick ice is often snow-covered young and first-year sea ice (>30cm) as well as bergy bits and icebergs. Using these data, we calculated normalised difference spectral indices for both 8-band and 4-band imagery. Coastal Blue-Green 1 and Blue-Red indices were determined to be the most effective for discriminating between the different categories, and optimum thresholds were identified. Applying these indices and thresholds in QGIS, we generated 233 maps covering the months of March to August for the years 2018 to 2023.
From the initial visual interpretation, the results showed credible classification of the images and revealed continuous seasonal patterns for all years of the study, with minimal ice coverage observed in March, May, and July through August, a peak in sea ice coverage in April, and a resurgence of thin ice in June. However, no observable multi-year trends could be identified from a preliminary analysis of the maps, other than a sharp decline in ice coverage in 2023. Quantitative analysis of the maps allows estimates of the sea and glacier ice extent within the AOI to be made.
This research enhances our understanding of seasonal and interannual sea and glacier ice distribution in the nearshore and coastal zone of Svalbard. These findings have the potential to inform future studies about sea ice distribution, with the PlanetScope Imagery maps to be made publicly available through the Svalbard Integrated Arctic Earth Observing System data portal at the end of the study. Future research will compare the relative advantages of PlanetScope and SAR imagery.
How to cite: Makhotina, E., Rees, G., Swirad, Z., and Tutubalina, O.: Mapping of sea and glacier ice distribution in 2018-2023 in the Hornsund fjord, Svalbard with PlanetScope imagery, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7127, https://doi.org/10.5194/egusphere-egu25-7127, 2025.