ROMFORSK-Program for romforskning

Sea ice is one of the most pronounced climate indicators in the Arctic Ocean and data on sea ice properties is therefore important in climate research. This project focused on providing methods for high-resolution sea ice drift retrieval and corresponding data sets using satellite Synthetic Aperture Radar (SAR) imagery. Ice drift causes formation of ridges and leads in the icepack, displacement of ice masses between regions and export of ice out of the Arctic Ocean. By utilizing repeated SAR images over sea ice areas, ice motion can be calculated with higher resolution than from other satellites. With Sentinel-1 A and B satellites the SAR coverage and repeat frequency is significantly improved, allowing ice drift to be calculated daily over large parts of the Arctic Ocean. To derive high resolution sea ice motion from consecutive SAR images, a paper introducing an open source feature-tracking algorithm for Sentinel-1 data was published in the framework of this project. The main advantages of the considered feature-tracking approach are the computational efficiency and the independence of the vectors in terms of position, lengths, direction and rotation. Traditional Pattern-matching on the other side allows better control over vector positioning and resolution. To combine the advantages from feature-tacking and patter-matching in the most meaningful way, a combined algorithm was published as part of the project. A feature-tracking procedure, based on the algorithm from the first publication, provides an unevenly distributed vector field. Outliers in this vector field are excluded in a second step. Based on the filtered feature-tracking vectors, drift and rotation on the entire SAR scene are estimated. This initial drift field limits the search area for a consecutive pattern-matching algorithm that provides small to medium scale adjustments. During a field experiment with KV Svalbard in Fram Strait in July 2016, three GPS transponder units were deployed on the sea ice to measure detailed ice drift with high temporal resolution, i.e. position recording intervals between 5-30 minutes. The data along with corresponding python scripts, plots and satellite images following the position of the trackers has been submitted for publication. These efforts shall help to gain a better understanding of short-term drift variability and how displacements from subsequent satellite images relate to instantaneous sea ice velocity. The data can also be used for validation of the ice drift algorithms in this region. More than 1500 Sentinel-1 SAR images, acquired over Fram Strait between 2014 and 2016, were used for sea ice drift retrieval using the open-source feature tracking algorithm from the first publication of this project. Based on this extensive SAR image data set, more than 5000 image pairs in HH and HV polarisation were used to calculate monthly mean sea ice velocities at 79N. This dataset, also submitted for publication along with python scripts and plots, shall serve for a better understand of the sea ice export through Fram Strait and illustrate the capabilities of a computationally efficient feature-tracking algorithm to derive sea ice drift from large Sentinel-1 SAR image data sets. These efforts shall help to prepare for an operational high-resolution sea ice drift product based on Sentinel-1 SAR imagery. The users of the operational products will be the sea ice services at Norwegian Meteorological Institute and the Copernicus Marine Services. A long-term goal is to supplement the ice motion data with SAR sea ice classification result to produce maps of ice drift, ice deformation and classification. These maps will provide significantly more detailed sea ice information for many user groups compared to what exists today.

Sea-ice motion is an essential variable to observe from EO data, because it strongly influences the distribution of sea-ice on different spatial and temporal scales. Ice drift causes advection of ice from one region to another and export of ice from the Arctic Ocean to the sub-Arctic seas. The proposal will exploit Synthetic Aperture Radar (SAR) data from Sentinel-1and Radarsat-2 for sea ice research and prepare for operational use of Sentinel-1 data. The main objective of IceMotion is to develop new SAR-derived sea ice motion data with high resolution (ca. 1 km) to observe and model sea ice dynamics, in particular ice deformation. The ice drift will be based on combined use of pattern tracking and Doppler shift data. Such high-resolution ice drift data are not yet provided and will be important for observation of ice divergence/convergence, leads, ridges and other regional and local ice processes. A long-term goal is to combine sea ice drift, including deformation, with ice type classification from SAR data to create innovative, high-resolution ice maps. The results will be exploited towards user groups in polar regions such as sea-ice modeller, operational ice forecasting services, ice navigation, offshore operation, climate research and environmental monitoring including tracking of pollution embedded in sea ice. The proposal will develop a sustainable processing system for ice drift retrieval from Sentinel-1 and Radarsat-2 data, to be operated under Copernicus marine services.