Logo 2016 Dragon 3 Final Results Symposium

Conference Agenda

Overview and details of the sessions and sub-session of this conference. Please select a date or session to show only sub-sessions at that day or location. Please select a single sub-session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
C2: ID.10501 Sea Ice Monitoring
Time:
Tuesday, 05/Jul/2016:
2:00pm - 3:00pm

Session Chair: Noel Gourmelen
Session Chair: Hui Lin
Workshop: Hydrology & Cryosphere
Location: Building 7-220#, School of Resources and Environmental Science, Wuhan University

Show help for 'Increase or decrease the abstract text size'
Presentations
Oral presentation

Selected European Studies On Sea Ice Classification And Drift Retrieval As Basis For Collaborative Projects During Dragon 4

Wolfgang Dierking1, Xi Zhang2

1Alfred Wegener Institute for Polar and Marine Research (Bremerhaven, Germany), Arctic University of Norway (Tromso, Norway); 2First Institute of Oceanography (Qingdao, China);

Selected European Studies On Sea Ice Classification And Drift Retrieval As Basis For Collaborative Projects During Dragon 4

Wolfgang Dierking

Alfred Wegener Institute Helmholtz Center for Polar- and Marine Research, Bremerhaven, Germany & Arctic University of Norway, Tromsø, Norway

Xi Zhang

First Institute of Oceanography, State Oceanic Administration, Qingdao, China

Marko Mäkynen, Markku Similä, Juha Karvonen

Finnish Meteorological Institute, Helsinki, Finland

Rasmus Tonboe, Leif Toudal Pedersen

Danish Meteorological Institute

Thomas Hollands, Stefanie Linow, Jakob Griebel

Alfred Wegener Institute, Germany

Anders Berg, Leif Eriksson

Chalmers University of Technology, Sweden

Roberto Saldo

Technical University of Denmark

Stefan Muckenhuber

Nansen Environmental and Remote Sensing Center

Anthony Doulgeris, Johannes Lohse, Ane S. Fors, Thomas Kræmer, Torbjørn Eltoft

The Arctic University of Norway (Tromsø)

Mari-Ann Moen

Kongsberg Satellite Services, Tromsø, Norway

Nick Hughes, Penelope Wagner

Norwegian Ice Service, Tromsø, Norway

The key element in the Dragon program is the utilization of remote sensing technologies for geo- and bio-scientific research. Considering the vast extent of the Polar Regions and the difficulties to access those, it is immediately clear that in particular the use of data from Earth Observing (EO) satellites is essential for monitoring ice sheets, ice shelves and sea ice. Recent studies on sea ice focus both on aspects regarding the interaction mechanisms between ocean, ice and atmosphere and their implications for weather and climate, and on information retrieval about ice mechanics and ice conditions for supporting marine traffic and offshore operations.

During the Dragon-2 and 3 phases, researchers from the First Institute of Oceanography, Qingdao, China, the Finnish Meteorological Institute, the Danish Meteorological Institute, and the Alfred Wegener Institute in Germany have successfully established a close information exchange with respect to sea ice classification and ice parameter retrieval. Direct collaboration projects were focused on ice thickness retrieval using polarimetric SAR data from the Bohai Sea, and the possibility to estimate ice thickness in the Arctic based on compact satellite radar polarimetry. The PIs Ji, Zhang, and Dierking as well as the other members of the Dragon sea ice team have been involved in various external collaborative projects, which have been positively influencing their work in the Dragon program. The objective of this presentation is to give an overview about important European projects and studies that were carried out during the Dragon-3 phase and are of importance for future activities in the Dragon-4 program. Here, we focus on sea ice classification and drift retrieval using synthetic aperture radar (SAR).

The separation of different ice types is needed for marine operations (mainly thin, smooth ice separated from thicker compacted ice) and for scientific process studies regarding, for example, heat exchange between ocean and atmosphere through the ice, or exchange of momentum between ice, on the one hand, and atmosphere and ocean, on the other hand. The major step in sea ice classification is to sub-divide sea ice SAR images into distinct regions based on similarities of parameters derived from the radar signal(s), and relate those regions to existing sea ice classification schemes. A special case is the separation of open water and ice for the determination of ice concentration. In this context, different research groups study different statistical models to adequately describe the distribution of radar parameter values typical for single ice types. Besides multi-polarization imagery, acquired with polarimetric SAR, multi-frequency data sets (that require combined acquisitions from SAR systems on different satellites) are in the focus of recent investigations in European groups. In order to improve the reliability of the classification and the retrieval of ice parameters, more advanced models for describing the multivariate dependencies are needed. The ultimate goal of these investigations is to make the whole classification process automatic, reliable and robust. Classification during the melting season, which is hampered by moist or wet snow layers and ice surfaces, and by melt ponds on the ice, is an important topic as well. The direct validation of sea ice maps and parameter retrievals is difficult because of the logistic difficulties to obtain the needed complementary data. The European sea ice remote sensing group hence closely collaborate with field researchers or participate themselves in field cruises to the Polar Regions.

The retrieval of sea ice thickness from remote sensing data is one of the holy grails of polar research. European partners are developing and demonstrating the regional mapping of sea ice thickness based on multi-sensor satellite data (e. g. Sentinel-1 SAR combined with AMSR2) and parallel thickness simulations of a sea ice model (CMEMS Topaz). Thin ice (<30 cm) areas are detected using AMSR2 data and are excluded to reduce ice classification ambiguities in SAR images. Also areas with ice concentrations less than 70% are not considered for retrieval. For thicker ice areas SAR data are employed to modulate locally the TOPAZ ice thickness field. Another example is the development of new sea ice classification and thickness products from SAR and radar altimeter data (the latter from Cryosat-2 and Sentinel-3) in the EU H2020 project SPICES. The ice thickness retrieval works under cold wintertime conditions.

The sea ice drift retrieval is carried out in three different ways: (1) pattern matching. For this method, at least two consecutive SAR images are needed. From normalized cross-correlation or phase correlation, or combinations of both, the displacement of the ice between the timings of the two image acquisitions is calculated. (2) feature tracking. In this case, structures are identified in two consecutive images, and their displacement is determined. (3) Doppler shift analysis. Here, only one image is needed, e. g. the Sentnel-1 radial surface velocity product. Whereas the results of (1) and (2) represent the 2-D average drift vectors for time steps between a few hours and days, approach 3 is a snapshot of the instantaneous line-of-sight motion. Different European groups have been working on the different methods and focus recently on problems of rotational ice movements, discontinuities of the drift field, evaluation of the accuracy of drift retrievals, spatial scaling of drift and deformation, and on increasing the computational speed of the retrieval algorithms.

Dierking-Selected European Studies On Sea Ice Classification And Drift Retrieval As Basis_ppt_present.pdf

Oral presentation

Techniques for Sea Ice Parameter Extraction and Sea Ice Monitoring Using Multi-Sensor Satellite Data in the Bohai Sea—Final Report

Xi Zhang, Wolfgang Dierking

The First Institute of Oceanography, State Oceanic Administration, China, People's Republic of;

The Bohai Sea and its coastal regions are significant economic areas in China. Sea ice poses a great threat to coastal construction and manufacturing industry, leading to severe economic loss to China. The objectives of this project are to extend the techniques and methods obtained as part of the Dragon-2 programme (ID: 5290) to form automatic or semi-automatic ice characteristics extraction methods using multi-sensor satellite data, and further improve techniques for sea-ice monitoring based on optical and SAR images. The technology developed in the project can be used in the operational sea-ice monitoring of the Bohai Sea and other ice-covered areas.

Each remote sensing sensor has its own advantage and disadvantage. SAR is capable of sensing the dielectric properties and surface and volume structures of sea ice. Optical sensors measure the spectral characteristics of sea ice in the visible and near-infrared range. For this reason, fusion of cloudless optical and SAR images can enhance the differences between sea-ice types and provide complementary information for sea-ice classification. An image fusion approach of SAR and multispectral data was proposed. ENVISAT ASAR and CBERS multispectral data were used for the experiment. The results demonstrate that spectral and SAR texture characteristics were preserved though this method, and the fusion result is effective for sea-ice interpretation and classification.

Research on sea-ice thickness retrieval is developed using both SAR and hyperspectral data in this project. From field measurements in the Bohai Sea, the reflectances of different sea-ice thickness were obtained. Based on the in-situ data, we developed a sea-ice thickness retrieval model for hyperspectral data and tested the model by using airborne hyperspectral data to retrieve sea-ice thickness. The results show that ice thickness in the test sites varied from 2.0cm to 30.0cm.

Making use of new remote sensing instrument to monitoring sea ice is another object in this project. GOCI (Geostationary Ocean Color Imager) acquires one image per an hour (8 images in every day time), which is the first geostationary observation satellite of Korea. GOCI data consists of 8 spectral bands with a spatial resolution of about 500 m. Sea-ice detection method of the Bohai Sea could be carried out to extract sea-ice parameters using continues-time GOCI data, for example sea-ice types, area, concentration et al. And based on the advantage of GOCI continuous-time observation, we have proposed a new method to retrieval the sea-ice drifting and spreading. The proposed method can accomplish tracing sea-ice movement.

Finally, in this paper, the methods of calculating sea-ice-hazard risk will be proposed. Moreover, the spatial distribution characteristics and occurrence probability of the sea-ice-hazard risk in the Bohai Sea also will be studied. The methods of calculating sea-ice-hazard risk would be expressed through the different sea-ice-hazard indexes for different hazard-bearing objects. In this paper, there are two hazard-bearing objects. One is the marine transportation, and the other is the offshore construction. The indexes will be constructed by the sea-ice parameters, such as sea-ice concentration, thickness and drift velocity, etc. The calculating function of the sea-ice-hazard indexes will be deduced.

Zhang-Techniques for Sea Ice Parameter Extraction and Sea Ice Monitoring Using Multi-Sensor Satellite Data_Cn_version.pdf

Poster

Sea ice edge detection by polarization basis transformation using full-polarization SAR

Yi Zhang1, Jie Zhang2, Xi Zhang3

1North China Sea Marine Forecasting Center, State Oceanic Administration; 2The First Institute of Oceanography, State Oceanic Administration, China, People's Republic of; 3The First Institute of Oceanography, State Oceanic Administration, China, People's Republic of;

The key of monitoring and detection sea ice is extracting the location and extent of sea ice edge. Extracting sea ice edge exactly has important significance for evaluating sea ice condition and ensuring navigation and marine operation safety. In this paper, Sea ice edge detection method is developed by full-polarization SAR and polarization basis transformation method. By analysis, different polarization basis can be used to extract different sea ice characteristics. H-V basis can be used to extract dielectric property and circular basis can be used to extract surface roughness characteristic. Multiple information is merged, and present a sea ice edge extraction method, and the result is compared with optical image to verify the reliability of this method.

Zhang-Sea ice edge detection by polarization basis transformation using full-polarization SAR_Cn_version.pdf


 
Contact and Legal Notice · Contact Address:
Conference: 2016 Dragon 3 Final Results Symposium
Conference Software - ConfTool Pro 2.6.94
© 2001 - 2016 by H. Weinreich, Hamburg, Germany