A novel method for detecting lake ice cover using optical satellite data

• Published in: International journal of applied earth observation and geoinformation 2021-12, Vol.104, p.102566, Article 102566
• Main Authors: Heinilä, Kirsikka, Mattila, Olli-Pekka, Metsämäki, Sari, Väkevä, Sakari, Luojus, Kari, Schwaizer, Gabriele, Koponen, Sampsa
• Format: Article
• Language: English
• Subjects: Lake ice; MSI; Optical; Reflectance; Sentinel; SLSTR
• ISSN: 1569-8432; 1872-826X
• DOI: 10.1016/j.jag.2021.102566

•The method uses gaussian mixture model (GMM) applied to optical data.•Unlike many other implementations, the method uses an integrated cloud classifier.•Even complex cases like fragmented ice and turbid water are successfully handled.•This computationally efficient method can be used for near real-time production.•∼13,000 lakes in Northern Hemisphere are covered with the method applying S3 SLSTR. Seasonal lake ice is sensitive to temperature fluctuations and long-term temperature trends. It is therefore a good indicator of climate warming, which will likely have dramatic impacts on lake ice phenology in northern latitudes. Beside the climate change aspect, lake ice data are important regarding transport and safety issues. In addition, changes in ice cover affect the ecology of the lake and water quality. A new method for accessing lake ice extent (LIE) using optical satellite data was developed at the Finnish Environment Institute. The resulting ICEmod method is based on multidimensional Gaussian distributions calculated for training data using several reflectance and thermal bands and their-related indices. Three alternative classes are provided for a lake ice pixel: (i) open water, (ii) ice cover, and (iii) cloud. The main advantages of ICEmod are the utilization of several spectral bands, inclusion of simultaneous cloud detection, definition of statistical probability for each pixel, the simplicity of processing and easy transition between different satellite sensors. This novel method is applied in the provision of the Sentinel-3 Sea and Land Surface Temperature Radiometer (SLSTR) -based 0.005° Lake Ice Extent product (LIE-NH) for the Northern Hemisphere under the Global Land Service, a component of the Copernicus Land Service. The 0.005° resolution LIE products based on ICEmod applied to SLSTR (an extensive subset corresponding to LIE-NH) were validated against high-resolution Sentinel-2 Multispectral Instrument -images covering lakes in the different parts of the Northern Hemisphere, most of them representing both open water and ice pixels. The gained overall classification accuracy is 96%. The resulting omission error for ice is 6% and omission mainly occurs in lakes where the ice is very dark and fragmented. The ICEmod method is not always able to separate dark ice from open water if there is water on the top of the dark ice, while misclassifications do not occur if the water layer is on white ice. The commission error for ice is 2.5% and is mostl