Sea ice thickness retrieval using microwave satellite observations from SMAP and SMOS
12 April 2019
Sea Ice plays a central role in the Earth’s climate system and thus, knowledge about its properties and extent is important for climate modeling and prediction as well as for planning shipping routes in the polar regions. The thickness of sea ice determines the heat exchange between the ocean and the atmosphere as well as the resistance against the deforming forces of wind and ocean currents. Already a thin layer of sea ice provides a surface for snow to deposit. Snow accumulation even further reduces the heat exchange and increases the albedo. The amount of thicker multiyear ice has been strongly decreasing during the last decades, but the winter sea ice maximum only little. Thus, today the area of thinner first-year ice makes up a considerably larger part of the sea ice covered region. Daily observations of the ice cover and its thickness are important to track these rapid changes and improve our understanding of the Arctic climate system.
One way of measuring sea ice thickness is using passive microwave satellite observations. More specifically, in the L-band (1.4 GHz) part of the electromagnetic sea ice emissions do not come only from the surface, but also from deeper layers underneath. For ice thinner than about 50 cm, therefore, a relation of emitted radiation to ice thickness can be exploited. Another advantage of L-band radiometry is that the atmosphere (clouds, precipitation, etc.) has small influence on the surface emission when traveling through the atmosphere towards the sensor and, as at all microwave frequencies, observations at L-band can be used during the long polar night when no light is available. With the launch of the new Soil Moisture Active Passive (SMAP) L-band radiometer in 2015, a previously developed retrieval algorithm for thin sea ice thickness (Huntemann et al., 2014), which uses data from the Soil Moisture Ocean Salinity (SMOS) sensor, has been adapted to use brightness temperature (TB) data from the new sensor (Pațilea et al., 2019).
Figure 1 represents an example of one day of data from both sensors used for ice thickness retrieval. At that time of year sea ice freeze-up has started and large parts of the marginal seas (e.g. Laptev and Kara Seas) of the Arctic Ocean are covered with thin sea ice. We can also observe that in some regions like the Laptev Sea, the sea ice along the shores already got thicker than 50 cm while in the central Laptev Sea still areas of thin sea ice prevail. There is an area around the North Pole, where there is no usable satellite data available. That area, however, is covered by thicker sea ice almost year-round. Thus, the complete L-band data timeseries allows to monitor the development of thin ice areas in the Arctic Ocean since the start of SMOS in 2010.
Figure 2 shows an example of a timeseries of L-band retrieved ice thickness data in the Laptev Sea. The start of each year is taken as 15th of September which is approximately around the Arctic sea ice minimum. We can see that by 15th to 20th of October most of the Laptev Sea is covered with sea ice (Figure 2 middle plot). By that time also the mean ice thickness increases and reaches the maximum ice thickness of 50 cm that can be retrieved only about one month later by 15 November (Figure 2 bottom plot). In the top plot we can observe that there is variability between the years, with 2011 having a late freeze-up and containing dominantly thin sea ice until the start of December, while in 2016, thick sea ice becomes dominant approximately three weeks before.
A method to quantify the uncertainty of the new data set has been developed based on the sensitivity of the retrieval to changes in brightness temperature and is shown in Figure 3. The uncertainty increases with ice thickness. For ice thickness of about 50 cm the average uncertainty reaches 30 cm due to high sensitivity to the change in brightness temperatures combined with uncertainty of the sea ice concentration that will impact it. Thus, here we only present ice thickness up to 50 cm. Above 50 cm ice thickness the measured brightness temperatures get close to saturation. Here the impact of sea ice concentration on the ice thickness uncertainty gets larger. For an ice thickness much higher than 50 cm combined with only slightly lower than 100% sea ice concentration the retrieved ice thickness can decrease below 50 cm. An additional source of uncertainty is the ice thickness change within one day. The SMAP/SMOS ice thickness dataset is a daily average of all measurements. However, for thin sea ice at low temperatures the ice thickness within one day can vary substantially. This effect is largest for thin ice and decreases with ice thickness.
In summary, here we present a new satellite dataset for thin sea ice thickness up to 50 cm that combines microwave radiometer measurements from the two sensors SMAP and SMOS. The combined dataset is available since 2015 and the SMOS only since 2010.
SMOS data on seaiceportal.de you can find here.
Cătălin Pațilea, University of Bremen, Institute of Environmental Physics
Georg Heygster, University of Bremen, Institute of Environmental Physics
Marcus Huntemann, University of Bremen, Institute of Environmental Physics
Gunnar Spreen, University of Bremen, Institute of Environmental Physics
Huntemann, M., Heygster, G., Kaleschke, L., Krumpen, T., Mäkynen, M., and Drusch, M.: Empirical sea ice thickness retrieval during the freeze-up period from SMOS high incident angle observations, The Cryosphere, 8, 439-451, doi: 10.5194/tc-8-439-2014, 2014.
Pațilea, C., Heygster, G., Huntemann, M., and Spreen, G.: Combined SMAP/SMOS Thin Sea Ice Thickness Retrieval, The Cryosphere, 13, 675–691, doi: 10.5194/tc-13-675-2019, 2019.