Advancing microwave radar retrievals of snow depth on sea ice: toward full characterisation of the snow and sea-ice layers

Abstract

Snow is a key factor in the sea-ice and Earth's climate systems that modifies the physical, climatic, and biogeochemical processes taking place. One of its most important impacts is in regulating sea-ice growth and melt. Despite its importance, little is known about the spatial and temporal distribution of snow depth on sea ice on the regional to global scales. Snow is tightly coupled to the highly dynamic sea-ice and atmospheric conditions and it is, therefore, very heterogeneous and constantly evolving both in space and in time. As a spatially and temporally representative, global, year-round product of snow depth observations on sea ice does not exist to this date, applications often have to rely on climatological values that do not necessarily hold true in the rapidly warming global climate. The unknown properties directly translate into the uncertainty of the result. This dissertation takes on the ambitious goal of working toward full characterisation of the snow and sea-ice layers. To achieve that, the focus is on advancing microwave radar retrievals of snow depth on sea ice. Enhanced snow depth observations will enable improving other measurements of sea-ice related parameters, most importantly sea-ice thickness, and in joint analysis of coincident sea-ice measurements estimating sea-ice bulk density becomes possible. In the first step, field experiments with ground-based C and K band pulse radars are carried out to investigate microwave penetration into the snow cover. The results show the K band microwaves expectedly reflect from the snow surface while the C band microwaves penetrate closer to the snow–sea-ice interface potentially enabling dual-frequency snow depth retrieval in less than half of the studied cases and only on first-year ice. In the second step, radar measurements of snow depth on sea ice are upscaled by using an airborne radar in the western Arctic Ocean in 2017–2019. A high-sensitivity, ultra-wideband, frequency-modulated continuous-wave (FMCW) radar is integrated to the instrument configuration of the Alfred Wegener Institute's (AWI) IceBird sea-ice campaigns. Snow depth retrievals with a custom algorithm based on signal peakiness from the radar measurements at a low altitude of 200 ft show good consistency against high altitude measurements at 1500 ft, which are comparable to previous acquisitions. At the nominal low altitude of the IceBird surveys, the small, two-metre radar footprint increases the spatial resolution and reduces the effect of off-nadir targets. Validation against ground measurements reveal a sub-centimetre mean bias, which is below the sensor resolution. As the main result of this step, the AWI IceBird surveys are now capable of discriminating between the snow and sea-ice layers. In the third step, the full AWI IceBird sensor configuration, including airborne laser, radar, and electromagnetic induction sounding instruments, is exploited by collocating the coincident thickness and freeboard measurements and tracking the locations of air–snow, snow–sea-ice, and sea-ice-water interfaces for more than 3000 km along survey paths over different sea-ice types. Assuming values for snow and sea-water densities and that the sea-ice cover is in isostatic equilibrium, it is possible to derive sea-ice bulk density. The results show that the ice-type averaged densities for first-year and multi-year ice are higher than and do not differ as much as widely used values from previous studies. This highlights the demand of algorithms to adapt to changing sea-ice density in satellite altimetry retrievals of sea-ice thickness. Finally, a negative-exponential parametrisation of sea-ice bulk density is derived using sea-ice freeboard as the predictor variable for future applications. In conclusion, this dissertation takes important advancing steps in characterising the snow and sea-ice layers. Previously, the airborne AWI IceBird surveys carried out in late-winter were only able to measure the combined thickness of the snow and sea-ice layers but now, after successful integration of the FMCW radar and in combination with the airborne laser scanner measurements, it is possible to track the locations of all three interfaces bounding the snow–sea-ice system. Such airborne multi-instrument measurements of snow depth, sea-ice thickness, and freeboard are important data sets in their own right to complement the scarce observations of sea-ice related parameters in remote areas of the polar regions, but a joint analysis allows deriving further key parameters like sea-ice density. The results of this dissertation can be applied to improve retrievals of geophysical sea-ice parameters from the soon 30-year long satellite altimetry data record, which in turn will contribute to enhance monitoring the climate-sensitive sea-ice cover and modelling future projections of the changing global climate.