Radiometrische Messungen im arktischen Ozean - Vergleich von Theorie und Experiment = Radiometric measurements in the Arctic Ocean

Arctic sea ice is going through a dramatic change in its extent and volume at an unprecedented rate. Sea-ice thickness (SIT) is a controlling geophysical variable that needs to be understood with greater accuracy. For the first time, a SIT-retrieval method that exclusively uses only airborne SIT data for training the empirical algorithm to retrieve SIT from Soil Moisture Ocean Salinity (SMOS) brightness temperature (TB) at different polarization is presented. A large amount of airborne SIT data has been used from various field campaigns in the Arctic conducted by different countries during 2011–15. The algorithm attempts to circumvent the issue related to discrimination between TB signatures of thin SIT versus low sea-ice concentration. The computed SIT has a rms error of 0.10 m, which seems reasonably good (as compared to the existing algorithms) for analysis at the used 25 km grid. This new SIT retrieval product is designed for direct operational application in ice prediction/climate models.

/pdf/on-the-retrieval-of-sea-ice-thickness-using-smos-2aady5v3mo.pdf

On the retrieval of sea-ice thickness using SMOS polarization differences

/pdf/lotus-preparing-land-and-ocean-take-up-from-sentinel-3-4e4cavpevu.pdf

LOTUS- Preparing Land and Ocean Take Up from Sentinel-3

In this study we have developed an empirical retrieval for thickness of young and first-year ice during the freeze up period for the L-band passive microwave radiometer Microwave Imaging Radiometer with Aperture Synthesis (MIRAS) on the Soil Moisture and Ocean Salinity (SMOS) satellite. The retrieval is based on intensity and polarization difference using the incidence angle range of 40◦ to 50◦ and is validated using data from airborne EM-Bird, Moderate-resolution Imaging Spectroradiometer (MODIS) thermal imagery, and self consistency checks for ice thicknesses up to 50 cm with an error of 30% on average. In addition, we modeled the microwave emission for Arctic first-year ice using the sea ice version of the Microwave Emission Model of Layered Snowpacks (MEMLS). The sea ice conditions used as input for MEMLS were generated using a thermodynamic energy balance model (based on the Crocus model) driven by reanalysis data from European Centre for Medium-Range Weather Forecasts (ECMWF). From unexpected features in the modeled microwave emission and disagreements with the empirically trained SMOS retrieval several shortcomings of the energy balance model and MEMLS were identified and corrected. The corrections include a treatment of mismatch of layer definition between the energy balance model and MEMLS, an adaptation of the reflection coefficient for lossy media in MEMLS, and several smaller corrections. For comparison, two simple models ignoring volume scattering, one incoherent and one coherent, were set up and were found to be able to reproduce the results of the more complex MEMLS model on average. With the simple models, the effects of thin coherent layers, the snow cover, the interface roughness and three different dielectric mixture models for sea ice were explored. It was found that the choice of the mixture model is essential for the relation of sea ice thickness to brightness temperatures in L-band, suggesting sea ice thickness sensitivities from few centimeters to several meters for salinity conditions of the global oceans. The interface properties, especially at the sea ice bottom, were found to be a major uncertainty source when modeling the microwave emission of thin sea ice. In addition, the variability in snow depth, the interface roughness, and the ice surface salinity and temperature were found to have a similar influence on the resulting brightness temperatures, with a strong effect on horizontally (up to 30K) and weak effect on vertically polarized radiation (up to 10K) for temperatures below 260K. A model for simulating coherent microwave emission for thickness distributions of ice and snow was prepared to overcome weaknesses from the single thickness coherent and incoherent models. Comparison to the incoherent model showed that for realistic snow depth distributions obtained from Operation IceBridge (OIB) coherence effects can change the brightness temperatures on the scale of a SMOS footprint up to 10K in horizontal polarization. These findings suggest that the retrieval for the thickness of thin sea ice with satellite based L-band sensors yield higher uncertainties than expected from earlier studies.

Marcus Huntemann

Papers

Thickness retrieval and emissivity modeling of thin sea ice at L-band for SMOS satellite observations