Suppl. 1 (Detailed Description of the Remote Sensing Analysis)

Suppl. 1 (Detailed description of the remote sensing analysis)

Sea-ice concentration and thickness data utilized in the present study are based on satellite microwave radiometer observations by the instruments Special Sensor Microwave/Imager (SSM/I) and Advanced Multi-frequency Scanning Radiometer (AMSR-E). Both instruments are onboard polar orbiting satellites providing twice-daily coverage of the polar hemispheres.

The sea-ice concentration has been calculated from SSM/I and AMSR-E brightness temperature polarization differences at a frequency of 85.5 GHz and 89.0 GHz, respectively, using the ARTIST Sea Ice (ASI) concentration retrieval algorithm (Kaleschke et al. 2001; Spreen et al. 2008). By using the high-frequency channels of the two instruments the grid resolution of the sea-ice concentration maps is 12.5 km x 12.5 km (SSM/I) and 6.25 km x 6.25 km (AMSR-E). The ASI algorithm takes into account the weather influence, which is larger at the high frequencies applied compared to the low SSM/I and AMSR-E frequencies, by implicitly including weather-induced changes of the brightness temperature polarization difference in an adaptive set of tie points, and by using additional weather filters (Kaleschke et al. 2001; Spreen et al. 2008). It has been shown that this finer spatial resolution allows a better representation of the sea-ice edge and of openings in the sea-ice cover compared to ice-concentration retrieval algorithms which are based on other SSM/I or AMSR-E frequencies (Kaleschke et al. 2001; Kern et al. 2003; Spreen et al. 2008). In addition, the sensitivity to thin ice at the higher frequencies applied is larger than at the other SSM/I or AMSR-E frequencies.

The sea-ice thickness d was calculated from the SSM/I brightness temperature polarization ratio R at 85.5 GHz following the scheme proposed by Martin et al. (2004, 2005). By exploiting the exponential decrease of R with decreasing sea-ice thickness below ~20 cm the ice thickness d can be computed with the following empirical relationship:

α = 248.4368; β = -258.4807; γ = 1.0088. The coefficients α, β, and γ were derived using independent ice-thickness estimates for thin ice in the Ross Sea, Antarctica. The value of γ was fine-tuned for a best match of thin-ice thickness distribution and AMSR-E ASI ice concentration. While Martin et al. (2004) used 37 GHz data, the work in this study is based on 85.5 GHz data for two reasons. First, using coefficients provided by Martin et al. (2004) led to an ice-thickness over-estimation by several centimeters compared to the co-incident in-situ observations (data not shown here). Second, the spatial resolution obtained at 85 GHz is finer and the sensitivity to thin ice is expected to be larger because of the smaller penetration depth into sea ice for microwave radiation at 85 GHz compared to, e.g., 19 or 37 GHz. Our upper detection limit for the sea-ice thickness is ~12 cm; above this value the error caused by sensor noise and uncertainty in the coefficients used in Equation 1 exceeds 4 cm. Resulting ice-thickness distribution maps are blended with AMSR-E ASI ice concentration maps.

References:

Kern S, Kaleschke L, Clausi DA (2003) A comparison of two 85GHz SSM/I ice concentration algorithms with AVHRR and ERS-2 SAR imagery, IEEE Trans. Geosci. Rem. Sens., 41(10), 2294-2306.

Martin S, Drucker R, Kwok R, Holt B (2004) Estimation of the thin ice thickness and heat flux for the Chukchi Sea Alaskan coast polynya from SSM/I data, 1990 – 2001, J. Geophys. Res. 108:C10012, doi:10.1029/2004JC002428.

Martin S (2005) Improvements in the estimates of ice thickness and production in the Chukchi Sea polynyas derived from AMSR-E. Geophys Res Lett 32:L05505.

Kaleschke L, Luepkes C, Vihma T, Haarpaintner J, Bochert A, Hartmann J, Heygster G (2001) SSM/I Sea Ice Remote Sensing for Mesoscale Ocean-Atmosphere Interaction Analysis. Can J Remote Sens 27(5):526–537

Spreen G, Kaleschke L, Heygster G (2008) Sea ice remote sensing using AMSR-E 89-GHz channels. J Geophys Res 113(C2): C02S03 doi:10.1029/2005JC003384

Online resource table 1: In situ measurements of thin-ice thickness compared to thin-ice thickness derived from 85 GHz SSM/I data. SSM/I-based values are derived as the average over 3x3 grid cells with 6.25 km x 6.25 km grid-cell size and are given together with one standard deviation. TI = Thin ice; Br = brash ice.

Online resource figure 1

Online resource figure 1: Distribution of thick and thin sea ice based on Special Sensor Microwave/Imager 85 GHz data for the region of interest for 17 September (A) and 18 September (B) 2007, 4:00 UTC. The thick black line indicates water depths of 1000 m. Thin-ice sample locations are denoted by white diamonds (please refer to Tab. 1 for in situ measured ice thickness). The area covered with thin ice (ice thickness < 12 cm) within the white box is 40 000 km² (A), within the pink box 75 000 km² on 17 September. The dotted black line in B denotes the approximate ice edge position on 30 September 2007 according to the respective AMSR-E ASI sea-ice concentration distribution map. Light yellow = open water; light grey = ice > 12 cm; dark grey = no satellite data; green = land.