Using both altimeter and radiometer, ocean surfaces potentially covered by sea ice can be classified into several sea ice types and age intervals.
The cryosphere plays an important role in moderating the global climate, in part due to its high albedo. Sea ice is seawater that has frozen (not to be mistaken with icebergs that are shed from continental ice).
Both altimeter and radiometer data are included in Geophysical Data Records. By giving a good overview of polar regions (-82.5°S / 82.5°N), Envisat mission is well adapted. However, since the period cycle is 35-days, the sea ice distribution can vary due to processes of formation, motion or melting. Envisat altimeter (RA-2) and radiometer (MWR) data are delivered on DVD-Roms or by FTP and can be obtained from EOHelp.
This dataset has another advantage. The Envisat GDR brings the opportunity to use algorithms other than the classical ocean-oriented ones, that are better suited for non-ocean surfaces. One of them is optimized for sea ice, the so-called sea ice retracker. We used the brightness temperature from the MWR which is a dual-channel radiometer operating at 23.8 GHz and 36.5 GHz and whose primary goal is to measure tropospheric water vapor path delay correction for RA-2 altimeter
The earliest altimetry missions were dedicated to studying the open ocean, but since then the ability of radar altimetry to monitor ice surfaces has also been demonstrated. We use the Broadview Radar Altimetry Toolbox to observe sea ice changes through time.
In order to compute some statistical parameters for the brightness temperatures only on polar regions, we limit the coordinates between 50°N and 82.5°N and between 50°S and 82.5°S. We also apply the land mask here to exclude all data over continents.
IIn the “Datasets” tab, click on “Add Dir” to download the numerous files for cycles 025 and 029 from the GDR Envisat in the same dataset (and use a meaningful name for the dataset to easily identify it later on, for example, Datasets_025_029).
A specific operation is made for statistical computation on each polar region. Select
longitude for “X“,
latitude for “Y”
and for the “Data Expression“, make the difference between the two brightness temperature (ΔTBs = interpole_365_temp_mwr – interpole_238_temp_mwr).
In the “Selection criteria” box:
The data editing uses the formula as follows: altim_landocean_flag <= 1 , in order to apply a land mask. To compute expression on either Arctic or Antarctic region, we add (still in the selection criteria) a selection based on latitude coordinates as defined above (&& is_bounded(50,lat,82.5) , for example, for the Arctic).
Finally, click on “Compute statistics” button. A file result gives the statistical value for Number of valid data, Mean, Standard deviation, Minimum and Maximum.
For a given cycle (here we choose the cycle 025), a second operation is created with the previous mean and the standard deviation values to normalize the ΔTBs as follows: ΔTBs-norm = ( ΔTBs – mean ΔTBs) / standard_deviation ΔTBs which corresponds to our new data expression. For the Arctic region, we have ΔTBs = ( ΔTBs – 4.82) / 10.77 .
In the views menu, we plot the file with the previous operation.
To map the backscatter coefficient (sigma0) for a given cycle (here we choose the Envisat cycle 025), we choose the backscatter coefficient computed from the sea ice retracking in the Ku-band and at 18 Hz. No statistical value is computed for this parameter. So, the operation can directly be made selecting longitude for “X”, latitude for “Y” and for the “Data expression”: 18 Hz Ku-band sea ice backscatter coefficient. The data editing uses the formula as follows :altim_landocean_flag <= 1 && is_bounded(0,hz18_ku_seaice_bscat,45).
Results and comments
|Difference of the two brightness temperature normalized, for the Envisat cycle 025 (2004/03/09 – 2004/04/12) over the polar regions.|
|Backscatter coefficient in Ku-band (ocean retracking) for the Envisat cycle 025 (2004/03/09 – 2004/04/12) over the polar regions.|
The distribution of the ΔTBs normalized values displays several structures with negative values (in dark colours), with neutral values(in light-blue colours) and positive values (in red colours). These structures are in accordance with those shown in Tran et al., 2009. Based on clustering method and reference maps, their study describe several sea ice type from Envisat altimetry and radiometry data. By using their conclusion, we can easily retrieve some of these sea ice category over our maps:
- ice-free ocean for the highest positive ΔTBs-norm values and the lowest sigma0 values (<10 dB),
- multi-year sea ice for the lowest ΔTBs-norm values and low sigma0 values (0-15 dB),
- first year sea ice for neutral ΔTBs-norm values and the highest sigma0 values (20-30 dB).