Bathymetry Grid for CATS/CADA Models

Accurate bathymetry is the single most important element in the success of our two tidal model families, the dynamical Circum-Antarctic Tidal Simulation (CATS) model and the inverse Circum-Antarctic Data Assimilation (CADA) model. Traditional trackline depth data is, however, sparse in many regions of the Antarctic. Furthermore, the tidal waves propagate under the floating glacial ice shelves, with the speed of propagation and the rate of energy loss depending on the water column thickness under these shelves.

The basic map for water column thickness under the Amery Ice Shelf (AIS) is from Williams et al. [1998]. The ice shelf has been extended further south, based on satellite observations of the grounding line, kindly provided by E. Rignot (JPL, Pasadena, CA). Water depth in the extension of the AIS is taken to be consistent with the nearby ice-radar-derived values, but there are no real data available for this region.

For the open Ross Sea, we obtained data from the National Geophysical Data Center ("NGDC") and, after quality control, interpolated these data to our CATS/CADA grid. The resultant map of water depth for the open water section of the Ross Sea is similar to that presented by Brancolini et al. [1995]. For the ocean cavity under the Ross Ice Shelf (RIS), we use a depth grid that was kindly provided by David Holland and based on the measurements by Greischar and Bentley [1980]. The RIS cavity geometry has been further updated by changing the grounding line to match the 1993 SCAR coastline [GEBCO-97 CD-ROM].

The major changes to the open-ocean Amundsen Sea is an increase in default shelf water depth from the 100 m used by ETOPO-5, to 400 m, which is consistent with typical measured water depth. We have also realigned the coastline based on the SCAR 1993 coastline data set as provided on the GEBCO-97 CD-ROM. Future work will use recent ship track data to improve the representation of Pine Island Bay, the terminus for the Pine Island Glacier linked to the stability of the West Antarctic Ice Sheet.

The major changes to the open-ocean Bellingshausen Sea is an increase in default shelf water depth from the 100 m used by ETOPO-5, to 400 m, which is consistent with typical measured water depth. We have also realigned some coastline based on the SCAR 1993 coastline data set as provided on the GEBCO-97 CD-ROM. The cavity under the George VI Ice Shelf (GIS), lying between the Antarctic Peninsula and Alexander Island, was also modified. In ETOPO-5 this cavity thickness is only 20 m, greatly limiting tidal propagation between Marguerite Bay (the northern terminus) and Ronne Entrance (southern terminus). A tide gauge at Ronne Entrance was very poorly represented by a model run with the ETOPO-5 bathymetry. Ice-radar measurements [Maslanyj, 1987] indicate, however, that the channel under the GIS has typical thicknesses of ~400-600 m. After updating the grid based on these measurements, the model agrees quite well with the tide gauge at Ronne Entrance.

This section of the CATS/CADA grid is based on the modified bathymetry developed by Robertson et al. [1998] for a Weddell Sea regional tidal model. Modifications to ETOPO-5 were made in several regions where more recent data have become available. Bathymetry based on satellite altimetry and aircraft gravimetric surveys, as well as depth measurements from Ice Station Weddell [ LaBrecque and Ghidella, 1993, 1997] was used in the western Weddell Sea in the region 73^oS to 65^oS, and 60^oW to 44^oW. Minor additional modifications were made along the edges of this region to smooth the transition between the two data sets. Under the FRIS, measurements of water column thickness were used instead of bathymetry [ Vaughan et al., 1994]. This has been further updated in CATS/CADA using the new water column thickness map developed by Johnson and Smith [1997]. Bathymetry in front of the FRIS was modified to agree with Gammelsrød et al. [1994]. The southwestern corner of the grid (Ronne Depression and General Belgrano Bank) was updated using the trackline depth data acquired during the 1998 ROPEX cruise [ Nicholls et al., 1998; Padman et al., 1998a]. Additional modifications were made based on satellite observations of grounded icebergs [ Viehoff and Li, 1995 ]. The addition of ROPEX data had a profound affect of tidal fields in the southern Weddell Sea [ Padman et al., 1998b]. Throughout the domain, depths less than 2 m have been converted to land to avoid the occurrence of negative water column thickness, which can otherwise result as tidal elevation fluctuates.