|
||||||
|
|
||||||
 |
Antarctic Ice Velocity Data
|
|||||
The above overview map is a subscene from the USGS 1-km AVHRR mosaic (Ferrigno et al., 1996). To view data set maps select Location/Velocity Maps, or use the link on the overview image.
Excerpts taken from Nereson, 1998 (see Related Publications)
Ninety steel survey poles were placed in the firn at Siple Dome in 1994 to a depth of about 1.5 meters in a 10 km square grid surrounding the summit, along a line traversing the north flank of the summit, and in a smaller grid configuration over the relict flow feature. The position of each pole was measured using Global Positioning System (GPS) and optical survey techniques. In November and December 1996, their positions were re-measured and 35 additional poles were placed in the firn along the south flank of Siple Dome, across several scar features near Ice Stream C, and near the original summit grid. The line of poles extending along the south flank of Siple Dome were re-surveyed in December 1997 by Bjorn Johns (UNAVCO) and Xin Chen (NASA).
GPS data were collected with TRIMBLE 4000 SSE dual frequency receivers and L1/L2 Geodetic Antennas using a combination of kinematic and fast static techniques (e.g., Dixon, 1991; Hulbe and Whillans, 1993, 1994). Kinematic techniques were used where the pole spacing was small, or in places where detailed topography was desired. Fast static techniques were used in sparse sections of the grid where survey time was limited by pole-to-pole transit time and where correcting for possible "loss of lock" to the satellites would have been excessively time consuming.
Excerpts taken from Nereson, 1998. See Related Publications.
GPS data were processed using the TRIMBLE GPSurvey 2.20 software package. Satellite position information was obtained from the broadcast ephemeris data file. All baselines were processed with an elevation mask of 10 degrees, and an ionosphere correction was applied to all baselines longer than 5 km. The software allows a wide range of solution types depending on the quality and type of data collected. The optimum solution type (iono-free-fixed) was obtained for all Siple Dome baselines. This means that all baseline solutions contain no ionospheric biases, with a calculated error typically less than 1 mm. The actual measurement uncertainty is greater (5-10 mm) because the error reported by GPSurvey is a measure of the scatter of the calculated solution, and does not reflect systematic errors arising from imprecise ephemeris data or imprecise coordinates of the reference station. Two baselines were measured with both GPS and optical survey techniques. Over a 1.5 km distance, the unadjusted baseline GPS and EDM distances agree to within 0.02 m. After separate adjustment of GPS and optical data, the agreement improves to within 0.01 m.
The network of over 300 baselines is used to calculate the optimal position of each survey pole relative to other poles in the survey. Closure discrepancies are calculated for sets of geometric figures in the grid. The positions of the poles in the network are then adjusted to minimize the closure discrepancies. This process is called "network adjustment." The GPSurvey Network Adjustment Module was used to adjust the Siple Dome network. All Siple Dome survey grids were adjusted according to a "minimally constrained adjustment" where one fixed point is specified, and the network is adjusted around this fixed point using inner constraints. An 8-hour occupation of the summit in 1994 established a baseline between Siple Dome Summit and McMurdo Station, giving a summit position (to the top of the survey pole) of 81.6543390 degrees S, 148.808100 degrees W and an elevation of 622.1 m relative to the WGS84 ellipsoid, with horizontal and vertical errors of 0.05 m and 0.1 m respectively. The summit is held fixed at these coordinates for the network adjustment.
The adjustment results give relative pole positions within the summit grid with residuals less than 0.01 m. Residuals for poles along the north flank and in the Siple Ice Stream grid were larger (0.1 to 0.2 m) since this grid did not have "strength of figure" and did not contain redundant measurements. Residuals along the south flank of Siple Dome were on average 0.02 m in 1996 and 0.03 m in 1997. Vertical positions are generally less well-determined where few baselines were measured. Average vertical residuals are about 0.01 m for the summit grid (1994 and 1996); 0.05 m for the north flank and Siple Ice Stream grid (1994 and 1996); and 0.02 m (1996) and 0.10 m (1997) for the south flank and Duck's Foot.
Additional notes by NSIDC:Estimates of error in speed and bearing listed in the tables were calculated as follows:
Speed:
Normal propagation of errors technique (see, e.g., Data Reduction and Error Analysis for the Physical Sciences, by Philip R. Bevington, McGraw Hill, 1969) was applied to the equations for calculating speed from the station positions and the intervening time. In general, if a measured quantity Y depends on some function of variables (x1, x2, ...) as Y = f(x1, x2, x3, ...), then the uncertainty in Y (call it S) is given by (assuming the x's are independent of each other)
S^2 = s1^2 (df/dx1)^2 + s2^2 (df/dx2)^2 + s3^2 (df/dx3)^2 + ...
For this application, the equation becomes
speed_error^2 = (s_x1^2 + s_x2^2) * ((x2-x1)/(T^2 V))^2
+ (s_y1^2 + s_y2^2) * ((y2-y1)/(T^2 V))^2
+ s_T^2 * (D/T^2)^2
where
x1, y1 are the x,y coordinates for a station in year 1
x2, y2 are the x,y coordinates for a station in year 2
s_x1, s_y1 are the uncertainties in the x,y coordinates in year 1
s_x2, s_y2 are the uncertainties in the x,y coordinates in year 2
T is the intervening time
D is the calculated distance that the station moved
Bearing:
Imagine the two positions of a station, in a pair of years, as two points. Circles could be drawn around these points to represent the uncertainty in determination of their location. The estimated error in bearing was found by determining the angle between the line connecting the two points (centers of the circles) -- call it the axis, and a line tangent to each circle, and crossing the axis (see the figure).
The angle is given by:
err = asin( (dx1 + dx2) / d )
Access the tabular velocity data for Siple Dome: