Determination of Ice Flow Velocities at the South Pole Using Measurements from the Global Positioning System (GPS)

M S Schenewerk & J R MacKay
NOAA, Silver Spring, MD

L D Hothem & G Shupe
U S Geological Survey, Geometrics Branch, Reston, VA

ABSTRACT

The motion of a GPS antenna located at south pole has been measured. GPS data from Amundsen-Scott, at the south pole, and McMurdo, 1354.6 km distant, have been used to compute the vector baseline between the two stations at 10 day intervals from December, 1991 to December, 1993. These data indicate that the ice sheet at the south pole is moving 10 m/yr along the 40° west longitude line. The ellipsoidal height of the Amundsen- Scott antenna changed -0.22 m/yr during this same period. A second baseline from Santiago, Chili to Amundsen-Scott, 6078.0 km in length, was used to confirm this result and gives excellent agreement.

INTRODUCTION

The U.S. Geological Survey (USGS) has been monitoring ionospheric activity in the Antarctic since December 1991 using a dual-frequency ASHTECH M-XII GPS receiver located at the south pole. Serendipitously, these data can also provide an estimate of the relative position of that antenna with respect to another GPS receiver. The baseline connecting two GPS receivers can be measured to approximately 0.01 parts per billion through a processing technique called double-differencing provided that sufficient observations from mutually visible GPS satellites are available. Double-differencing of GPS phase measurements is presented in the literature and will not be described here. Data from the International GPS and Geodynamics Service (IGS) dual-frequency, geodetic quality ROGUE receiver at McMurdo span the period of interest here and, when combined with the Amundsen-Scott data, can be used to estimate of the position of the antenna at the south pole with an uncertainty of a few centimeters.

PROCESSING

The relative positions of and distances separating the sites used in this experiment are shown in Figure 1. The data were processed using the PAGE3 orbit production software developed at NOAA. Ionosphere-free, double- differenced phase measurements were used as the observable. Observation elevation dependent phase corrections were included to minimize the differences between GPS antenna types. An observation elevation limit of 15° was imposed in post- processing to minimize multipathing effects. Data from March 18 - May 17, 1992 were taken at a 120 second interval at the south pole. Thereafter, a 30 second data interval was used. Four tropospheric adjustments per day were estimated along with the baseline. All data within a 24 hour span were used to make a single baseline estimate. The Scripps daily precise ephemerides were used for data taken before January, 1993. After that time, the NOAA precise ephemerides are available and were used. No additional orbit adjustment was done. The coordinates of the McMurdo and Santiago sites are those provided by the IGS and are consistent with the ITRF 1991.0 system. The NUVEL-1 tectonic plate velocity was assumed for the south pole site. Hereafter, the Amundsen-Scott site will be referred to as amun, McMurdo as mcmu, and Santiago as sant.

RESULTS

Figure 2 shows the differences from the a priori, 1993.0 south pole antenna position for the amun-mcmu baseline. The 1 sigma formal error bars are drawn in each graph but are concealed in the X and Y plots because of the scale. The Z error bars are typical of X and Y also. Lines were fit to each component using a least-squares algorithm which weights each measurement by its formal error. The best- fit lines are draw dashed in Figure 2 and the line intercept, slope, and linear correlation coefficient are given beneath each plot. These line-fits imply a consistent, linear velocity of 9.99±0.001 m/yr along the 39°50'34"±0°.01 west longitude line for the amun antenna. When the fit velocity was removed, the scatter indicates a baseline repeatability of 0.02 PPM. The change in the Z component seen in Figure 2 implies a change in the ellipsoidal height of the antenna at the south pole. The amun- mcmu measurements were rotated into local, topocentric coordinates revealing a -0.219±0.005 m/yr velocity component in the vertical direction.

A second, independent baseline between amun and sant was used to verify the amun-mcmu results. The amun-sant baseline is 6078.0 km in length and roughly perpendicular to the amun-mcmu baseline. The greater baseline length implies correspondingly fewer mutual observations of GPS satellites, but all other details of the data processing were identical to those of the amun-mcmu data. Figure 3 shows the differences between the estimated and the a priori, 1993.0 amun position. Once again, weighted, least-squares lines were fit to the components. The lines and line-fit information are also shown in Figure 3. The baseline repeatability after the site motion was removed is 0.023 PPM. The velocity components measured using these data are 9.74±0.006 m/yr along the 39°23'31"±0°.003 west longitude line; -0.189±0.01 m/yr in the local vertical. Both results are in excellent agreement with those from amun-mcmu.

COMPARISON TO OTHER TECHNIQUES

DOPPLER data from the Amundsen-Scott site are available for 1975 to 1982 and 1987 to 1990. Position estimates of the DOPPLER antenna have been made using 30 day averages were made and the horizontal site velocity measured. These results, 9.88±0.138 m/yr along the 41°11'26"±7°.09 west longitude line (1975-1982) and 10.09±0.002 m/yr along the 40°11'23"±0°.38 west longitude line (1987-1990), are in excellent agreement with the GPS determined velocity. Figure 4 shows the agreement between these techniques after the offset between the DOPPLER and GPS antennas, measured from a site sketch, is removed. Taken as a whole, these two techniques have measured a very consistent velocity during the last 20 years for this region of the ice sheet .

A program to measure the secular change in gravity at the south pole has been underway since 1957 (Greischar, et al., 1992). These measurements show 0.10±0.03 mGal/yr change in gravity at the "new" site between 1974 and 1978. This increase in gravity is roughly equivalent to a -0.33±0.10 m/yr change in height, in reasonable agreement with the GPS determined rate of -0.219±0.005 m/yr.

CONCLUSIONS

The position of a GPS antenna located at Amundsen-Scott site, near the south pole has been measured from two global GPS tracking stations. The velocity measured is 9.99±0.001 m/yr along the 39°50'34"±0°.01 west longitude line with a vertical component of -0.219±0.005 m/yr. The horizontal component show excellent agreement with historical DOPPLER measurements, and the vertical is consistent with on-going gravity measurements at that site. The simplicity, portability, and all weather accessibility of GPS has already proven this tool to be extremely valuable in this harsh environment. However, this work demonstrates that discrete GPS measurements can be usefully linked together using ties to existing global GPS tracking sites and enhancing the value of all measurements.

REFERENCES

Greishar, L.L., Bentley, C.R., and Whiting, L.R., 1992, "An Analysis of Gravity Measurements on the Ross Ice Shelf, Antarctica", Contributions to Antarctic Research III, 57, pp. 105-155.

Mark Schenewerk (mark@tony.grdl.noaa.gov)


Return to the GRD homepage