Scientific Investigations Report 2007–5251
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5251
Version 2.0, June 2013
GPS is a U.S. Department of Defense satellite-based navigation system designed to provide continuous worldwide positioning and navigation capability. For this study, GPS surveys were done to determine the three-dimensional position of monuments in the geodetic monitoring network. The network was established in 1996 by the USGS to determine changes in land-surface elevations in the network (Ikehara and others, 1997) and to establish baseline values for comparisons with results of future surveys.
The geodetic monitoring network, henceforth referred to as the land-subsidence monitoring network, consists of geodetic monuments used as GPS stations (fig. 3). Most geodetic monuments are flat metal disks that are anchored in the ground or to a structure and that can be used in making repeated surveying measurements of horizontal and (or) vertical positions. During the 1996 study by Ikehara and others (1997), historical data for monuments in the southern Coachella Valley were compiled and reviewed to determine the location and the quality of the vertical-control data. Sources of the data include NOAA’s (National Oceanic and Atmospheric Administration) National Geodetic Survey (NGS [formerly the U.S. Coast and Geodetic Survey]), the California Department of Transportation (Caltrans), the U.S. Bureau of Reclamation, and the CVWD (Ikehara and others, 1997). The geodetic monuments were examined before each of the GPS surveys in 1996, 1998, 2000, and 2005 to determine whether any had been damaged or destroyed and to evaluate their suitability for GPS observations.
The original subsidence monitoring network in the southern Coachella Valley was established in 1996 and consisted of 17 geodetic monuments. The network was modified for the 1998 GPS survey by replacing two monuments that had been destroyed (D1299 Tie [D12T] and Caltrans 14.3 Reset 1994 [C143]) with two nearby monuments (G70 1928 [G70] and Caltrans 13.2 1986 [C132]) (fig. 3). The network was again modified for the 2000 GPS survey because monument 54JA was horizontally unstable; the replacement monument (JA54) was installed about 6 m (20 ft) northwest of monument 54JA (fig. 3). In addition, four new monuments—MAGF, MANI, OSDO, and DEEP—were constructed and added in the Palm Desert and Indian Wells areas for the 2000 GPS survey (fig. 3) because the InSAR maps processed for 1996–2000 showed subsidence in these areas (Sneed and others, 2001, 2002). The monument SWC was destroyed by flooding in the Whitewater Stormwater Channel in early 2005, and thus could not be included in the 2005 survey (fig. 3). The spacing between the monuments meets the generalized network design criterion established by Zilkoski and others (1997), which requires that the distance between local network points not exceed 10 km (6 mi).
GPS measurements were made at the geodetic monuments to determine their horizontal positions and ellipsoid heights. Ellipsoid height is the vertical coordinate relative to a geodetically defined reference ellipse; the ellipsoid that closely approximates the Earth’s shape in the study area is the North American Datum of 1983 (NAD83). To determine changes in ellipsoid heights, the heights from successive GPS surveys are compared; then, the differences in the heights are used to determine the location and magnitude of any vertical land-surface changes. The GPS surveying in 1996, 1998, 2000, and 2005 was done in accordance with “Guidelines for Establishing GPS-Derived Ellipsoid Height” by Zilkoski and others (1997) with one minor variation common to all 4 GPS surveys: single baseline, rather than multi-baseline, processing software was used for postprocessing. There are no known conclusive tests that objectively evaluate the effect of using single-baseline, rather than multi-baseline, processing software (Craymer and Beck, 1992). Other variations to the guidelines are specific to particular surveys and are described in the following sections. All of the GPS survey data were recomputed during this current study to eliminate the effects of variable processing methodologies used by assorted software and to standardize processing procedures for an improved comparison of three-dimensional positions derived from data collected during each GPS survey. Software used for the baseline and least-squares adjustment computations was Trimble Geomatics Office version 1.63.
GPS measurements for the 1996 survey were made using six dual-frequency, half-wavelength P-code GPS receivers (Ashtech LD-XII and Ashtech MD-XII) and choke-ring antennas (Dorne-Margolin) at the 17 geodetic monuments between June 3 and 14, 1996, to determine horizontal positions and ellipsoid heights (Ikehara and others, 1997). For this survey, the duration of the GPS measurements was nearly tripled compared with the duration specified by Zilkoski and others (1997) to compensate for using half-wavelength GPS receivers rather than the full-wavelength GPS receivers (Ikehara and others, 1997). GPS measurements were made at the 17 geodetic monuments on at least 2 different days, and data were recorded during 2.5- to 3-hour observation periods (Ikehara and others, 1997). Six of the 17 geodetic monuments—C101, CAHU, COCH, PAIN, D12T, and DUNE (fig. 3)—also were network control stations; GPS measurements were made at these six stations on 3 additional days, and data were recorded during 6-hour observation periods.
Determining the ellipsoid heights of the 17 geodetic monuments in the network involved two phases of least-squares adjustments. During the first phase of least-squares adjustments, the horizontal coordinates and ellipsoid heights for the six Coachella Valley network control monuments were determined by processing the GPS measurements made at these monuments with measurements made simultaneously at three Continuous Global Positioning System (CGPS) stations (DHLG, PIN1, and CRFP) in southern California (fig. 1), and by using precise satellite orbital data and accurate coordinates of the CGPS stations produced by the International GPS Service (IGS) and the Scripps Orbit and Permanent Array Center (SOPAC), respectively. The GPS observations of the CGPS stations were recorded continually (at 30-second intervals), and were archived by SOPAC, a member of the Southern California Integrated GPS network (SCIGN). The network control monuments were selected on the basis of geographic distribution; they are at the perimeters of the monitoring network. During the second phase of least-squares adjustments, the positions of the six Coachella Valley network control monuments were held fixed at the positions determined during the first phase, and the horizontal coordinates and ellipsoid heights for the other 11 monuments were determined. The level of uncertainty for these heights is ±50 mm (±0.16 ft) at the 95-percent confidence level.
GPS measurements for the 1998 survey were made using five dual-frequency, full-wavelength P-code GPS receivers (Ashtech MD-XII) and choke-ring antennas (Dorne-Margolin) at the 17 geodetic monuments between October 5 and 9, 1998, to determine horizontal positions and ellipsoid heights. GPS measurements were made at the 17 geodetic monuments on at least 2 different days, and data were recorded during 45-minute observation periods. Five of the 17 geodetic monuments—COCH, CAHU, PAIN, C101, and G70 (fig. 3)—also were network control stations; GPS measurements were made at these five stations on 3 additional days, and data were recorded during 4.5-hour observation periods.
Determining the ellipsoid heights of the 17 geodetic monuments in the network involved two phases of least-squares adjustments. During the first phase of least-squares adjustments, horizontal coordinates and ellipsoid heights for the five Coachella Valley network control monuments were determined by processing the GPS measurements made at these monuments using measurements made simultaneously at three CGPS stations (DHLG, PIN1, and WIDC) in southern California (fig. 1), and by using precise satellite orbital data and accurate coordinates of the CGPS stations produced by IGS and SOPAC, respectively. The GPS observations of the CGPS stations were recorded continually (at 30-second intervals) and were archived by SOPAC. During the second phase of least-squares adjustments, the positions of the five Coachella Valley network control monuments were held fixed at the positions determined during the first phase, and the horizontal coordinates and ellipsoid heights for the other 12 monuments were determined. The level of uncertainty for these heights is ±20 mm (±0.07 ft) at a 95-percent confidence level.
GPS measurements for the 2000 survey were made using six dual-frequency, full-wavelength, P-code GPS receivers (5 Trimble 4000SSIs and 1 Trimble 4000SSE) and compact L1/L2 Trimble antennas (with ground plane) at the 21 geodetic monuments between August 28 and September 1, 2000, to determine horizontal positions and ellipsoid heights. GPS measurements were made at the monuments on at least 2 different days, and data were recorded during 35-minute observation periods. Six of the 21 geodetic monuments were used as network control stations—COCH, DEEP, CAHU, PAIN, C101, and G70 (fig. 3). GPS measurements were made at these six stations on 3 additional days, and data were recorded during 5-hour observation periods.
Determining the ellipsoid heights of the 21 geodetic monuments in the network involved two phases of least-squares adjustments. During the first phase of least-squares adjustments, horizontal coordinates and ellipsoid heights of the six Coachella Valley network control monuments were determined by processing the GPS measurements made at these monuments with measurements made simultaneously at the same three CGPS stations (DHLG, PIN1, and WIDC) used in processing the 1998 GPS survey data (fig. 1), and by using precise satellite orbital data and accurate coordinates of the CGPS stations produced by IGS and SOPAC, respectively. The GPS observation frequency of the CGPS stations was set at 30 seconds, and the observations were archived by SOPAC. During the second phase of least-squares adjustments, the positions of the six network control monuments were held fixed at the positions determined during the first phase, and the horizontal coordinates and ellipsoid heights for the other 15 monuments were determined. The accuracy of these ellipsoid heights is ±30 mm (±0.10 ft) at the 95-percent confidence level.
GPS measurements for the 2005 survey were made using six dual-frequency, full-wavelength, P-code GPS receivers (Topcon GB1000) and compact antennas (with ground plane) (Topcon PG-A1 Geodetic) at the 20 geodetic monuments between August 15 and 19, 2005, to determine horizontal positions and ellipsoid heights. GPS measurements were made at the monuments on at least 2 different days, and data were recorded during 1-hour observation periods. Six of the 20 geodetic monuments were used as network control stations—COCH, DEEP, CAHU, PAIN, C101, and G70 (fig. 3); GPS measurements were made at these six stations on 3 additional days, and data were recorded during 6.5-hour observation periods.
Determining the ellipsoid heights of the 20 geodetic monuments in the network involved two phases of least-squares adjustments. During the first phase of least-squares adjustments, horizontal coordinates and ellipsoid heights of the six Coachella Valley network control monuments were determined by processing the GPS measurements made at these monuments using measurements made simultaneously at the same three CGPS stations (DHLG, PIN1, and WIDC) used in processing both the 1998 and 2000 GPS survey data, and by using precise satellite orbital data and accurate coordinates of the CGPS stations produced by IGS and SOPAC, respectively. The GPS observation frequency of the CGPS stations was set at 30 seconds, and the observations were archived by SOPAC. During the second phase of least-squares adjustments, the positions of the six network control monuments were held fixed at the positions determined during the first phase, and the horizontal coordinates and ellipsoid heights for the other 14 monuments were determined. The accuracy of these ellipsoid heights is ±20 mm (±0.07 ft) at the 95-percent confidence level.
The horizontal coordinates and the ellipsoid heights of the monuments determined from each of the four GPS surveys were compared to determine the magnitude of horizontal and vertical land-surface changes at the monuments, respectively. The horizontal changes at the monuments were consistent with the northwest movement of the Pacific Plate (with respect to the North American plate) (Shen, Z.-K. and others, 2003). The ellipsoid heights are given in table 1, and ellipsoid-height changes, adjusted to show ellipsoid height values relative to the first GPS measurement for a particular monument (that is, the first measurement was set to equal 0), are given in table 1 and shown in figure 4A.
Comparison of GPS measurements made at the 13 geodetic monuments surveyed in 1996 and in 2005 in the southern Coachella Valley indicate that the elevation of the land surface had a net decline of 124 to 9 mm ±54 mm (0.41 to 0.03 ft ±0.18 ft) during the 9-year period (table 1). Changes at 9 of the 13 monuments (DUNE, R70R, 5211, CAHU, S753, K572, C101, P572, and C427) exceeded the maximum expected uncertainty of ±54 mm (±0.18 ft) at the 95-percent confidence level, indicating that subsidence occurred at these monuments between June 1996 and August 2005. Changes at 4 of the 13 monuments (COCH, PAIN, JOHN, and K70) did not exceed the maximum expected uncertainty of ±54 mm (±0.18 ft) at the 95-percent confidence level, indicating that the vertical positions of these monuments in June 1996 and in August 2005 were similar.
Comparisons of GPS measurements made at 20 geodetic monuments in 2000 and in 2005 indicate that the elevation of the land surface changed –192 to +51 mm ±36 mm (–0.63 to +0.17 ft ±0.12 ft) during the 5-year period (table 1). Changes at 6 of the 20 monuments (MANI, MAGF, OSDO, R70R, JA54, and COCH) exceeded the maximum expected uncertainty of ±36 mm (±0.12 ft) at the 95-percent confidence level—subsidence occurred at five of these monuments (MANI, MAGF, OSDO, R70R, and JA54) and uplift occurred at COCH between August 2000 and August 2005. Changes at 14 of the 20 monuments (DEEP, DUNE, 5211, CAHU, S753, PAIN, C132, K572, JOHN, C101, K70, P572, G70, and C427) did not exceed the maximum expected uncertainty of ±36 mm (±0.12 ft) at the 95-percent confidence level, indicating that the vertical positions of these monuments in August 2000 and in August 2005 were similar.
GPS measurements were made at 16 geodetic monuments (DUNE, COCH, R70R, 5211, CAHU, S753, C132, PAIN, C101, K572, JOHN, P572, SWC, K70, C427, and G70) during at least 3 of the 4 surveys, thus permitting calculations and comparisons of subsidence rates. A seventeenth monument, 54JA, was surveyed in 1996 and 1998 and was then replaced with monument JA54 (about 6 m [20 ft] northwest from 54JA) for the 2000 and 2005 surveys; although the monument location was changed slightly, rates of subsidence were compared. Because GPS data indicated that the heights of PAIN, JOHN, K70, SWC, and G70 were fairly stable (fig. 4A), the subsidence rates of the other 12 monuments are discussed. Eleven of the twelve monuments (DUNE, COCH, R70R, 5211, 54JA/JA54, CAHU, S753, C132, C101, K572, and P572) show decreased subsidence rates between 2000 and 2005 compared with the subsidence rates between 1998 and 2000, a period when the largest rates were computed at these 11 monuments (fig. 4A). The most marked subsidence-rate decreases between these two periods occurred at monuments COCH, 5211, CAHU, and S753, and the least marked subsidence-rate changes occurred at 54JA/JA54. Subsidence rates between 2000 and 2005 at DUNE, R70R, and P572 decreased by factors ranging from 2 to 3, and C132 decreased by a factor of about 5, compared to rates between 1998 and 2000. Calculated subsidence rates at monuments C101 and K572 continually decreased during 1996–2005. GPS results for the twelfth monument—C427— indicated subsidence between 1996 and 1998 and between 2000 and 2005, and uplift between 1998 and 2000, resulting in insignificant vertical position differences between 1998 and 2005. A power pole adjacent to C427 may have degraded the quality of the GPS observations; it is suspected that GPS observations at C427 had higher-than-expected errors.
In the northwestern part of the geodetic network (near MANI, MAGF, OSDO, R70R, and JA54), where significant land subsidence was measured by 2005, and near DUNE, where small amounts of subsidence were measured by 2005, water levels generally showed seasonal fluctuations superimposed on longer-term water-level declines during 1995–2005 (fig. 4A). In the northeastern part of the network, at COCH, where significant uplift was measured by 2005, water levels in nearby wells also showed seasonal fluctuations superimposed on longer-term water-level declines during 1995–2005 (fig. 4A). However, wells 5S/8E-17N1 and 5S/8E-20C2, in which water levels were measured near COCH, are on the west side of the San Andreas Fault (fig. 4A) and the geodetic monument COCH is on the east side of the fault (fig. 3); thus, hydrologic conditions in the two wells probably do not represent hydrologic conditions at COCH.
In the southwestern part of the network near monuments C101, C132, P572, S753, and C427, where small amounts of subsidence were measured by 2005, water levels showed seasonal fluctuations superimposed on longer-term water-level declines during 1995–2005 (fig. 4A).
In the southeastern part of the network near K572, JOHN, K70, and G70, and in the northeastern part of the network near 5211, where small or insignificant changes in elevation occurred, water levels generally showed seasonal variability superimposed on fairly stable or rising water levels during 1995 and 2005 (fig. 4A).
The relationship between ground-water-level changes and concurrent vertical changes at the geodetic monuments is not clearly defined. Complications that contribute to the difficulty in deciphering the relationship include the low frequency of both GPS and water-level measurements, and the complex, often significantly delayed mechanical responses of the aquifer system to ground-water-level changes. However, the locations of the monuments that subsided between 1996 and 2005 are coincident with areas where water levels generally declined during this period and during most of the last century (fig. 4). Similarly, the locations of the monuments that were fairly stable between 1996 and 2005 are coincident with areas where water levels generally were stable or recovered during this period. The coincident areas of subsidence and declining water levels, and of elevation stability and stable or recovering water levels, indicate that aquifer-system compaction may be causing subsidence. In the areas of subsidence, if water levels have declined sufficiently such that the effective stresses exceeded the preconsolidation stresses as is suggested by the historically low water levels shown in figure 4B, the subsidence may be permanent. Although this longer term relationship seems more straightforward, tectonically-induced subsidence also occurs on longer time scales. However, the CGPS stations used as constraints for each of the GPS datasets were fairly stable between 1996 and 2005, indicating that tectonic-induced vertical crustal motion was not included in the subsidence measurements at the geodetic monuments in the network.
U.S. Department of the Interior | U.S. Geological Survey
URL: https://pubs.usgs.gov/sir/2007/5251
Page Contact Information: Publications Team
Page Last Modified: Thursday, 01-Dec-2016 19:55:48 EST
