ELEVATION DETERMINATIONS ALONG A CANAL

Transcription

1 ELEVATION DETERMINATIONS ALONG A CANAL IN CENTRAL CALIFORNIA N. Barbella and M. Berber Department of Civil and Geomatics Engineering California State University, Fresno, California, U.S.A. 1. Introduction Subsidence is a phenomenon across the globe. According to USGS (United States Geological Survey) in the U.S.A., California, Texas and Florida are the states that suffer from it most. The major source of subsidence is exploitation of groundwater. Due to groundwater depletion, the water layer that holds the ground disappears and rocks compact. If this compaction happens in a small area, subsidence appears in the form of a sinkhole. If compaction covers large areas like the central valley in California, it makes itself visible as a small land drop over time. According to USGS, subsidence has been a major concern in the central valley of California since the 1950s. Growing populations and demand on agriculture increased the use of groundwater in the valley, thereby exacerbating the already dire situation. USGS studies indicated that by 1970, significant land subsidence (more than 30 cm) had occurred in about half of the valley, or about km 2, and, locally, some areas had subsided by as much as 8.5 m. Reduced surface-water availability during , , and caused increases in groundwater pumping, declines in water levels to near or beyond historic lows, and renewed aquifer compaction. In the central valley, subsidence has caused costly infrastructure damage such as canal buckling and reduced freeboard on canals and bridges. At some points, up to 28 cm of land subsidence was meas ured from 2012 to Scientists have been putting effort into monitoring and understanding subsidence. For this purpose, various types of instrumentation have been used, includ ing InSAR (Interferometric Synthetic Aperture Radar), GPS (Global Positioning System), differential levelling, extensometers, piezometers, etc. InSAR relies on satellite images taken at different times. The images are then processed to reveal relative groundele v ation change over time. InSAR is preferred for projects covering large areas because of its cost and image processing. In comparison, an extensometer measures the compaction and expansion of the aquifer system to a specified depth. It is ideal to have extensometer measurements as well; yet, considering drilling and the expense for the device itself, these cost a great deal in terms of investment. A piezometer is a specialized well that is used to measure water levels at specific depths. Again, installation of piezometers is rather expensive. Differential levelling using automatic or digital level is a precise way of obtaining elevation information, provided that the benchmarks used are outside of the subsidence zone. This technique allows carrying elevation from a known benchmark to other points of interest with high pre ci sion. However, it is not a fea sible Figure 1: Differential levelling procedure. technique for large project areas covering a state or a continent since it requires sur vey crews to take measurements one setup at a time along roads, railroads, etc. When obtaining information on large scales, the use of GPS is more efficient. GPS provides three-dimensional positioning of a point wherein, by observing variations in vertical coordinate, subsidence is determined. In this study, elevations along a canal in Madera County, California are determined using GPS, RTK (Real Time Kinematic) GPS, Network RTK and differential levelling. In addition, subsidence in the area is examined. 2. Methods Currently, differential levelling is accomplished using an automatic or digital level along with one or two levelling rods. A level is attached to a tri pod, and rod readings are taken by observing the rods held on backsight and foresight points. The basic procedure is illustrated in Figure 1. The level is set up approximately halfway between the points. Assume that the 132 GEOMATICA dx.doi.org/ /cig Vol. 70, No. 2, 2016

2 ele vation of point A is known. After lev elling the instrument, backsight and foresight rod readings are taken. Height difference (Δh AB ) between these points is calculated as Δh AB = backsight foresight. Finally, the elevation of point B is calculated as H B =H A +Δh AB. With differential levelling, monumented points called benchmarks are used. Heights of these points are usually given with respect to mean sea level, or more precisely, the geoid (Figure 2). Mean sea level is established through the use of tide gauges, and elevations are determined working inland from a tide gauge. Levelling lines are usually along roads or railways due to the nature of the measurements. This means that elevations of benchmarks are orthometric heights. GPS provides latitude, longitude and ellipsoidal height of a point. Ellipsoidal height is determined with respect to WGS84 (World Geodetic System 1984) ellipsoid, since GPS uses WGS84 ellipsoid as the reference datum. NGS (National Geodetic Survey) has been producing geoid mod els to convert ellipsoidal height obtained from GPS to orthometric height of a specific vertical datum. In the conterminous United States, North American Vertical Datum of 1988 (NAVD 88) is used, wherein by generating refined geoid models, NGS provides geoidal height N, which is the height of the geoid from the reference ellipsoid (Figure 2). h=h+n (1) where h is ellipsoidal height, H is orthometric height and N is geoidal height. Therefore, in using this formula one can move from ellipsoidal height to orthometric height or vice versa. RTK GPS has been widely used in surveying. Yet, in RTK GPS, as the Figure 2: Geoidal height. dis tance between the rover and the ref er ence station increases, the atmospheric conditions at the rover and ref er ence station become increasingly different. This decreases the accuracy and makes it more difficult for the rover to fix the ambiguities. To extend the capabilities of RTK positioning over long distances, the Network RTK concept was investigated by Wubbena et al. [1996]. An RTK Network is a network of permanent GNSS receivers that continuously streams satellite observations to a central server in which combined data is used to gen er ate RTK corrections for precise positioning use within the coverage area of the reference stations. These net work-gen er ated RTK corrections are called Network RTK. Major GPS manufacturing companies have established their own Network RTK systems at local, state and nation al levels. These networks vary in size, from small local networks consisting of only a few reference stations, to dozens of reference stations spanning a whole country. These permanent reference stations stream corrected RTK positioning information to rover receivers. The rover connects to the Network RTK server via radio modem, GSM or Internet, and once it receives the RTK data, it computes its position [Leica 2012]. In this study, the Leica net work in California is used. Figure 3: Kenney irrigation canal (Project Area) runs about 2 km along the south side of Avenue 11 in Madera County, California, U.S.A. (image from Google). The blue line depicts the path taken to carry out differential levelling. Vol. 70, No. 2, 2016 GEOMATICA 133

3 3. Application and Results Control points (CP1 and CP2, established for this study, and existing point USGS 13EGS) placed along the Kenney irrigation canal, which runs about 2 km along the south side of Avenue 11 in Madera County (Figure 3), are surveyed using static GPS method for a period of approximately four hours. The closest NGS point (AC6103), which has precise 3D coordinate information, is surveyed again using static GPS method for approximately eight hours (see Figure 3). To collect the measurements, Trimble R8 receivers are used. Antenna and receiver are mounted on a fixedheight 2-m range pole. The first approach to surveying the points along the canal involves a Trimble R8 set up as a base station on CP1, which is then connected to the radio antenna (Figure 4a). Two other Trimble R8 receivers are used as rovers. Site shots are taken (Figure 4b). With the second approach, the Leica receiver is used as a rover to utilize the Network RTK system established by Leica in California (Figure 5a). This approach is used to complete the survey of the canal (Figure 3). The rationale behind an eight-hour survey made at AC6103 is to obtain precise elevation for this point, as this point is used as the reference point for differential levelling (see section 2). Moreover, subsidence at this point was also found to be a concern. With the static GPS measurements collected at control points (CP1, CP2 and USGS 13EGS), elevations of these points are determined. On the other hand, as well known, GPS is not an effective technique in determining precise elevations if the observation peri od is not long enough. Furthermore, elevation determinations are two to three times worse than horizontal coordinate determinations [Berber et al. 2012]. Therefore, elevation from point AC6103 is carried to the control points in our proj ect area, which utilize differential lev el ling for comparison purposes. For differential levelling, a Sokkia SDL30 digital level (Figure 6a) is used Figure 4: a) Trimble R8 base station setup, b) Trimble R8 rover. Figure 5: a) Leica network in California (image from smartnetna.com), b) Leica Viva rover. with its barcoded rod mounted on a turning plate (Figure 6b). As mentioned above, sur vey began at point AC6103 and con tinued to the control points in our project area. The entire length of the levelling was approximately 10 km (see Figure 3) and took three days to complete. The precision obtained is classified as First Order Class I by Federal Geodetic Control Committee because all our misclosures were below the First Order Class I limit val ues pre scribed by this committee. Elevations determined for three con trol points in our project area are listed in Table 1 for both differential levelling and GPS. Since differential levelling is the most precise method and highest precision is achieved with the survey per formed, it is surmised that the rather large difference at CP2 is a result of errors 134 GEOMATICA Vol. 70, No. 2, 2016

4 associated with GPS measurements at this point. At the other two points, differences are congruent between differential levelling and GPS elevations. GPS ellipsoidal heights are con verted to orthometric heights by adding the geoidal height (see Section 2). This shows the precision of the geoid model (GEOID12B) disseminated by NGS. Table 1: Differences between differential levelling and GPS elevation results (m). Differential levelling GPS Further analysis of elevations are deter mined using differential levelling; RTK GPS and Network RTK are done at the following eight points. These points are close to the base station and on a concrete headwall and this is the reason that these stable points are chosen. In addition, using both differential levelling and GPS, these points are coordinated relative Figure 6: a) Sokkia SDL30 digital level, b) Sokkia SDL30 barcoded rod on a turning plate. Difference USGS 13EGS CP CP Table 2: Discrepancies of RTK GPS and Network RTK from differential levelling (m). to CP1. The results are tabulated in Table 2 and portrayed in Figure 7 for easier interpretation. Leica Network RTK system is in NAD83(2011) epoch. Therefore, to be consistent with comparisons, static GPS survey results and RTK GPS results are transformed into this epoch. As can be seen in Table 2 and Figure 7, the discrepancies between differential levelling and RTK GPS vary between 9 and 13 cm; whereas the differences between differential levelling and Network RTK vary between 4 and 7 cm. Better results obtained with Network RTK can be attributed to the power of the software used by Leica Network RTK system, in that to min i mize the influence of distance-dependent errors, Network RTK sys tem software computes Network RTK corrections, interpolates these corrections, and then transmits them [Lachapelle and Alves 2002] allow ing higher precision to be achieved. The headwall, on which the USGS 13EGS point is set, was damaged. Hence, AC6103 was the only control point which had precise elevation information. As mentioned above, in order to obtain reliable elevation information, this point is surveyed for approximately eight hours using GPS. The elevation of this point obtained using GPS is compared against the RTK GPS Network RTK Figure 7: Discrepancies of RTK GPS and Network RTK from differential levelling. In this figure, the hor i zontal axis shows the number of points used and the vertical axis shows the ele va tion differences (m). Vol. 70, No. 2, 2016 GEOMATICA 135

5 ele vation obtained from the NGS data sheet and the result indicated m subsidence based on NAD83(2011) epoch, which is just below that of 25 mm in Figure 8. As can be seen in Figure 8, the proj ect area subsidence monitored by USGS is about 25 mm. The survey indicated 23 mm subsidence. USGS not only uses permanent GPS stations but also runs other systems (i.e., InSAR, extensometers, piezometers and monitoring wells). The independent result, which is in agreement with USGS sur veys, confirms the subsidence phe nomenon in the cen tral valley. In the end, using the Network RTK system, a topographic survey of the canal is produced (Figure 9). For this project, cross sections of the entire length of the canal including any utilities, utility poles, concrete/stone struc tures, inlets/outlets, nearby gates, side roads, dirt build up, and any major ero sion are surveyed. As mentioned above, the length of this canal is about 2 km. Since it is not possible to show the entire canal, a portion of it is displayed in Figure Conclusions In this project, to determine ele va tions of points of interest, static GPS survey, RTK GPS, Network RTK and differential levelling techniques are used. Differential levelling is still the most precise technique for elevation determinations; however, it requires more time in the field and more than one personnel to complete the survey. Network RTK measurements can be taken with one person using just one GPS receiver. Our results indicated that Network RTK measurements produced comparable results. This can be ascribed to the power of the software used by the Network RTK system. RTK GPS results produced decimeter-level precision for point elevations, which are not suitable for most engineering proj ects. Furthermore, GPS survey con ducted at the NGS point confirmed the subsidence phenomenon in our project area. Figure 8: Subsidence contours in Madera County, CA, U.S.A. (image from USGS). Figure 9: A portion of topographic map of Kenney irrigation canal. Acknowledgements The authors thank Dixon & Associates, Inc. for providing the dig i tal level used for this project. References Berber, M., A. Ustun and M. Yetkin, Comparison of accuracy of GPS techniques. Measurement 45: Lachapelle, G., and P. Alves Multiple reference station approach: overview and current research. Journal of Global Positioning Systems. 1(2): Leica RTK Networks an introduction. System 1200 Newsletter. 52, SmartNet, EU. Wübbena, G., A. Bagge, G. Seeber, V. Böder, P. Hankemeier Reducing distance dependent errors for real-time precise DGPS applications by establishing reference station networks. 9th Int. Tech. Meeting of the Satellite Div. of the U.S. Institute of Navigation, Kansas City, Missouri, September q 136 GEOMATICA Vol. 70, No. 2, 2016