Journal of Surveying Engineering

The digital elevation products produced by the Shuttle Radar Topography Mission (SRTM) are disseminated with different resolutions. This study investigates the differences in information content between different resolutions. Three 1° cells are selected to represent flat area, moderate terrain, and mountain ranges. All of these cells are located in the continental territory of the United States. The theory of information loss caused by the sampling density is shown with the increasing deviation from the reference in the experiment. The error resulting from bilinear interpolation is also assessed. The experimental result also indicates that if the 1-arcsecond SRTM, digital elevation model (DEM) is released for global coverage, the height accuracy would be improved $\sim 10\%$ over the 3-arcsecond product.

As part of the Geoid Slope Validation Survey of 2011 (GSVS11), the global positioning system (GPS) was used in a variety of ways to determine the geodetic coordinates of nearly 400 passive geodetic control marks. The Texas DOT (TXDOT) real-time network (RTN) was one of the GPS positioning tools used. Both 6- and 180-s occupations were performed with the TXDOT RTN, as well as static 20-min occupations [processed using the National Geodetic Survey (NGS) Online Positioning User Service, Rapid-Static (OPUS-RS) tool], and $48+\text{h}$ occupations were adjusted as a network (using the NGS OPUS-Projects tool). Although the original 20-min occupations processed by OPUS-RS had an unforeseen systematic error (invalidating their use in this comparison), a direct comparison between OPUS-RS and the other techniques was still possible by subdividing the OPUS-Projects $48+\text{h}$ receiver independent exchange format (RINEX) files on 218 points into more than 30,000 individual 20-min segments. A comparison between all techniques, after accounting for various systematics such as antenna calibrations, different reference frames, and epochs, showed that the RTN had small biases, but, overall, its accuracy ($1\sigma$) was subcentimeter in longitude, less than 2 cm in latitude, and less than 3 cm in ellipsoid height. An investigation into outliers indicated that the RTN could occasionally have unexplained outliers as large as 10 cm, possibly, but not definitely dependent, on time of day or local environment. The analysis also showed that, for this in-the-field environment, the accuracy of OPUS-RS was somewhat better than expected (biases at the 0.1-cm level; SDs at 0.6 cm in latitude and longitude and 3.1 to 3.4 cm in height) but outliers ($>5\sigma$) did occur in approximately 1.5% of all runs, the largest of which was approximately 1.5 m off in latitude and height, and 0.4 m in longitude. Part, but not all, of this could be explained by being near the coastline. Finally, a significant cautionary result came from this study—users of the NGS OPUS-RS software should make certain that their receivers are generating P2 data (pseudoranges on L2, but not those from L2C). Without that data, OPUS-RS will not function at its highest accuracy. Although this has always been a requirement of OPUS-RS and was historically not an issue, the authors discovered that some GPS receivers inadvertently turn P2 off when L2C is turned on.

The main goal of this paper is to investigate error sources and their influence on measured distance in reflectorless electronic distance measurement (EDM) surveys. The results of field surveys previously performed by many authors does not give the full answer to this problem. Therefore this paper presents theoretical analysis augmented with field survey data obtained by other authors. The following error sources are described: impact of target color and material, atmospheric correction, EDM centering error, beam divergence, and instrumental errors. The results of theoretical analysis resolve the question about the source of errors in reflectorless EDM surveys. These results are broadly consistent with the results of field tests.