Survey of Stapleton Hill East

Transcription

1 31 January 2018 Survey of Stapleton Hill East The Team: John Barnard, Graham Jackson and Richard Mclellan 1) Introduction Stapleton Hill East (Hill Number 17017, Section 38B, OS 1:50000 Map 137, 148, 149, OS 1:25000 Map 201N, Grid Ref. SO349665) is listed as a Tump (Thirty & Upward Metre Prominence) in the Database of British and Irish Hills (DoBIH). The summit area comprises a long fairly narrow ridge situated in open woodland and two visitors have recorded ten-figure grid references for the highest point, both for the same feature. One of the team, Richard Mclellan, also used an Abney level to determine the summit position. More recently, another team member, Dave Marshall studied LIDAR (Light Detection and Ranging) data for Stapleton Hill East and found a summit position lying about 40m SW of this position. The purpose of the survey was to determine the highest point of Stapleton Hill East and then to measure accurately its height and position. This would also allow a comparison between survey data and LIDAR data and forms part of a wider project to determine the usefulness of the latter for the Database of British and Irish Hills. 2) Equipment used and Conditions for Survey A Leica NA730 Professional Automatic level (X30 telescopic system)/tripod system and a 1m E- staff extendable to 5m were used to determine the relative heights of candidate positions for the summit. The absolute height of the summit position was measured using a Leica Viva GS15 receiver. This receiver is a dual-frequency, multi-channel instrument, which means it is capable of locking on to a maximum of 12 GPS and 8 GLONASS satellites, as availability dictates, and receiving two signals (at different frequencies) from each of these satellites. The latter feature reduces inaccuracies that result from atmospheric degradation of the satellite signals. As a stand-alone instrument, it is capable of giving position and height to an accuracy of about two metres and five metres respectively. Despite the on-board features of the GS15 receiver, there are still sources that create residual errors. To obtain accurate positions and heights, corrections were made to the GNSS (Global Navigation Satellite System) data via imported RINEX data from Ordnance Survey which were post-processed using Leica GeoOffice 8.3. Repeated measurements with the Leica Viva GS15 instrument made on the same point yield a height precision of ±0.06m. Note that small hand-held GPS receivers used for general navigation can only receive up to 12 GPS satellites and each at a single frequency and therefore these instruments have a poorer positional accuracy of ±8 metres and a height accuracy of no better than ±15 metres. Some recently produced hand-held GPS Garmin receivers can also receive signals from GLONASS satellites which greatly improve the speed at which these instruments can achieve a satellite fix. Conditions for the survey, which took place between 11.30hr and 15.00hr GMT, were fair. The temperature was about 6 degrees Celsius, but a north westerly wind averaging 16mph, but gusting up to 26mph made conditions feel much colder. It remained dry during the survey, with periods of sunshine, and visibility was good. 1

2 3) Character of Hill Stapleton Hill East lies about 2km WSW of Lingen, a small picturesque hamlet that nestles at the foot of the hill. The hamlet of Stapleton, from which the hill derives its name, lies 3km to its WSW. Good tracks from both hamlets give access to the hill, but that from Lingen is slightly shorter and there is parking (SO ) for about six cars where a path leaves the road and leads to the church. From the parking area, the road is followed South for about 350m where a good track leads up the lower slopes of the hill to the edge of Lingen Vallet Wood. The gated track continues through the wood but heading right at the gate, reaches a grassy ride ascending through mixed woodland. Nearing the summit area, the woods become more open and predominantly of oak. The woodland, until just before the summit itself, is Open Access land. An extract of the Ordnance Survey 1:25000 scale map (Crown Copyright Ordnance Survey) is shown below. The hill has a spot height of 324m near the East end of the summit contour. Just to the North of the summit is a meadow which was identified as a potential position for setting up the Leica Viva GS15. A fence runs along the summit ridge through open oak woodland. To the unaided eye the summit itself is level for about 50m and is not marked by a cairn or other manmade feature. 4) LIDAR Analysis Analysis of the Environment Agency s LIDAR (Light Detection and Ranging) data had previously been conducted by Dave Marshall. Only data at a spatial resolution of 2m were available. These showed three separate areas above 323m. A section of the Excel worksheet used to map the DTM (Digital Terrain Model) is shown below, covering an area 76m x 48m. Grid references submitted to the Database are shown in purple squares and the LIDAR summit is shown in a red square. The green area represents ground below 320m, the beige colour represents ground between 320m and 323m and the brown areas are ground above 323m. 2

3 The grid references (OSGB) and LIDAR heights for the three positions, together with the names of the submitters, are as follows: RichardM m SO RHW m SO LIDAR m SO The LIDAR DSM (Digital Surface Model) was compared with the DTM in the summit area. Differences between the two models were zero at the LIDAR summit location and immediately around it. Differences were also at or close to zero around the two database submissions. All three locations are in an area of deciduous woodland as shown in the satellite image below. The blue rectangle in the image shows the approximate limits of the Excel extract detailed above. Although the area looks heavily wooded in the satellite image, the ground was found to be relatively open during the site visit. The meadow in the left-hand corner of the image is the area identified from the 1:25000 map where the Leica Viva GS15 could be set up to measure the height of Stapleton Hill East (see Section 4.1). A position sufficiently far away from the trees was chosen in order to obtain an unimpeded view of the southern sky and the height difference between this position and the summit was obtained by a line survey. The LIDAR data gives the height of this position as 319.2m. 3

4 4.1) Survey of the Summit The first task was to locate the positions of the two candidates for the summit that previous visitors to the hill (ENE top) and LIDAR (WSW top) had found. This was done using the Go To feature on a Garmin Oregon 450 receiver into which their recorded grid references had been previously entered. These positions were marked with flags. That found by visitors to the hill (ENE top) was an obvious high point by the fence, while that identified by LIDAR (WSW top) was ground 40m WSW immediately adjacent to a large tree and also close to the fence. No other candidates for summit position were identified. These positions are shown in Appendix 1. Next the Leica NA730 level was set up on a tripod a few metres from the ENE top and staff readings were taken on the ENE top, after removal of approximately 10cm of leaf litter to expose natural ground, and the WSW top. Prior inspection of the WSW top revealed that the highest ground was immediately adjacent to the tree, but this raised ground was the result of root growth and therefore was dismissed. The point chosen was about 1m from the tree and near the fence. Staff reading at ENE top = 0.58m Staff reading at WSW top = 0.50m The WSW top (the LIDAR position) is the highest point of the hill by 0.08m Next the Leica NA730 was repositioned to a position just a few metres from the WSW top, the LIDAR summit, in order to explore the ground around the tree in more detail. All staff readings were in the range 0.79±0.01m showing the ground to be level all around the tree. The highest point adjacent to the trunk of the tree gave a staff reading of 0.512m; root growth had raised the ground here by 0.28m. The sequence of photographs showing this set of measurements is given in Appendix 2. 4

5 The density of tree cover was sufficient to preclude our setting up the Leica Viva GS15 on the summit to determine its height. Consequently, staff readings were taken once again on the chosen summit position and also at a position in the open meadow away from the trees as referenced in Section 4. Staff reading at summit position = 0.79m Staff reading at GS15 set-up position= 4.801m The Leica Viva GS15 was set up = 4.01m below the summit To obtain an absolute determination of height the Leica Viva GS15 was set up in the meadow using the short tripod configuration (see photo in Appendix 3 and Section 4). The height of the receiver above the ground was then measured with the integral tape. The vertical offset from measuring point to the ground was 0.578m plus 0.255m for the tribrach/hook system. GNSS data were collected for 1hr with an epoch time of 15 seconds. The data for the Leica Viva GS15 were processed in Leica GeoOffice 8.3 using the ten nearest base stations. The results are given in the table below: - System Easting error(1sd) Northing error(1sd) Height(m) error(1sd) GS The height of Stapleton Hill East is = m (this result takes into account the height difference of the set-up position relative to the summit and the height of the Leica Viva GS15 above the ground at the set-up position). This height calculated from the LIDAR height at the GS15 collection point is 323.2m The height of the lower top 40m ENE is = m Finally, in order to obtain a metre accuracy grid reference to compare with that from LIDAR, the Leica Viva GS15 was set up over the summit position using the short tripod configuration (see photo in Appendix 3). The height of the receiver above the ground was then measured with the integral tape. The vertical offset from measuring point to the ground was 0.578m plus 0.255m for the tribrach/hook system. GNSS data were collected for 15min with an epoch time of 15 seconds. When GNSS data are processed within GeoOffice 8.3 the maximum distance between different solutions for both height and position are set to a default of 0.075m for all Base stations. This means that if this difference is exceeded for a solution from a particular Base station, then that data are rejected from the final calculation. Since the Leica GS15 was positioned adjacent to a tree and in a wood with a short data collection time, the position and height defaults of 0.075m were exceeded for all Base stations. Therefore, the defaults for position and height were changed to 0.5m and 1.0m respectively. This allowed all the data to process and for us to obtain a position to the nearest metre, more accurate than from hand-held Garmin units, to be compared with the LIDAR data. The results are tabulated below: System Easting error(1sd) Northing error(1sd) Height(m) error(1sd) GS N/A N/A The grid reference for the summit as measured by the Leica GS15 is SO (OSTN15) with an estimated error of ±0.2m (3SD). This agrees well with the LIDAR position of SO

6 66508 given that the ground around the tree is flat. This grid reference equates to SO for a Garmin receiver. Note that this measurement was only carried out to determine a position and not a height. The ten figure grid references recorded by hand-held Garmin GNSS receivers for the LIDAR summit (the WSW top) were: - Garmin Oregon 450 SO Accuracy: averaged Height = 336m Garmin Montana 600 SO Accuracy: averaged Height = 327m Garmin Etrex 20 SO Accuracy: averaged Height = 336m The ten-figure grid references for the lower ENE top were: - Garmin Oregon 450 SO Accuracy: averaged Height = 336m Garmin Montana 600 SO Accuracy: averaged Height = 327m Garmin Etrex 20 SO Accuracy: averaged Height = 336m The LIDAR position for the ENE summit is SO (OSTN15) which equates to SO for a Garmin receiver. Again, this agreement is relatively good given the tree cover which obstructs and reflects the satellite signals. 5) Summary of Operating Conditions The latest geoid model was employed and 60 minutes of GNSS data were collected. The cut-off angle of 15 degrees prevents data from satellites close to the horizon being employed in the processing; at these low angles errors due to atmospheric effects compromise the accuracy of the data. Parameter GS15 (set-up) GS15 (summit) Data Collection summit (min) Number of Base Stations used in Processing for all points Epoch Time (sec) Tropospheric Model Geoid Model Hopfield OSGM15 Hopfield OSGM15 Cut off Angle (degs) Max distance between different solutions Position (m) Max distance between different solutions Height (m) ) Coordinate Recovery Analysis In order to verify the precision and consistency of a GNSS dataset, Ordnance Survey recommends a procedure called Coordinate Recovery Analysis. Instead of processing the data with reference to all the nearest OS Base Stations under approximately100km distance, as used in this report, the data 6

7 are first processed with reference to only the nearest Base Station, in this case Shobdon. The data are then reprocessed with the survey point taken as a Reference Point and all the remaining Base stations taken as survey points. These measured values for the OS Base Stations can then be compared directly with the actual OS values for position and height. (This has been carried out via an Excel spreadsheet supplied to us by Ordnance Survey). Although the spreadsheet calculates a number of different parameters, two important ones are presented in the tables below. Height Difference U metres is the vertical height difference between the height of the Base Station as measured in this survey compared with the actual OS value. Separation D ij metres is the distance in 3-d space between the measured and actual OS values for each Base Station. The results for the summit are presented below. Base Station Code Distance to Survey Point km. Height Difference U metres Separation D ij metres Shobdon SHOD 8 Reference Reference Brecon BREC Shrewsbury SHRE Droitwich DROW Machynlleth MACY Stroud STRU Stone STON Cardiff CARI Birmingham BIAX Lichfield LICF The results show a consistent dataset, as all measured OS Base Stations are within 0.05m distance and height of the OS actual values, even for Base Stations up to 100km distance. 7) Discussion of Results For the Leica Viva GS15, a 60-minute data collection time gives results with a measurement uncertainty in height of ±0.06m (3 standard deviations). The height error due to uncertainty in the summit position is estimated as ±0.01m. Therefore the overall measurement uncertainty for summit determination from the Leica Viva GS15 is ±0.06m [ (0.062²+0.012²)]. The Environmental Agency s published specification for the accuracy of their LIDAR data states that the root mean square error is required to be within 0.15m for absolute height and is expected to be within 0.05m for relative height (random error). It furthermore states that in recent surveys the absolute height rms error is seen to fall to c.0.05m. These statistics imply that for practical purposes the accuracy of absolute heights would be ±0.15 for recent data and up to ±0.45m for older data, and ±0.2m for the difference between two proximal heights. We assume the specification pertains to the DSM data, as DTM heights will have additional modelling error which varies considerably with the nature of the terrain. For the purpose of determining the height of a summit, additional error 7

8 will accrue from the spatial resolution of the grid, i.e. the horizontal distance between sampling points (2m in this case), especially in uneven ground. The highest point in the LIDAR data in the vicinity of the summit is m, close to the surveyed height of m. However the two positions are about 3m apart. At the point in the grid closest to the surveyed summit (0.6m away) the LIDAR height is m. At both points the DTM and DSM heights are identical, so tree cover is evidently not interfering here. Note that the survey showed ground in an approximately 5m x 5m area at the summit to be flat to within +/-0.02m; the LIDAR gives variations between m and m (see photographs in Appendix 2). At the lower ENE top, the DTM and DSM heights are both m compared to m from the survey. As we only have Garmin GPS measurements for the location of this top, we cannot determine how far apart the positions are. The height difference between the two summits is 0.22m in the LIDAR and 0.08±0.01m in the survey. It is reasonable to assume that only random error will contribute to the LIDAR figure. The error of 0.14m in the LIDAR result is within the expected range even without taking into account the limitation imposed by the 2m sampling grid, which was probably a factor here. It is perhaps fortuitous that the presence of trees did not lead to a greater discrepancy. On Chatterley Whitfield Hill, surveyed by us in December 2016, the LIDAR DTM summit was 0.31m lower than the surveyed height. This was probably due to the summit being covered by an oak tree, as three other surveyed locations on the same hill that did not suffer from tree cover had errors of 0.12, 0.11 and 0.01m. In dense forest the error in LIDAR heights can be much greater. 8) Summary and Conclusions The summit of Stapleton Hill East is at grid reference * SO and is ground by a tree in open deciduous woodland. Its height is 323.4m ± 0.06m. The lower top 40m ENE has a height of 323.3m. *NB: Grid references for Garmin receivers are quoted in the summary. John Barnard, Chris Crocker, Graham Jackson, Dave Marshall & Richard Mclellan 3 Feb

9 Appendix 1 On the ENE top of Stapleton Hill East The WSW top of Stapleton Hill East with the ENE top in the distance 9

10 Appendix 2 Taking staff readings around the tree on the summit Taking staff reading at base of tree 10

11 Appendix 3 Leica Viva GS15 set up on summit to determine position Tape offset for Leica Viva GS15 set up on summit 11

12 Leica Viva GS15 set up in meadow for determination of summit height Tape offset for Leica Viva GS15 set up in meadow 12