While making some alterations to the SMP tracks page, I happened to load some tracks from Phil G4OBK’s recent Madeira expedition, and noticed that the summit markers were not being placed exactly where they “should” be. This is usually nothing to worry about, and often happens where a summit position is off by a few seconds of arc, and is only a minor inconvenience. What struck me about the Madeira summits, however, is that ALL the summit-markers were placed consistently about 100 meters to the SSE of the “actual” summit position as seen in the satellite view in Google maps.
This consistent displacement made me wonder whether the Madeira summit positions as stored in the SOTA database are not based on the WGS84 Global Spherical Mercator projection (as used by Google Maps, Google Earth, Bing maps, Yahoo maps and others), but rather on some other, older, coordinate system?
As it happens, accurate positions of these (and other) summits in the WGS84 projection can easily be ascertained from within the SMP by zooming in to the summit and activating the Position -> Latitude/Longitude Grid controls at the top of the map area (other means of doing this, of course, also exist…). Perhaps the Region Manager for Madeira could rectify the summit positions?
Interesting. Jim G0CQK produced the locations from SRTM data which is how we normally check summit locations. That is WGS84 based ISTR. A consistent error offset suggests that maybe two differing coordinate systems have been used.
I know that most UK walkers have their GPS set to BNG (British National Grid) which is OSGB36 based not WGS84. What I don’t know is how a GPS handles the conversion to and from WGS84. If I was writing the software then I would do everything internally as WGS84, the native data format and only convert to the local format when displaying the results. Using OSGB36 datum will be wrong for somewhere so far from the UK as Madeira and if the results were saved in OSGB36, putting them into your code could result in an fixed error.
It would be interesting to download the track which is in error into the same typeof GPS that created it, switch the datum to the correct one for Madeira and then re-upload the track to see if it now aligns with the existing summit position. Maybe the error is now preserved and this would have no effect.
Or maybe the original data we generated for CT3 is wrong.
I think you are over complicating the issue. (Too complicated for me to understand anyway!) I realised that the SOTA quoted positions of the summits were incorrect when I fed Rob’s GPX data into Open Street Map running in Base Camp before our trip (assuming OSM was correct - caveat being that it is a community project). So I adjusted them to what I found was the highest point of contour in OSM and then loaded them back into my GPS.
I think the data is slightly out but like you and Rob don’t know why. I also found the contour data for MI-011 in OSM was less high that it should have been compared to my 1:40000 paper map and the quoted height in the SOTA summits database, having said that MI-011 wasn’t a summit I activated, although Stephan and Jana and Jurg did activate that one last month before we got there.
I’m not saying Phil’s GPS tracks are in the “wrong” position - on the contrary, they seem to agree very well with the positioning of both Google maps’ satellite images and, as well as can be determined, from the Google map contours. So your GPS appears to be working well, Phil.
The SOTA database coordinates for the summits on Madeira, however, seem to be consistently displaced relative to the [map + contours + GPS tracks] group. Hence the question as to whether the SOTA db positions for the Madeira summits might have been derived from a dataset not based on the WGS84 projection - perhaps from old, pre-WGS84 maps of Madeira?
It’s just my amateur programmer’s curiosity coming to the fore - I see a discrepancy, would like to understand where it comes from and, if possible, correct it. It kind of bothers me to see well-laid out GPS tracks arriving at the correct point on the map, only to see the summit marker-icon (the position of which is determined by the SOTA-published coordinates) hanging 100 meters out into space.
And I mentioned the SMP’s ability to accurately report WGS84 coordinates (as provided by the Google Maps API) simply to hint that the summit coordinates may easily be checked there (or indeed elsewhere…) and altered accordingly.
If there’s a constant offset then it usually means a datum difference. To show the effect I set my GPS to OSGB36 and noted the lat&long. Then I set it to SE BASE for Madeira and noted the position. The 2 positions are very different. If you do the maths then the difference in reported position by the GPS is 523.4m!
It’s not a big issue to fix the positions but they should be correct anyway as they came from SRTM data. It will be interesting to find where the error comes from. The position error is not as bad as some Italian summits which we know are badly in error.
It’s good reminder for people to ensure they have their GPS set to the correct datum for the map data they intend to use.
All noted both - in Madeira I had the GPS datum set to WGS84, and will likely do the same when in mainland EU later this month. When I arrive on a summit I will always make for the highest point irrespective of what the GPS says. From there (on HF) I tend to pick out what appears to be the best place to set up within the AZ. That can depend on if there is anywhere to fix the antenna to with bungies / rope rather than having to fiddle with an on pole string support to tent pegs. I expect others do the same, whether using GPS or not.
on our trip in Madeira we observed a similar inaccuracy between the GPS coordinates given in the CT3-ARM and on Jana’s Garmin GPS receiver. She used a Openstreet-Map based map on here GPS device. We didn’t analyze further, whether it is a fixed offset, maybe due to wrong datum or map calibration, or a random deviation.
Jana is going to summarize all our observations and will send soon an e-mail to Andy with all our measured GPS-coordinates. So hopefully the exact GPS coordinates can be updated in a future version of the CT3-ARM.
Of course she will soon upload all GPS tracks of our SOTA-activations to the SOTA-Mapping project.
All of our 7 activated summits could be found very easily, as most of them have a cairn on top, including Pico Penha de Aguia. One exception was the Pico Grande, but it has such a prominent crown that you cannot miss it.
The only summit, we had no clue at all, was the Pico Poio Moledo, according to our GPS track, we were hiking in a distance of only 50-70m nearby its GPS-position, but we didn’t found the summit. So it’s still left for a future SOTA-activation.
Many thanks Andy, for authoring the CT3-ARM and also many thanks Phil for all the up-to-date information. Madeira is really a fantastic region for hiking and SOTA-activations.