I just finished a control survey that runs east-west for 2-1/2 miles with control points spaced +- 1500' apart. The points were located horizontaly using static GPS. Vertically they were leveled. The elevation for point 56 was held. When Geoid03 is applied to the points the results are not as compatible with leveled elevations as when Geiod09 is used. Here is a quick look at the results. Pt 20 is at the west end of the project and 56 is at the east end. I wonder if others are seeing similar results between 09 and 03.
PT No. LEVEL GPS-09 GPS-03
20 6918.17 6918.19 6918.06
1 6914.40 6914.43 6914.30
2 6879.91 6879.94 6879.83
3 6866.61 6866.63 6866.54
4 6860.63 6860.65 6860.56
5 6823.37 6823.38 6823.30
6 6833.58 6833.56 6833.50
7 6785.74 6785.75 6785.71
8 6741.65 6741.66 6741.64
9 6704.18 6704.20 6704.18
56 6686.99 6686.99 6686.99
What you should be comparing is the geoid separation. Then let us know how far you have to go between control points to get a different computed separation, and then tell us if that corresponds to the NGS geoid models not being modeled to anything more than 1 arc minute.
09 geoid height at pt 56=-47.86'
03 geoid height at pt 56=-47.62'
09 geoid height at pt 20=-48.10'
03 geoid height at pt 20=-47.73'
Geoid 09 did a much better job matching the real world than Geoid 03. I was not suprised to see a difference between the two models; but I was a bit suprised to see 03 fail to model the correct slope in such a short distance. The elevation at pt 56 had to be held as there was already existing topo from that point and that elevation. What I was then interested in was which Geoid model would match the actual elevation difference between the control points.
I have found generally that every new geoid model is better in the relative sense, you should really ALWAYS tie to some local GOOD orthometric heights to verify how good in the absolute sense it is if you actually want to agree with local points. It could be that the local points are all bogus too and the geoid model is going to really be better. The new GRAV-D project is all about getting away from local BM's totally.
SHG
In the absolute sense the 09 elevation was a few tenths from the held bench mark. However, for this project that was not important. I was interested in how the geoid actually modeled the elevation difference between control points. NGS bench marks do not exist in this area. Nearest one is along an intersecting highway 15 miles south.
Where might this be? What state?
It's in Wyoming. It's not unuaual to see Geoid contour lines that are close in this part of the world. It can get very frustrating matching GPS and leveled elevations. Working with Geoid 09 has greatly improved process.
