The dynamic height listed on NGS Data Sheets is listed as Geoid 12B.
(1) Is the NGS geoid simply a PER POINT calculation based on gravity measured/calculated at the point's location -or- an actual mathematical/statistical surface based on a series of points where gravity was measured/calculated at each point?
(2) When using GPS with project vertical datum set on NAVD88/Geoid 12A/B, how close do you typically measure to the reported dynamic height?
(3) Except for bad measurements or disturbed monuments, to what in your opinion do you attribute the difference? How much do you accept? how do you address it in your survey control?
I don't worry too much about the dynamic height and use the ortho height. Geoid 12 A/B is not a true gravimetric geoid. It is based on a gravimetric geoid that is then warped to fit NAVD 88 benchmarks.
ScottySantaFe, is it your policy to constrain to the ortho height? Do you check to see if you are "close enough" and just hold GPS solution?
Dynamic heights are appropriate for work with water surfaces over large areas. Most other work in North America is done with Orthometric. There are also other height definitions elsewhere.
Most equipment and software will default to ellipsoidal height for GNSS or use a geoid model to report ortho.
Good description of Geoid 12 given above. The model is interpolated for other points. It will be replaced in 2022 with a best estimate model of a true geoid.
How close it needs to be depends on the application and how close you can get depends on many factors. Averaging time, stability of the GNSS reference station (s) e.g. CORS, space weather, and local reflection environment are all significant.
A cm or two horizontal and two to four vertical are usually achievable for long GNSS static sessions, plus geoid uncertainty.
RTK work is strongly affected by distance from a base and ant bad reception environment.
As stated above, Geoid12B is a hybrid geoid; i.e. it's constrained to physical benchmarks. Beoid22 (or whatever it's called) will be a purely gravity model.
The geoid model (BIN file) itself consists of latitudes, longitudes, and ellipsoid heights in a grid with (I believe) one arc minute spacing; the value for your position is interpolated from the four surrounding grid points.
The following is from the NGS Geoid12B page:
"The relative accuracy of GEOID12B to NAVD88 is characterized by a misfit of +/-1.7 centimeters nationwide"
The algorithm for calculating dynamic height is here: https://www.ngs.noaa.gov/TOOLS/DYNHT/dynamic_ht.pdf
It's better to calculate in the metric system and then convert to feet at the end. Doing that gives 281.55193 meters as the dynamic height for the given example, which, when rounded, hits the value on the data sheet.
The relationship between orthometric height (H), ellipsoidal height (h), and geoid height (N) is: H = h - N. Folks better informed that I am can fill in here, but I think that continual drift from the correctness of this equation on data sheets led to height adjustments.
Here's a good source for more information rather quickly: https://www.ngs.noaa.gov/GEOID/PRESENTATIONS/2007_02_24_CCPS/Roman_A_PLSC2007notes.pdf
Note how similar the gravity numbers are to the old g = 9.8 meters per second squared that we learned in Newtonian physics. That's not a coincidence. A Gal is 1 cm per second squared, so g = 980 Gals. An mGal is 1000 times that, making the gravity numbers on the data sheet correspond to 100,000 time Newton's g, much more accurate, though, and highly localized.
My response (not using your numbering):
1. Regarding your starting comment: No. The Dynamic Height listed on the NGS Data Sheet is not listed as GEOID12B. A geoid model is NOT involved in the determination of orthometric, normal, dynamic, or Helmert of orthometric heights. These heights use raw leveled heights in combination with gravity.
2. Most users outside of the Great Lakes do NOT use dynamic heights. As stated above, a geoid model is not involved in determining dynamic heights.
3. The published GEOID-Ellipsoid separation suitable for use with NAD83 ellipsoid heights is computed from the GEOID12B model. This hybrid geoid model is computed with the gravimetric model as a basis with a surface developed from datum shifts as well as a surface developed from the GPS on Benchmark data set. The rationale for the hybrid geoid model is to provide compatibility with NAVD88.
4. The NGS Data Sheets are created from the NGS data base. The relationship h - H - N = 0 (where h = ellipsoid, H are Helmert orthometric heights and N is the geoid-ellipsoid separation) would be correct if all three heights were accurately determined. As the data sheet retrieves these heights from the data base individually (and NOT forcing them to satisfy the equality) they do not sum to zero.
There is uncertainty in the determination of h, H and N. As techniques, equipment and modeling improve the quantities become better determined.
5. Regarding your point 3, due to the the inability to monitor the national benchmark network, the NGS is largely unaware of changes due to geophysical effects (fluid withdrawal, tectonics, sediment loading, etc) and mark disturbance. Without repeated observations we cannot characterize geophysical effects.
6. If I were to make GPS observations on a monument with a known NAVD height and find a discrepancy more than a few centimeters, I would suspect the published NAVD88 height. I would ideally have included a number of points with NGS-published adjusted NAVD88 heights and thereby be able to isolate problem benchmarks. On this issue it is useful to examine information on the NGS site related to the GPS on BM campaign. It identifies the points excluded from the dataset and the reason. It is useful to note that the GEOID12B model exists because some bad GPS on BM points were included in the original GEOID12.
To get a sense of whether there is a potential problem in my project area, I like to compute what the geoid-ellipsoid separation should be by computing N given good H (i.e. adjusted NAVD88 heights) and h (i.e. well-determined ellipsoid heights) values. A comparison with the N values from the data sheet will identify outliers.
The NGS has been very good at providing information about the upcoming new datums. Check their web site. Check out the presentations posted to their site. Many were developed for non-geodesist audiences.
For some somewhat dated info on aspects of the wonderful world of geoid modeling see: http://geodesyattamucc.pbworks.com/w/file/fetch/65388670/lecture22_Geoid_2013.pdf
GeeOddMike, post: 454039, member: 677 wrote: Dynamic Height listed on the NGS Data Sheet is not listed as GEOID12B. A geoid model is NOT involved in the determination of orthometric, normal, dynamic, or Helmert of orthometric heights. These heights use raw leveled heights in combination with gravity
That's for the data sheets. But remember that any calculation of those values from a GNSS-derived ellipsoidal height WILL depend on a geoid model.
Bill93, post: 454042, member: 87 wrote: That's for the data sheets. But remember that any calculation of those values from a GNSS-derived ellipsoidal height WILL depend on a geoid model.
I agree deriving a height consistent with NAVD88 using GPS requires that a model of the geoid-ellipsoid separation (in the US currently GEOID12B) is needed.
I was mostly thinking of the definitions seen in http://geodesyattamucc.pbworks.com/f/HeightSystemsSneeuw.pdf
and overlooked the reality that these various heights are overwhelmingly now determined via GPS/GNSS in some way.
