Large Difference In Geoid Height

I don't know the answer to your question but I have gotten gas many times at the St. Helena Forest Fire Station (mentioned in the drive to notes) many times. The Emergency Command Center is on the right (call sign "St. Helena" on the radio).

The NGS Geoid Models have come a LONG WAY since GEOID93.

I doubt that .75 meters is really as uncommon as you may think (at least here in Western CONUS). I haven't really investigated GEOID93 v. the later models, but my past experience with GEOID93 indicated that it was NOT very "accurate" in the Mountain West. Obviously some areas were better than others, but as a "rule," GEOID99 was much better, and things (GEOID models), have steadily improved with each new model.

Bear in mind, IT (GEOID Models), are still MODELS, so don't expect perfection (especially around mountains).

I can't wait to see what GRAV_D does to the new GEOID Models here in the Great Basin.

Loyal

I've never seen one that bad in Florida. Mountains do create "geoid undulations" that take an extra dense network of control points to model correctly. If this in an area of sparse population where little 1st order leveling and GPS control control existed to incorporate into the Geoid93, then while it's a little surprising, it's certainly possible. I'd guess that after Geoid93 came out, the locals have noticed that OPUS (or whatever) was missing the local leveled control by a wide margin, and NGS put some extra effort into modeling the geoid in that region.

Very common where I'm from. No cause for alarm. Indeed, geoid models have come a long way, baby.

> A friend of mine was recently working on a project with a NGS datasheet that showed a large difference (2.5 feet) in Geoid Height between GEOID 93 and GEOID12A. I was surprised to see this large of a difference. The point is near the mountains, which may have something to do with it. possibly
>
> In theory, this means you could have two GPS derived NAVD 88 elevations for the same monument that are 2.5 feet different, depending on the GEOID model used.
read up NGS GEOIDS PAGE
>
> How common is this? Am I completely missing something? hopefully someone like loyal will weigh in on this
>

This really isn't a mountainous area ...
Interesting fact is that JT9631 was not used in the 2012 national re-adjustment.
Don't know if it was rejected or not.
One clue seems to be the reference to ortho height being determined with GEOID93.
I haven't seen this reference on a newer datasheet as of yet.
Could be an area subject to vertical displacement.

Examining the data for the point you specified indicates that there was no differential leveling associated with the site. As you note, the NAVD 88 height was derived by GPS OBS meaning its ellipsoid height was used in combination with the current geoid model to derive an approximate NAVD 88 value.

Note that the data sheet shows three ellipsoid heights. The first was determined in 1994 with the subsequent values in NAD 83 (2007) and NAD 83 (2011) the results of adjustments not re-observations (as far as I can tell). I have not examined the NGS site for possible OPUS-DB ellipsoid heights. It should be remembered that in 1994 ellipsoid heights were not the priority they subsequently became. Equipment and processing have tremendously improved since these days.

While some posts speculate that something geophysical might have happened to the point, the decimeter-level difference between the original and subsequent ellipsoid heights is the result of better processing and adjustment techniques with respect to ellipsoid heights.

As for the issue of geoid heights, there have been a number of models used by NGS. Since GEOID 1993 there have been models in 1996, 1999, 2003, 2009 and 2012(A). Each model represents new data, new algorithms and new approaches. If you examine the table of geoid models on the NGS site you will see that GEOID 1993 is listed with the "Gravimetric Geoids" subsequent models labeled "GEOID" and year are "Hybrid Geoids." See http://www.ngs.noaa.gov/GEOID/models.shtml

Hybrid models differ from gravimetric models by the use of additional data set consisting of GPS observations on benchmarks. The equation linking the ellipsoid (h), orthometric (H) and geoid (N) heights is h - H - N = 0. Making GPS observations yields an ellipsoid height, if we make these observations on a benchmark with an orthometric height we can derive the expected value of the ellipsoid-geoid separation which is N.

We expect from using this GPS on BM data to obtain a model that will yield better agreement with the level network. You can see the improvement in the numbers of these points in more recent models. BTW, the latest hybrid model, GEOID 2012A, also including GPS observations submitted to the OPUS-DB tool.

There are a lot of very interesting technical issues with respect to the hybrid process. I find Professor Featherstone's papers on these and related issues to be especially clear. See: http://www.cage.curtin.edu.au/~will/publications.html

While Canada has beaten the US in the replacing their height system (see other posts on this board), the US GRAV-D program seeks to replace the existing NAVD88 network with a more rigorously determined and readily monitored vertical reference system.

Be advised that there is always the possibility of model problems related to poor quality NAVD 88 heights, poor ellipsoid heights, poor density of control and the like. Take a look at some of the posts to this site related to problems with GEOID 2012 (replaced by GEOID 2012A).

HTH,

DMM

Excellent, excellent post Mr. Mike.

Gelid 93 was not useful for me back in the day, so I didn't ever use it and it wasn't recommend for use. 96 wasn't great 99 was better but still messy, the first one that I found that worked was 03 and it still holds up fairly well. The best way I've found to test them is to work in an area with steep geoid slope and compare them to level runs. I know this is time consuming but so what? Get western!!!

I'd never heard of any job done with geoid93, I suppose it may have happened but nothing I ever worked with and I go back to 94 with RTK work, I'm sure there are differences like that just shows the importance of using what is relevant to your job.

Thank you for all of the responses. I've learned a lot. I guess I'm not an old dog just yet.

Landon