I appreciate the good words posted above about me, but I think you are being more impressed than is warranted. I probably wouldn't be a good crew member because my surveying knowledge is spotty.?ÿ I like to explore topics and try to understand the fundamentals behind them, but tend to get a little beyond my knowledge sometimes in trying to apply what I've read.?ÿ So it's good to have folks like Loyal, John Hamilton, GeeOddMike, JoeGeodesist, and several others to rein me in when I get too far afield on the theory, and a lot more of you on the practice.
That anomaly in the gravitational field, known as the Midcontinental Rift, is one of the primary reasons that Iowa was selected for the 2014 Geoid Slope Validation Survey performed by the NGS.
https://www.ngs.noaa.gov/GEOID/GSVS14/GSVS14SurveyOverview.pdf
https://www.ngs.noaa.gov/GEOID/GSVS14/
https://www.ngs.noaa.gov/web/science_edu/webinar_series/2014_GSVS.shtml
?ÿ
I've added a couple measurements and revised some ultra-rapid to rapid results (with little change).?ÿ It's looking like I need to repeat the session on the bridge.
?ÿ
?ÿ
The pain is real......
"Surveying isn't for those who don't keep up. The exception is upper management, where little to no surveying happens anyway."
?ÿ
I sure have a lot of upper management solo operators working out of their garages here...?ÿ?ÿ ??ÿ?ÿ
Updated with latest session and recovery info.
Sorry for being so dense, but I had problems following your graphic.?ÿI plotted the points on Google Earth Pro and used its tool to determine the distances from the northernmost to each of the subsequent points. I then plotted the GEOID12B heights with respect to the distance from NJ1008. Using meters as units on both axes yields a more reasonable shape.?ÿ
Looking at data for the points, I make the following observations.
1. Three of the points (NJ1008, NJ0626 and NJ1003) had other than scaled or hand-held positions and NAVD88 published heights determined by differential leveling of some sort.?ÿ
2. Of interest to me ?ÿis that the position of NJ1003 is indicated to be ??NO CHECK? for the most recent nationwide adjustment but was a ??B-order? point in the Iowa 1996 HARN survey. Ellipsoid heights determined prior to the FBN enhancement surveys are in general poorly determined.
3. The three points above ALL participated in the most recent hybrid geoid model. See file: GPSBM2012_CONUS.xls at: https://www.ngs.noaa.gov/GEOID/GEOID12B/GPSonBM12B.shtml
4. The other two points (NJ1005 and NJ0563) have scaled positions which means no ellipsoid heights are available in the NGSIDB. The heights Bill93 determined will contribute otherwise unknown information.
5. Reviewing the information on the data sheet for NJ0626, one can see a 30mm discrepancy in the summation of the published h and H values. The sum of (h - H yields a computed value for N which should agree with the GEOID12B model). Given this discrepancy, I am surprised it was included. Given the paucity of points in Iowa, it may have been considered necessary.
Using data published on its data sheet:
275.439 = ellipsoid height (h)
306.28?ÿ = NAVD88 height (H)
-30.841 = h - H = computed value for the ellipsoid to geoid separation (N)
-30.811 = value for N on the data sheet
computed value differs by 3 cm
6. Using the NGS Project Identifier page to determine the projects in which it appears, I only see two. NJ0626 was NOT included in the original IA HARN survey (GPS1121). That project was observed between Dec 1996 and Feb 1997. The two GPS projects in the NGS data base where NJ0626 appears are GPS179 that appears to be an FAA project to locate points on airports in Iowa then in the FBN enhancement survey (GPS1752) circa 2002. Both surveys obviously included observations of the monument already reported to be ??tilting.?
7. Regarding the RESET points: NJ1003 and NJ1005, while they cannot be assumed to be as accurate as the original monuments (which were part of a level project not merely a one setup height transfer) they are most likely valid.?ÿ?ÿ
8. I hopefully include a graphic showing the extent of glaciation and one showing the likely height differences between the current and future vertical datums; it shows Iowa to see about a 70 cm change.
?ÿ
Wrong graphic in previous post. The Laurentide Ice Sheet extent is shown here.
BTW, here is a better graphic showing Glacial Isostatic Adjustment (GIA ) in N America.
Thanks for taking an interest in the technicalities. That is what I hoped this thread would generate, but it mostly went sideways.
Please ask about any aspects of my plot that you didn??t get figured out.
---
The method of calculating distance has little effect. I used distance along straight line segments connecting the measured points (starting at 1 km because that looked prettier) versus your direct distance from the first point, but that makes less than 3 km difference in the overall distance and doesn??t change the curve shape much, as shown by my re-plot below.
It looks like you fitted a smooth curve to the geoid data at only the measured points, whereas I looked up and entered geoid values at more points along the line segments to get better detail. That accounts for the difference in shape.
I think your Geoid numbers are in IGS. I used the values in NAD83 from the data sheets. That shouldn??t affect the shape of the curve significantly over this span.
---
2. NO CHECK on NJ1003 for the horizontal must mean the Ellip Ht is also unchecked. Does that mean they didn??t submit enough GPS sessions? Its rms value doesn??t indicate a seriously poorer accuracy than the others.?ÿIt is distressing that all of my measurements on NJ1003, even moderately long ones, came up with an Ellip Ht more than 2 cm lower than a B order measurement. Random error? Systematic error? (I??ve checked some others closer) Settlement?
---
5. I note a small mistake in your calculations
NJ0626* NAVD 88 ORTHO HEIGHT 306.268 (meters) 1004.81 (feet) ADJUSTED
Not 306.28 :?ÿ?ÿ?ÿ?ÿ?ÿ 275.439 - 306.268 = -30.829and ??30.829 vs ??30.811 is a difference = 1.8 cm (not 3 cm)
which doesn??t change the conclusion that the mark may have moved since the level run, given its tilt. Perhaps after the end-of-August deadline for one application of GPSonBM, I??ll find a day to get a session on that mark for comparison to see if it looks like it has settled more.?ÿ
Regarding ??paucity of points,? the points plotted across the bottom of my figure show the status of the other bench marks in the area. That was to show that very few remain in areas with open sky. ?ÿThat is typical of my experience in several other counties of Iowa. Road widening has wiped out a great many. Active railroads usually have the disks too close to the tracks for occupation. Even if abandoned, they have most marks at bridges and culverts where trees are plentiful.
---
6. I need investigate the NGS Project data, as you extract information from there I hadn??t seen. NJ0626 is indeed the closest HARN point to the airport. NJ1005 is closer but didn??t have GPS data.
Sorry for the delayed reply. Hopefully I have not made any more transcription or math errors.
With respect to your graphic, you indicate it shows the Geoid 12B surface as well as a surface described as "xGeoid17B - 0.70 m". On this graphic you also depict the values you determined.
Am I correct in assuming that you are computing values of the ellipsoid-geoid separation (geoid height (N) ) as a means to examine the agreement of the model to observations??ÿ
I was curious about the shape of your surfaces. When I plot the N values with respect to either distance from NL1008 or with respect to the decimal latitude, I see a more "regular" shape.?ÿ
The dotted line surface labeled "xGeoid17B - 0.70 m" confuses me. It implies that there is a constant offset between the two surfaces and that it is 0.7 meters. While orthometric heights will change from NAVD88 to NAPGD2022 by around 0.7 meters, the difference in the ellipsoid-geoid separation between the Geoid12B and XGeoid18B models is about 0.3 meters.?ÿ
As the new horizontal datum will approximate the current IGS frame, the expected horizontal differences in coordinates will be between one and two meters. I considered them equivalent when using the XGeoid18B interactive computation tool to determine Geoid12B and xGeoid18B heights.
As two of your points did not have ellipsoid heights: NJ1005 and NJ0563, I used the published NAVD88 heights and ellipsoid-geoid separations (Geoid12B) to compute their ellipsoid heights. The tool requires LLh as input and does not seem to accept on 2D geodetic coordinates. I used the XGeoid12B interactive geoid computation tool at:?ÿ https://beta.ngs.noaa.gov/GEOID/xGEOID18/index.shtml ?ÿ
The inputs for the tool are the NAD83 Latitude, Longitude and ellipsoid height for each point. XYZ coordinates can also be used. The outputs from the tool include the IGS08 positions and ellipsoid heights, Geoid12B height, USGG2012, XGeoid18A and XGeoid18B heights. Also included are a computed a orthometric height from each model and the difference between the modeled Ortho height and the NAVD88 height from the Geoid12B model.
The output from the xGEOID18B tool is here: http://geodesyattamucc.pbworks.com/w/file/128362248/XGeoid18B_output.txt?ÿ
Note that the difference in ortho heights from (NAVD88 and Geoid12B) and (IGS08 and XGeoid18B) is 0.7 meters but also note that this is not equal to the difference either ellipsoid heights or geoid models.
For illustration, let us examine results for NJ1008.
IGS08 ellipsoid height (h)?ÿ ?ÿ = 275.365?ÿm?ÿ ?ÿ NAD83(2011) ellipsoid height (h) = 275.382 m?ÿ ?ÿ ?ÿDifference = 0.017 m?ÿ ?ÿ
XGeoid12B geoid height (N) = -31.121 m?ÿ ?ÿ ?ÿ ?ÿ ?ÿ ?ÿ ?ÿGeoid12B geoid height (N)?ÿ = -30.805 m?ÿ?ÿ ?ÿ Difference = 0.316 m
Using the relationship h - H - N = 0 and solving for H, we rearrange to: h - N = H yielding:
274.365 - (-31.121) = 305.486 based on IGS08 and XGeoid12B approximates NAPGD2022?ÿ
275.382 - (-30.805) = 306.187 based on NAD83(2011) and Geoid12B
difference in orthometric heights: 0.701 m?ÿ?ÿ
I would imagine that the NGS would welcome any observations on NAVD88 published bench marks in their data base.
You also mentioned my use of the Project Identifier tool to determine which projects in the NGSIDB are associated with a PID. This tool is accessed via the data sheet retrieval tool's "Project identifier" option. When choosing this option while not knowing the identifier look for the link to find a project identifier by PID.
The value of finding which projects are associated with a point are that it allows the researcher to retrieve data sheets for all points observed or established during a project. At one time, the NGS had a page called "Line Tool" which allowed users to plot points in a project and retrieve detailed information about the project. Unfortunately, this tool became an orphan and no longer exists. I made queries and have no hope that the tool will return. Drat.
Advise me if I have misunderstood your data and presentation.?ÿ
BTW, the graphics were screen captures of 2D plots using Matlab. No scaling or smoothing was applied. Note that the heights were plotted with respect to latitude resulting in the left-most point being the lowest latitude. ONLY the five points were plotted.
?ÿ?ÿ
>>Am I correct in assuming that you are computing values of the ellipsoid-geoid separation (geoid height (N) ) as a means to examine the agreement of the model to observations??ÿ
Yes. Comparing the discrepancy in ElHt values measured vs NAVD88 and Geoid12b, and also comparing (just) the shape of xGeoid17B.
>>xGeoid17B - 0.70 m
The 0.7 meter was an arbitrary choice that approximates the difference in numerical values of the geoid models when you look them up, in order to get them on the same graph and see how much the shapes resemble each other.?ÿ They won't always have the same difference, but over 30 km you can see the are pretty close to the same difference.?ÿ Since they track each other pretty close in shape, I wasn't worried about deriving the real difference.?ÿ As you point out some of the difference is real and some comes from mixing NAD83 vs IGS/NAPGD2022.
>>I see a more "regular" shape.?ÿ
You aren't plotting enough points (versus whatever) to see the detail in the geoid models.?ÿ I picked a lot more points on lines connecting the observation points, looked up the geoid model values at them, and plotted them connected on the graphs with straight line segments that approximate the shape of the curve along that path on the ground.?ÿ With more points you see a not-so-gentle curve.
I do hope that there will be a good number of GPS observations on bench marks and commend those undertaking this service to the NSRS.?ÿI have always regretted that a more uniform and systematic approach was not possible, there are significant differences in state-level participation. I wonder what impact the surfeit of points in Minnesota has on modeling in neighboring Iowa??ÿ
Reviewing the NGS site for info about the Geoid12B (the latest hybrid model),?ÿ I did not see the the level of detail in reports about earlier models. Two recommended articles are shown below. In any event, the "distortion" of the gravimetric model to fit the GPSonBM data is hardly a simple exercise. Creating the surface from different wavelength data is challenging.?ÿ
I did plot Geoid 12B heights at 30-second intervals along the same meridian spanning the extent of your project area. Both the Geoid12B and XGeoid18B are offset by almost constant 0.3m.?ÿ ?ÿ
?ÿ?ÿ
