1. Horizontal Datums in the United States: NAD 27 and flavors of NAD 83 (1986), NAD 83(HARN) and NAD 83 (NSRS2007)
My introduction
NGS – TSRC – NHMP
GISC3325
10 March 2008

2. Class Update Exam Wednesday during class and lab periods.
Covers all labs, required readings, lectures and Chapters 5-7 of text.
Computations like those demonstrated in class will be part of the test.

7. www.ngs.noaa.gov Gravity For the American Vertical Datum Obtain Data
Obtain Software
Interactive (OPUS, Online adjustment utilities etc)
Information about NGS activities and programs
Height Mod
national readjustment
Geoid development

9. What did NAD 83 replace?

10. NAD 27
1927 – NAD 27 becomes the 2nd continuous, continental scale datum of the United States
Terrestrial
measurements
Distances measured
manually
Origin point in Kansas
Meades Ranch

11. Problems with NAD 27
First-order surveys designed with an accuracy goal of 1: 25,000
Network was adjusted piecewise rather than simultaneously.
Underlying measurements used were of relatively poor precision.
It was not geocentric
Reference ellipsoid used (Clarke 1866) was not appropriate for space-based positioning.

12. NAD 83 (1986) Took place from 1974 to 1986
Digitized observations from 1800’s to 1980’s
300 persons worked on the project
$37M (in 1974 dollars)
1 million simultaneous equations
Helmert blocking
Doppler and VLBI used but mostly terrestrial observations. NO GPS!!!!
NAD 83 datum moved to Center-Of-Mass of Earth

13. When it came time to determine a new datum, All available Extraterrestrial data were incorporated Through VLBI and --- next slide ----
Although before we leave this one-these sites were the beginnings of our modern CORS network-many of them converted to that very quickly upon completion of the NAD83

14. Why here is a multi-sensor site foreground is a Doppler receiver background a VLBI radio telescope.

15. NAD83(86) DOPPLER STATIONS
Apologize that this slide does not have the Canadian Doppler-this (and all our slides) were created for US audiences
Doppler to tie the North and Central American countries, including Hawaii and the Carribean islands together.
The TCT (explain) and many new Astro Azimuths were also incorporated to further determine Scale and orientation for the new datum.
Orientation of the ellipsoid was determined by the BIH Terrestrial System of 1984 (BTS(84).
Note: GPS was not yet developed to be a significant part- ONLY 5 observations
At this point in time NAD83 and The original WGS84 were equivalent.
When it came time to decide on a datum, we had the foresight to make it earth centered
Unfortunately, we now know we were about 2 meters off ( 1 meter horizontal and 1 meter vertical). (my opinion about why-weaknesses in Doppler as control and lack of further iterations)

16. GPS with NAD 83 (1986) There were no GPS observations in NAD 83 (1986)
With GPS we see:
Accuracy greatly exceeds EDM
Allows connection between points that are not intervisible
Provides ellipsoid heights
Required creation of geoid models (first: Geoid 90) to link ellipsoid to NAVD 88

17. COMPARISON OF DATUM ELEMENTS
NAD NAD 83
ELLIPSOID CLARKE GRS80
a=6,378,206.4 m a=6,378,137 m
1/f = /f =
DATUM POINT Triangulation Station NONE
Meades Ranch, Kansas Earth Mass Center
ADJUSTMENT k Stations k Stations
few hundred Base lines k EDMI base lines
few hundred Astro Az k Astro Azimuths
Doppler and VLBI
This slide gives you an idea of how many stations went into each adjustment and the characteristics of the datum.
The size and shape of the ellipsoid was recommended by the IAG (International Association of Geodesy).
The new adjustment will be defined Exactly the same-ie earth centered and with the GRS80 ellipsoid.
We will continue to use NAD83 as the datum because of legal issues in the US.
To jump ahead slightly, we will also publish ITRF coordinates as part of the transition for the future

18. NAD 27 and NAD 83
The magnitude of the shifts across the US from nearly zero to 100 meters on the west coast. And this doesn’t even show Alaska and Hawaii-several hundred meters.
Because of the magnitude, USGS maps had to be modified as well as many other programs (ie. State Plane Coordinate programs).
Despite our many efforts to inform and prepare everyone for this change, many people were still caught off guard and SURPRISED. Fewer venues for publicizing the adjustment. ACSM and major professional journals were our primary focus.
We are making every effort to inform the public and this time it is enhanced by the electronic age. I can an article to the local newsletters of every state with the push of a button, e.g.
Anticipated shifts for the new adjustment will be no where near this magnitude. Dale will talk more on this later.

19. Network Accuracy - 1 Meter Local Accuracy – First-Order
Results of NAD83 (1986)
Network Accuracy Meter
Local Accuracy – First-Order
(1 part in 100,000)
Through the effort of many people, We achieved a Very consistent and accurate network based on classical (ie terrestrial) observations-and putting it into today’s terms-
After 12 years of hard labor, we were done.
We celebrated, I took a month long vacation and lounged in our success. I was going to have it easy for the next 20 years!
But No sooner had I returned than projects with these strange new 3 dimensional observations began to arrive and the phone started to ring-why weren’t this surveys checking the control-both were first order!...

20. PROBLEMS with NAD 83 (1986) Not “GPSABLE” POOR STATION ACCESSIBILITY
IRREGULARLY SPACED
POSITIONAL ACCURACY inadequate for GPS observations
COMPLAINTS from this new technology called ‘GPS’ were pointing out problems in our brand new datum already.
Existing marks were hard to reach, especially with a few hundred pound GPS receiver.
Questionable satellite visibility at existing stations.
GPS was already pointing out inadequacies in the network. Not just logistical but technical also-why were there decimeter differences in the coordinates using GPS and what was published as first order?
We had always prided ourselves that our network was a degree of accuracy better than the technology that surveyors were using so that it would serve as control. And it was for those that were using ‘old’ technology. But not for a very vocal group who were investing in the ‘new’ technology and for those in the research end of preparing us for the new generation of technology.

21. HIGH PRECISION GPS NETWORKS
“GPSABLE”
Clear Horizons for use with Satellites
EASY ACCESSIBILITY
Few Special Vehicle or Property Access Requirements
REGULARLY SPACED
Always within Km
HIGH HORIZONTAL ACCURACY
A-Order (5 mm + 1:10,000,000)
B-Order (8mm + 1:1,000,000)
Dale tells the story of a meeting with Tenn. Delegates who developed a new network in their state.
I had received their GPS data, observed by a Macrometer and reduced through our new program,OMNI, producing 4 minute epochs. POOR NETWORK DESIGN ALSO
You could almost lose a whole satellite between 4 minute epochs.
He, not wanting to upset our brand new network, did manage to crowbar in their new vectors into our existing network with just a small, little distortion of the vectors while producing close to a 1:1,000,000 accuracy.
Unfortunately for him and NGS, fortunately for the future and NGS, Tenn. Refused to allow him to distort their brand new vectors. Thus a new network was born. The HARN.
And Tenn was not unique, Florida and Wisconsin were struggling with the same issues as they observed new statewide surveys (I might mention quickly and easily) with even better equipment. I remember meetings about how in the world were we going to handle the Wisconsin data with the marked differences to the established network and our desire not to upset all these obviously wonderful NAD83 coordinates that we had worked so hard to produce but being faced with differences that simply could not be ignored or, as Dale says, crowbared in?
And so NGS took on the job of providing a high accuracy GPS network in each state to which all other stations coordinates in the state would be adjusted. It would still be NAD83 but an updated version.

22. A-Order Adjusted to VLBI-Existing FBN-CORS
HARN/HPGN ADJUSTMENT
A-Order Adjusted to VLBI-Existing FBN-CORS
B-Order Adjusted to A-Order
Existing Horizontal (Conventional & GPS) Readjusted to A/B-Order
Adopted Adjustment Date Tags e.g. NAD83(1991)
CA HPGN 1991 only held ONE CORS (MOJAVE) along with existing VLBI
This slide basically shows the adjustment process.
The A-Order was determined by very long GPS vectors to ‘CORS’ in Florida, Mass., and California.
The original eastern strain network- also was used to define the A-order. Unfortunately, the original eastern stain network turned out to have many problems. It was simply done too early and not enough was known (and not enough satellites were available) to produce high quality results.
LAYERED APPROACH, ending in a complete readjustment of ALL data within each state. Including classical
DATE TAG was born. Each state with a different tag for the year it was observed. And while we were struggling to maintain consistency, even when we did it didn’t look like we had it because adjacent stations across borders appeared to be different.
FEATHERED SHIFTS- To further complicate the issue, because not all states were done at the same time or even sequentially with their neighbors, we feathered the residual shifts into the neighboring states creating further inconsistencies in the network (although we were trying to maintain consistency without a general readjustment).
INCONSISTENCY BETWEEN STATES BEGAN!!!

23. HIGH ACCURACY REFERENCE NETWORK
This slide shows the “HARN” network in each state-generally the more points, the more involved the state itself was in the reobservations.

24. Old Datasheet format
Sample datasheet produced from interactive map.

25. New and improved!

26. DX = 0.5378 DY = -1.4980 DZ = 0.1726 - or - DN = -0.6743 DE = 0.4393
DU =
GPS orbits are in ITRF, ground control (mostly) in NAD 83

27. IMPROVING POSITIONAL ACCURACY
TIME NETWORK LOCAL
NETWORK SPAN ACCURACY ACCURACY
NAD METERS (1 part in 100,000)
NAD83(1986) METER (1 part in 100,000)
HARN METER B-order (1.0 ppm)
A-order (0.1 ppm)
CORS meter meter
ACURRACY BUT NOT CONSISTENCY!!
To summarize: This slide shows how far we’ve come since 1986, really since 1927.
Unfortunately, --- Next slide ----

28. New Adjustment Network Accuracies
Median horizontal network accuracy: 1 cm
Median vertical network accuracy: 2 cm

29. Reasons for NAD 83 (2007) Readjustment
Multiple epoch dates
Inconsistencies between states
Need to be Consistent with CORS
Compute Network and Local accuracies
September 24, 2003 NGS Executive Steering Committee approved a plan for the readjustment of the horizontal positions and ellipsoid heights for GPS stations in the contiguous United States.
Orthometric Height adjustment will not be attempted
Our multiple adjustments over time have produced Inconsistencies between the states
Mass confusion with multiple epoch dates,
Consistency with the CORS, very important when we begin to utilize OPUS solutions and as people generally begin to use CORS more in their lower order surveys.
But most of all, we will now be able to remove the Order/class codes which were somewhat misleading or in error with a new way of representing the accuracy of the coordinates.
We Will readjust ALL horizontal positions and ellipsoid heights for only GPS stations.

30. What about Orthometric Heights?
Decision was made not to perform a National Readjustment of orthometric heights at this time
Control would be NAVD 88 not the CORS
No analysis of NAVD 88 control
Very few CORS have NAVD88 heights.
We will not attempt to do a GPS derived orthometric height adjustment at this time.
Bottom line, we would need to know the network accuracies of the NAVD88 marks in existence.
Have you ever tried to do an orthometric height adjustment on just a small project or county? You would then see the incredible and probably impossible task that would be involved.
HEIGHT MOD will be the answer.

31. Heights
h – H – N = 0 + errors in components

32. GPS on Benchmark data for GEOID 2003
h – H – N = 0 + errors in components

33. “Perfect” CORS
Site specific problems

34. Pre-position update trend analysis for TXBM Beaumont, Texas
Pre-position update trend analysis for TXBM Beaumont, Texas. Note UP component showing clear upward trend. The divergence from published value is 1.45 cm with an uncertainty of 2.68 cm. Therefore decision was made to update coordinates.

36. Tectonic Plates NUVEL-1A

39. Geoid Modeling Issues
The new ellipsoid heights will be used in the next hybrid Geoid model.
Future efforts will center on improvements to our gravity data.
NGS has acquired an airborne gravity meter.
We plan a comprehensive set of observations to capture gravity throughout the United States and our possessions.
Airborne gravity data supplements terrestrial and satellite based gravity.

40. (L) New NGS airborne meter (R) Portable absolute meter (A 10)

41. NAD 83 (NSRS 2007) READJUSTMENT
ONLY GPS DATA WAS USED
CORS (started in 1994)
FEDERAL BASE NETWORK (A & B)
COOPERATIVE BASE NETWORK (B)
USER DENSIFICATION NETWORK (First)
AIRPORT SURVEYS (B & First)
A simultaneous rather than layered adjustment strategy was adopted using Helmert blocking.
Basically ALL GPS data loaded through the following networks and projects
I’ve already mentioned the FBN. Critical network tying system to CORS
– next slide --

42. Readjustment General Comments
The CORS/CGPS were the control
Only GPS projects participated
The FBN/CBN Surveys are a key element
Tied to the CORS
Provide more accurate ellipsoid heights
We will make every effort to RIGIDLY constrain the CORS coordinates which are based on a MULTIYEAR solution.
MULTIYEAR SOLUTION This will be done in the ITRF reference frame and transformed into NAD83 (NSRS). More on that later. The multiyear solution is comprised of “rotating” 3-day (72hour) data sets reduced with the PAGES processing software on a computer running under the UNIX operating system, using the Precise Ephemeris over a period Of at least 2 ½ years. From analysis of these sequential data sets not only does NGS derive a very good set of coordinates, but also change over time. You might consider this a 4-dimensional adjustment.
Stress the fact - only GPS projects will be included!
A retrieval of every project from NGSIDB has been completed -- next slide ---

43. More Comments
Computation and database loading of scale factors for each project.
Projects re-weighted in both horizontal and vertical (ellipsoid).
Statewide GPS Readjustments provided extra information about the network on a state-by-state basis.
Work in NAD 83
Transform to ITRF later
Scale factors-basically rescale the uncertainties associated with the horizontal and vertical components of the GPS vectors.
Important step if we are to obtain realistic accuracies on both the horizontal and ellipsoidal components.
The process of determining relative weights by assigning appropriate apriori standard errors to vector components is underway right now. It is a process of pulling individual projects out of the data base, running a free adjustment and then scaling by some form of the “variance of unit weight” rotated into local horizontal and vertical components – remember, vectors are stored in a cartesian system with the Z component parallel to a virtual N-S polar axis, and X and Y aren’t State Plane X and Y, but rather oriented towards or 90 degrees from the greenwich meridian.
Statewide readjustments are giving us an indication of what to expect when all states are combined.

44. Network Accuracy Network accuracy of a control point
A value that represents the uncertainty of its coordinates with respect to the geodetic datum at the 95-percent confidence level
Datum is considered to be best expressed by the Continuous Operating Reference Stations (CORS)
Local and Network accuracy values at CORS sites are considered to be infinitesimal (approach zero)
These accuracies were implemented with the National Readjustment
Network – how well the absolute coordinates are known
A deciding factor for approving a new national readjustment is because of new accuracy standards.
Neither of these accuracy values can be calculated without the covariance matrix for the observations and this matrix can only be calculated from a simultaneous adjustment of all the observational data.

45. Local Accuracy Local Accuracy of a control point:
A value that represents the uncertainty of its coordinates relative to other directly connected, adjacent control points at the 95-percent confidence level
An approximate average of the individual local accuracy values between this control point and other observed control points used to establish its coordinates
Local Accuracies– show how well the coordinates are defined relative to other points in the network
NGS will no longer indicate Order/Class on datasheets.

46. NEW STANDARDS FOR GEODETIC CONTROL(
local accuracy adjacent points
network accuracy relative to CORS
Numeric quantities, units in cm (or mm)
Both are relative accuracy measures
Will not use distance dependent expression
Order/Class codes will no longer be used
The new standards can be found in the following web site.
ERROR ELLIPSES

47. N HEIGHT is precisely determined but only tied to one published benchmark

49. Computations of accuracies
This is an extract from an OPUS extended solution.

50. Annoying details X = ATWA-1*ATWL V = A*X-L σ2 = (VTWV)/DF
A is the design matrix
L is a matrix of observations
W is a weight matrix
X is the vector of unknowns
Q is the variance-covariance matrix (ATA)-1
X = ATWA-1*ATWL
V = A*X-L
σ2 = (VTWV)/DF
Accuracy of a quantity is determined by multiplying standard deviation of variances by diagonal element of the Q matrix.

51. SHIFTS (National results)
< 5 cm. in the horizontal component with an
average shift of 2.2 cm.
< 10 cm. in the vertical component with
an average shift of 4.6 cm.
I was hoping to have a plot in time for ACSM. Unfortunately, I didn’t have time to create it.
I can tell you, we did do a preliminary adjustment of just the FBN.
Results were less than 5 cm. in horizontal with an average of 2 cm.
less than 10 cm. in the up component with an average shift of 4.6 cm.

52. ISSUES
New realization of the CORS coordinates are not available-we recommend using the 2002 coordinates as published. The new Policy statement is posted on the NSRS Readjustment webpage.
Another adjustment MAY be run after the next CORS realization (2005) is released and new CORS coordinates published.
Question- With Absolute antenna calibrations now implemented with ITRF05, what effect will this have on previously reduced vectors which used relative antenna calibrations when combined with new reductions?

56. ITRF and WGS 84
WGS 84 from 1987 to 1994, WGS 84(G730) from 1994 to 1996, WGS 84(G873) from 1997 to 2002 and WGS 84(G1150) from 2002 to present.

57. ITRF to NAD 83 (????)

58. NAD 83 (NSRS 2007) v NAD 83 (CORS 96)
“… approximates but is not equivalent to the more rigorously defined NAD 83 (CORS 96) realization in which CORS coordinates are distributed.”
“ …should be consistent with corresponding NAD 83 (CORS 96) positional coordinates to within the accuracy of GPS data used and the accuracy of the corrections applied…”
NAD 83 (NSRS 2007) coordinates do not account for vertical crustal motion.
NGS did not compute velocities for any of the passive monuments involved in the adjustment.

60. NSRS 2007 adjustment 3,411 projects 67,693 points 283,691 vectors
673 CORS fixed
No. of constrained parameters = 2055
No. of unknown parameters = 203,076
Degrees 0f Freedom = 650,049
Constraints in adjustment: 471 National CORS, 3 Canadian CACS, 1 Mexican CORS and 213 California CGPS sites

61. Overall NAD 83 NSRS 2007 results
Median horizontal network accuracy: 1 cm
Median hz. adj vector residual: 0.52 cm
Median vertical network accuracy: 2 cm
Median vt. adj vector residual: 0.54 cm

62. Same data in datasheet format

63. Time-line for National Readjustment Positions and Ellipsoid Heights
All projects loaded in the NGS database prior to November 15, 2005 were included in the National Readjustment
Projects submitted after 11/15/2005 were accepted and loaded into the database but were not included in the readjustment
Completed adjustment by February 10, 2007 deadline.
New Datasheets including new positions and network accuracies are available NOW.
Project Report (pending)
February 10, 2007 will be the 200th anniversary of the Coast and Geodetic Survey established by Thomas Jefferson.