method and, as a result, a national adjustment,
NAD 83
(
HARN
)
was developed based on a state-by-state high accuracy refer-
ence network (
HARN
)
GPS
campaign. Simultaneously, another
realization
NAD 83
(
CORS96
) was obtained based on the network
of Continuously Operating Reference Stations (
CORS
) (Snay
and Soler, 2008). Due to discrepancies between
NAD
83(
HARN
)
and
NAD 83
(
CORS96
), another readjustment combining all
HARN
and
CORS
observations evolved as
NAD 83
(
NSRS2007
) (Pursell
and Potterfield, 2008). The latest realization
NAD 83
(2011) was
adjusted from multiyear
CORS
data from 1994 to 2010 (Weston
et al
, 2012).
GPS
positioning is computed in the frame of satellite orbits
such as the World Geodetic System of 1984 (
WGS 84
) which
employs WGS 84 as its reference ellipsoid. The geocenter
for
WGS 84
is known to be offset by roughly two meters from
the
NAD 83
geocenter (NIMA, 2000; Soler and Snay, 2004).
WGS 84
maintains its orientation and scale consistent with
the International Terrestrial Reference System (
ITRS
) and its
realizations of the International Terrestrial Reference Frame
(
ITRF
). Consequently,
GPS
-derived coordinates are converted
from
ITRS
to
NAD 83
coordinates.
ITRF
accounts for the mo-
tion of tectonic plates using the no net rotation (
NNR
) model
which assumes the angular momentum caused by the motion
of any tectonic plate is compensated for by the combined
angular momentum of the rest of the tectonic plates (Snay
and Soler, 2000b). In
NAD 83
, the North American Plate is held
fixed where positions do not move. Time-varying parameters
computed in
ITRF
are incorporated in
NAD 83
using a 14-pa-
rameter transformation (7 parameters include 3 translations, 3
rotations and 1 scale parameter in an Earth-centered, Earth-
fixed (ECEF) cartesian coordinate system, and time-varying
components of these parameters result in the reported 14-pa-
rameters) so as to cancel time-varying motion in
NAD 83
thus
effectively holding the North American Plate fixed. However,
it is evident that even after cancelling global tectonic velocity
at
NAD
83, there are residual motions, particularly due to the
active pacific plate, Juan de Fuca Plate, and Glacial Isostatic
Adjustment (
GIA
) across the Western US (Calais
et al
., 2006;
Sella
et al
., 2007).
Since
ITRS
addresses global tectonic velocity, any updates
of this and other technological improvements result in newer
realizations of
ITRF
. Accordingly,
NAD 83
realizations are made
consistent with respective
ITRF
realizations. For example,
NAD
83(2011), 2010.0 is consistent with ITRF2008, 2005.0. The
decimal year represents an epoch which refers to a moment
at which positions and velocity are conceptualized to ex-
ist (Meyer, 2010).
NAD 83
will be replaced in 2022 with the
new North American Terrestrial Reference Frame of 2022
(NATRF2022) which will have no geocentric offset compared
to
ITRF
,
WGS 84
, or IGS frame (
NGS 62
, 2017).
The North American Vertical Datum of 1988 (
NAVD
88) was
the result of the adjustment of levelling networks in the US,
Canada, and Mexico with a single control of tidal benchmark
located at Father Point, Rimouski, Quebec (Zilkoski, 1992).
Hybrid geoid models were used with
GPS
-based techniques to
convert
NAD 83
ellipsoid heights into levelled
NAVD 88
ortho-
metric heights (NGS 58, 1997; NGS 59, 2008). The equation
that relates
NAD 83
ellipsoid heights (
h
) and
NAVD 88
heights
(
H
) is,
H
=
h
–
N
; where
N
is the geoid height derived from
geoid models. The gravimetric geoid (e.g., Earth Gravitational
Model 1996 (EGM96) or Earth Gravitational Model 2008
(EGM2008)) is adjusted with
NAVD 88
heights at benchmarks to
create a hybrid geoid model (Arifuzzaman and Hintz, 2016).
The newer hybrid geoid models result in improved spatial
resolutions compared to the older models and are based
upon more gravity data along with improved computational
techniques (Roman, 2004 and 2009). The successive hybrid
geoid models are
GEOID96
,
GEOID99
,
GEOID03
,
GEOID09
,
GEOID12A
,
and
GEOID12B
.
GEOID12A
and
GEOID12B
are identical everywhere
across the conterminous US except Puerto Rico and the US
Virgin Islands region (Technical Details for
GEOID12
/
GEOID12A
/
GEOID12B
,
NGS
). The geoid models that have finer spatial reso-
lution match closely with
NAVD 88
heights. There will be one
additional hybrid geoid model (
GEOID19
) which will convert
ellipsoid heights to
NAVD
88 heights (Brian Shaw,
NGS
, person-
al communication). Following this,
NAVD 88
will be replaced
in 2022 with a completely gravity-based vertical datum, North
American-Pacific Geopotential Datum of 2022 (
NAPGD2022
),
where ellipsoid heights and
GEOID2022
will be used to derive
orthometric heights (
NGS
64, 2017).
The
NGS
recommends using the most current realizations
of the national datum and geoid model. However, since
there are a plethora of
NAD 83
realizations, and hybrid geoid
models, surveyors, and mapping
GIS
professionals may not be
consistent in their use of the most current datum realization
and geoid model. Indeed, the
MCPD
contains control points
derived from various realizations of horizontal datums as well
as geoid models.
The question arises then, how long (across time) do posi-
tions assigned to control points hold true? Although the basic
problem of the use of various datum realizations needs to be
understood in the context of their developments, three impor-
tant factors that can be analyzed to assess coordinate quality of
the control points are the (a) stability of the physical setting of
each control point, (b) effect of changes in the realization of the
National Datum, and (c) displacement of positions due to resid-
ual velocity caused by the regional tectonic setting. Any posi-
tion specified to a control point will be affected due to changes
in any or all of these factors. This study examines each of these
factors using the
MCPD
as its source data. The analysis of these
aspects of control point positions is relevant in the context of
the quality of the
MCPD
database and can be considered a life-
cycle analysis of control point coordinates in general.
Methods
Survey control points are often subject to ground motion
due to traffic/construction activities nearby, or the soil and/
or geologic conditions of the site. The vertical component of
control points tend to be affected more than the horizontal
component because more often than not ground motion is in
the vertical direction aside from tectonic activity (Fisher and
Conway, 2009). One way to determine if there was a distur-
bance of the physical setting of any control point is to analyze
changes in its position over time. Control points in the
MCPD
contain what are considered data overlaps where the same
control point was surveyed by more than one surveyor over
time. Analyzing these data overlaps across time provides in-
sight describing a control point’s change in position. For this
study, control points were selected based on collection tech-
nique and datum realization and only those control points
acquired using the same realization and collection technique
over different times were used for analysis.
There are also control points that have been surveyed us-
ing two different realizations of a given horizontal datum and,
in some cases, several realizations have been used. These data
overlaps are an ideal source to determine the discrepancy
of coordinates derived from different realizations. Various
datum realizations were able to be compared using the
MCPD
;
NAD 83
(1986) versus
NAD 83
(2011),
NAD 83
(1986) versus
NAD
83
(
NSRS2007
), and
NAD 83
(
CORS96
) versus
NAD
83(2011). To
compare vertical coordinates, available combinations of geoid
models in
MCPD
were similarly selected and analyzed;
GEOID99
versus
GEOID12A
, and
GEOID03
versus
GEOID12A
. Control points
were reported with a stated accuracy convention such as local
accuracy or network accuracy for both horizontal and vertical
216
April 2018
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
