British Journal of Environmental Sciences

Vol.7, No.2, pp. 70-86, May 2019

Published by European Centre for Research Training and Development UK (www.eajournals.org)
DEVELOPMENT OF 3D DATUM TRANSFORMATION MODEL BETWEEN WGS 84
AND CLARKE 1880 FOR CROSS RIVERS STATE, NIGERIA

Aniekan Eyoh1, Onuwa Okwuashi 2 and Akwaowo Ekpa3

Department of Geoinformatics & Surveying, Faculty of Environmental Studies, University of
Uyo, Nigeria 1, 2 & 3

ABSTRACT: The need to have unified 3D datum transformation parameters for Nigeria for
converting coordinates from Minna to WGS84 datum and vice-versa in order to overcome the
ambiguity, inconsistency and non-conformity of existing traditional reference frames within
national and international mapping system is long overdue. This study therefore develops the
optimal transformation parameters between Clarke 1880 and WGS84 datums and vice-versa for
Cross River State in Nigeria using the Molodensky-Badekas model. One hundred (100) first
order 3D geodetic controls common in the Clarke 1880 and WGS84 datums were used for the
study. Least squares solutions of the model was solved using MATLAB programming software.
The datum shift parameters derived in the study were ΔX=99.388653795075243m ±
2.453509278, ΔY = 15.027733957346365m ± 2.450564809, ΔZ = -60.390012806020579m ±
2.450556881,α=-0.000000601338389±0.000004394,β=0.000021566705811 ± 0.00004133728,
γ = 0.000034795781381 ± 0.00007348844, S(ppm) = 0.9999325233 ± 0.00003047930445. The
results of the computation showed roughly good estimates of the datum shift parameters (dX, dY,
dZ, K, RX , RY , RZ, K ) and standard deviation of the parameters. The computed residuals of the
XYZ parameters were relatively good. The result of the test computation of the shift parameters
using the entire 107 points were however not significantly different from those obtained with the
100 points, as the results showed good agreement between them. Seven reserved points (xsw148,
xsw117, xsw126, xsw97, xsw82, xsw64, xsw155) were used to validate the model.

KEYWORDS: Datum; transformation; MATLAB; Clark 1880; WGS84.

INTRODUCTION

Background of study
The science of geodesy has provided us with two different types of coordinate systems. These
are geocentric and regional (local) coordinate systems (Sella et al., 2002). The origins and axes
of these coordinate systems are different. While the geocentric coordinate system has its origin
at the centre of the mass of the earth and the regional coordinate system has its centre different
from the geocentre. These coordinate systems are associated with the term ‘datum’, which uses
coordinates referred to the surface of defined ellipsoid of revolution. Historically, different
ellipsoids have been chosen by different countries of the world in order to simplify surveying
and mapping in their region and as such these ellipsoids are not necessarily geocentric (Rollins
& Avouac, 2019). In Cross Rivers, Nigeria, the regional (local) coordinate system is the Minna
Datum based on CLARKE 1880 ellipsoid. The geocentric system of Cross Rivers, Nigeria is the
WGS84 ellipsoid. These datums are defined using two parameters i.e. Semi-major axis (a) and
flattening (f). Several assumptions were made in the definition.
Geospatial Cartesian coordinate is a geocentric coordinate system having the earth centre of mass
at its origin. This makes it a valid and unified reference system for the World Geodetic Systems
70

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
1984, which is an earth-centered, earth fixed coordinate system. Datum transformation using
Molodensky equations are becoming increasingly important, because of the growing use of
GNSS data (Trojanowiczl et al., 2018). Very often the spatial data is captured using GNSS whose
reference datum is the earth-centered WGS84 ellipsoid, and have to be transformed to a local
projection with its own ellipsoid and datum (e.g. Clarke 1880 Ellipsoid and Minna datum).
Heiskanen and Moritz (1967) gave the forward transformation from geodetic coordinates (φ,λ,h)
to cartesian coordinates (X,Y,Z) as,

X = (v + h) cosφ cosλ
Y = (v + h) cosφ sinλ (1)
Z = [v(1 – e2) + h] sinφ

Where the prime vertical radius of curvature (v) is given by,

v = a(1 – e2sin2φ)-1/2 (2)

Where a and e represent the semi-major axis and the first eccentricity of the reference ellipsoid
respectively. The parameters in Table 1 were used in equations 1 and 2 to compute forward
transformation from geodetic coordinates (φ,λ,h) to Cartesian coordinates (X,Y,Z). For Minna
datum, it is assumed that ellipsoidal height (h) is equal to orthometric height (h). This implies
that geoidal height (H) is zero; the normal and vertical coincided. To fully describe positions in
relation to the earth, the geodetic coordinate system and Cartesian coordinate system is
employed. The geodetic coordinate system comprises a right–handed orthogonal three-
dimensional coordinates made up of geodetic latitude (φ), geodetic longitude (λ) and ellipsoidal
height (h). They refer to the surface of specific ellipsoid of revolution about its minor axis. The
Cartesian coordinate system is the three-dimensional orthogonal axes in the x, y, and z directions.
Thus, corresponding triplets of Cartesian coordinates refer to these axes. The x-axis is directed
towards the intersection of the Greenwich meridian and equatorial plane. The z-axis is aligned
towards the north pole of the Earth’s rotation. The y-axis is orthogonal to x and z axes and
completes the right–handed coordinate system (Figure 1).

71

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)

Figure 1: A corresponding triplet of Cartesian coordinates

Usually, the more easily potable and understandable coordinates are the Eastings (E) and
Northings (N), which leads us again to another system, called plane (rectangular) coordinate
system. The geodetic coordinate system or Cartesian coordinate system can be projected to plane
coordinate system through appropriate projection models. The need to integrate data between the
WGS84 (UTM) ellipsoid and the Clarke 1880 (NTM) ellipsoid has been an issue of major
concern in the country over the years (Uzodinma, 2005). Unfortunately, there is no generally
acceptable transformation parameter to perform this transformation yet as there has been no
consensus on the particular set of parameters to use (Uzodinma, 2013). Therefore this study aims
to concentrate on a smaller division of the country in order to have a more concentrated area and
increase accuracy.

Due to the advent of modern space based method, data in Nigeria is now being captured in the
WGS84 reference system as opposed to the local datum capture which has been the norm in data
capture in Nigeria. This therefore calls for a major need to integrate the data in both systems in
order to create homogeneity in our reference systems as well as ensuring the integrity of
geospatial information while promoting the sharing and exchange of data across ministries,
agencies, and between the public and private sector and most importantly ensure that end users
achieve the transformation of geospatial information using only one methodology (or tool set)
which will result in transformations of known accuracy, with repeatable and consistent results
that are compatible across state boundaries. (Okeke, 2013).

A coordinate system forms a common frame of reference for the description of positions and on
the other hand, coordinates are simply an ordered set of numbers that are used to describe the
positions or features in a coordinate system (Moritz, 1980, Laundal & Richmond, 2017).
Transformation parameters are required to move from one system to another. In Nigeria, we have
different coordinate systems based on different origins which are used for various mapping
purposes. Also, new technologies like global positioning system have provided new methods of
coordinates’ determination. The map production, update and revision are based on geographical

72

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
coordinates; map-grid coordinates or coordinates in an arbitrary system. Some others are based
on old (local) system. There are no truly accepted transformation parameters the consequences
are obvious confusion and misrepresentation of features. In fact surveyors and survey
practitioners are already using the new technology based on geocentric system while most
available maps and map coordinates are in local system. The effect or implication is multiple
data sets on different systems (Ojigi & Dodo, 2013).

Thus, this project has emphasized on a clear choice of coordinate systems and coordinates
especially as new methods of spatial information capture emerge. It described the methodology
of making different coordinates compatible to be employed in spatial referencing by
determination of transformation parameters. This therefore will help in generalizing features for
representation in two dimensions on flat piece of paper. Hence, a recommendation has been made
to unify all the different coordinates or made to be compatible and flexible by employing least
squares adjustment principles to determine the transformation parameters. Seven parameters of
Helmert transformation are estimated using three-dimensional Cartesian coordinates in Nigeria.
Here, two cases are studied. Cartesian coordinates of WGS 84 and Clarke 1880 in mm level
accuracy, where the information from both cases was generated from secondary source. It
provides the coordinates in millimeter level accuracy. Helmert transformation parameters are
estimated by applying MATLAB code. Seven parameters of Helmert transformation between
WGS 84 and Clarke 1880 datums, and Clarke 1880 and WGS84 datums, and vice-versa are
estimated. Due to lack of large data, the estimation for this project might not be exactly accurate
as estimations of Helmert transformation parameters requires sizable numbers of Cartesian
coordinates based on the project area with high accuracy.

Statement of the problem

The existence of multiple coordinate systems has proved to be a major setback in our map
production, map updates and map revision in our country as well as states. Due to lack of a
perfectly universally adopted set of transformation parameters, it has been problematic
transferring coordinates from one system to another. This has in fact led to a situation where
known points in a certain coordinate system have to be re-observed just to determine their
coordinates in a different coordinates system.

It is therefore necessary to create a good relationship between systems so that we can move from
one system to another easily and maintain a certain level of accuracy. The development of the 7
parameters which would most likely coincide between two systems have been a major project
for Geodesists over the years and therefore this project bases at the realization of an acceptable
connection between the Minna Datum based on the Clarke 1880 Ellipsoid and the Global Datum
based on the WGS84 ellipsoid in Cross Rivers State so that coordinate conversion between the
two systems using certain parameters coincides exactly or very closely to each other.

Aim
This work was aimed at developing a simple, efficient, unique and accurate transformation model
for coordinates between WGS 84 and Clarke 1880 in Cross Rivers State.

73

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
Objectives
The following were the main objectives of this work:
i. The organization of data containing 100 common points in both Clarke 1880 (φ, λ,H) and
WGS 84 (φ, λ,h) geodetic systems.
ii. Compute and transform from geodetic coordinates (φ, λ,h) to Cartesian coordinates
(X,Y,Z)
iii. Develop adequate algorithms and simulations for the determination of the seven
Parameter datum transformation in Cross Rivers State, Nigeria.
iv. Determine the seven optimal transformation parameters between Clarke 1880 and
WGS84 Reference Ellipsoids for Cross Rivers State using the Molodensky Badekas
model.
v. Carry out the validation of the developed transformation parameters with some reserved
control points (testing points) in the network.
vi. Submission of findings and recommendations

METHODOLOGY

Study area
Cross River is a coastal state in southern part of Nigeria, bordering Cameroon to the east. It has
Its capital city at Calabar, and it derived its named from the Cross River, which passes through
the state. Its coordinates are 5°45'0" N and 8°30'0" E in DMS (Degrees Minutes Seconds) or 5.75
and 8.5 (in decimal degrees). The state was created in 1967 from part of the former Eastern
Region, and was known as the 'South-Eastern State until 1976 when it adopted its present name.
The state originally included what is now Akwa Ibom State. Cross River State is located in
Nigeria's Delta region. It is bounded on the North by Benue State, on the South by Akwa Ibom
State, on the East by Cameroun Repulic and on the West by Anambra and Imo States.

Selection of points
A set of 107 points involving coordinates in both the Nigerian Geodetic Network/System
established on the Clarke 1880 spheroid and the world/Global System located on the WGS 84
spheroid were used for this project where seven points was reserved to be used as validation
points (testing samples) therefore, only one hundred (100) points were used in the estimation
(Figure 2). The data was in the geographic units and the ellipsoidal heights were in meters.
However some assumptions were made during the process of this computation as stated in
Table1.

Table 1: Ellipsoid parameters of clarke 1880 and WGS84

Ellipsoid Semi-major 1/f f e2

axis (a) (m)
Clarke 1880 6378249.145 293.465 0.003407561 0.006803511
WGS 84 6378137.000 298.257223563 0.003352811 0.006694380

74

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)

Figure 2: Map showing spread of first order controls in Cross River State

Data Acquisition and Evaluation

The data are first order controls across Cross Rivers State acquired in 2000. The data was
carefully evaluated before using it for the purpose of this exercise.

Transformation of 3D Cartesian coordinates between two Datums

One of the many ways to mathematically transform positions from one datum to another is the
requirement of 'common points'. The common points are surveyed points that have coordinates
in both local and global datums. The achievable accuracy of the datum transformation will be
determined by the number, distribution and transformation technique adopted. Hence, the greater
accuracy required, the more common points are needed.
Further, the choice also depends on the following:

i. The extents of the area for which it is to be applied

ii. The presence of distortion in either of the reference systems
iii. The dimensions of the reference systems i.e. whether it is two or three dimensional and
again
iv. The accuracy requirements.

The 3D similarity transformation was chosen for this study for the obvious reasons that include
the following: it preserves shape and angles while lengths of lines and the positions of points
may be changed. Also, it assumes that there are systematic distortions within the two networks.
The general similarity transformation is the one defined as:

75

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
𝑋𝐵 𝑋𝐴 𝑇𝑋
[ 𝐵 ] = 𝑆𝐹 𝑅 [ 𝐴 ] + [𝑇𝑌 ]
𝑌 𝑌 (3)
𝑍𝐵 𝑍𝐴 𝑇𝑍

Where,

XB,YB, ZB = Coordinates in coordinate system B

XA, YA, ZA = Coordinates in coordinate system A
TX, TY, TZ = Translations terms in the three respective axes
R = 3×3 orthogonal matrix
SF = Scale factor = 1+Δs where Δs is the differential scale

There are seven parameters which are usually associated with a similarity transformation; three
rotation angles, three translational components and one scale factor (Shen et al., 2006). For small
rotation angles, the rotation matrix R is approximated by,

1 𝛼𝑧 −𝛼𝑦
R = [−𝛼𝑧 1 𝛼𝑥 ] (4)
𝛼𝑦 −𝛼𝑥 1

Where αx , αy , αz are the rotation angles in the three axes.

The least squares estimation of the transformation parameters

The least squares solution of the unknown parameters or the estimate of the correction to
approximate parameter vector (x) is given by (Mahboub, 2012).

Ẋ = - (ATPA)-1ATPL (5)

Where, ATPA is a non-singular matrix called Normal equation coefficient matrix, and ATPL is
the normal equations constant (or absolute) term vector. Equation 5 was given by Ghilani (2000)
as equation 6 in which the weight (P) and the column matrix (L) were presented as W and b
respectively,

x = - (ATWA)-1ATWb (6)

Also, the least squares observation equation for a linear mathematical model is given as,

Ax – b = V (7)

V = Ax – b (8)

Where x = column matrix of the unknown parameters [in this case, dX, dY, dZ; Rx, Ry, Rz,
K]
b = column matrix of absolute or differences in Cartesian coordinates local Clarke
1880 system
v = column matrix of the residuals
76

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
A = Design coefficient matrix of the unknown parameters in the observation equation. The
Molodensky-Badekas model as,

𝑑𝑋
𝑑𝑌
𝑉𝑋 1 0 0 𝐷𝑋 0 −𝐷𝑍 𝐷𝑌 𝑑𝑍 𝑋𝑊𝐺𝑆84 − 𝑋𝐶𝐿𝐾
[𝑉𝑌 ] = [0 1 0 𝐷𝑌 𝐷𝑍 0 −𝐷𝑋 ] 𝐾 − [ 𝑌𝑊𝐺𝑆84 − 𝑌𝐶𝐿𝐾 ] (9)
𝑉𝑍 0 0 1 𝐷𝑍 −𝐷𝑌 𝐷𝑋 0 𝑅𝑋 𝑍𝑊𝐺𝑆84 − 𝑍𝐶𝐿𝐾
𝑅𝑌
[ 𝑅𝑍 ]

Where,

𝐷𝑋 𝑋𝐶𝐿𝐾 − 𝑋𝑀
[𝐷𝑌 ] = [ 𝑌𝐶𝐿𝐾 − 𝑌𝑀 ]
𝐷𝑍 𝑍𝐶𝐿𝐾 − 𝑍𝑀

Computing the variables and coefficients of the design matrix

Putting the dimension of each matrix in equation 9 yields,

X1 = (mAnT nWn nAm)-1 mAnT nWn nb1 (10)

In order to extend the matrix to accommodate the number of common points, the number of
observation equations has to be determined. Each point provides 3 observation equations; hence
the number of observation equations (n) were 300, with seven (7) unknown parameters (dX, dY,
dZ; Rx, Ry, Rz and K). From equation 10 the least squares solution for the unknown parameters
(x) together with their statistics can be computed. However, the weight (W) in this study is
assumed same (unit); hence equation 10 becomes,

x = - (ATA)-1ATb (11)

Therefore equation 11 becomes the solution vector to the normal equation for the determination
of the estimates of the approximate or probable parameters (x). The coordinates of 107 common
points in Clarke 1880 and WGS 84 were fully used for an initial quick computation of the mean
shift parameters (dX, dY, dZ) in order to ascertain if the exclusion of 7 points from the network
would make significant difference in the values of the transformation parameters in comparison
with those derived from 100 control points. The main round of computation used 100 common
points, while seven (xsw148, xsw117, xsw126, xsw97, xsw82, xsw64, xsw155) were reserved
for validation or check computations. All computations were carried out using the MATLAB
programming software.

Variance factor and standard deviation

A basic procedure in error analysis is finding the variance factor (𝜎02 ) as derived from the
adjustment that allows calculation of standard deviation of an observation. The variance of unit
weight for weighted observation (A-posteriori) is given as,

77

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
𝑉 𝑇 𝑊𝑉
𝜎02 = (12)
𝑛−𝑚

Where 𝜎02 is the estimated variance factor, n is the number of observation and m is the number
of unknown transformation parameters; hence n-m is the degree of freedom, VTWV is the
weighted sum of the residuals. The observations were assumed to carry equal weights (unit
weight); hence the sum of the residual shall be without weight matrix. Therefore, the standard
deviation of unit weight for the observation is the square root of equation 12, but without the
weight element,

𝑉 𝑉 𝑇
√𝜎02 = √ (13)
𝑛−𝑚

Determination of accuracy by variance-covariance estimation

Variances and covariance of transformation parameters in the adjustment is the basis for
calculation of the absolute and relative error and accuracy. Apart from the solution vector x, the
matrix (ATA)-1 is of great importance. Now to compute the estimated variances and covariance
for the determination of standard deviations of estimated parameters and their residual, the
equations 12 and 13 were used respectively. The variance-covariance matrix is the product of
the unit variance and the inverse of the normal matrix, given by equation 14,

𝐶𝑋 = 𝜎02 (𝐴𝑇 𝐴)−1 = 𝜎02 𝑄 (14)

Therefore, equation 14 is the variance-covariance matrix of the least squares solution in the
adjustment computation.

Validation of the transformation parameters (back-substitution)

In this work, the test of the accuracy of the Molodensky Badekas computed transformation
parameters for Cross River State was achieved by comparing the observed coordinates of the
reserved seven (7) points (xsw148, xsw117, xsw126, xsw97, xsw82, xsw64, xsw155) (Figure 3)
with the back computation coordinates using MATLAB programming software.

RESULTS AND DISCUSSION

The geodetic coordinates were converted to Cartesian coordinates as they have the same ellipsoid
and the axes are aligned with minor and major axis of the ellipsoid. Heiskanen and Moritz (1967)
gave the conversion models as,

X = (N + h) cosφ cosλ
Y = (N + h) cosφ sinλ (15)
Z = [N(1 – (2f - f2)) + h] sinφ

Where N is the radius of curvature in the prime vertical given as,

N = a(1 – e2sin2φ)-1/2 (16)

78

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
The constants a and f are the dimensional parameters of either the regional or geocentric
ellipsoids. In local ellipsoids, the parameter h is not known but if Geoid-ellipsoid separation is
known along with orthometric height (H), then we can use the relation below to find h, as given
by Heiskanen and Moritz (1967).

h=H+N (17)

Where H = Orthometric height, and N = Geoid – ellipsoid separation. This formula was applied
to the data in Microsoft Excel to derive the following set of data where the necessary conversions
were made before computation.

Computation of ellipsoidal heights for Minna Datum

In order to derive the ellipsoidal height for the sets of Minna Datum coordinate provided by
"source", the undulations of all 107 points were computed using the 5-parameter molodensky
standard formula (Agajelu & Moka, 1989),

𝑎 𝑏
𝛥ℎ = 𝑇𝑥 𝑐𝑜𝑠𝜑𝑐𝑜𝑠𝜆 + 𝑇𝑦 𝑐𝑜𝑠𝜑𝑠𝑖𝑛𝜆 + 𝑇𝑧 𝑠𝑖𝑛𝜑 − 𝛥𝑎 (𝑅 ) + 𝛥𝑓 (𝑎) 𝑅𝑁 𝑠𝑖𝑛2 𝜑 . (18)
𝑁
Where,

Φ,λ = WGS84 coordinates of the station

Tx , Ty , Tz = Datum shifts to transform WGS84 datum to Minna datum
Δa, Δf = (Minna minus WGS84) semi-major and flattening respectively
aWGS84 = semi-major axis radius of WGS84 ellipsoid = 6378137m
fWGS84 = flattening of WGS84 ellipsoid = 1/298.257223563
aMINNA = 6378249.145m
fMINNA = 1/293.465
Δa = 112.145m
Δf*104 = 0.54750714 (Δf = 0.000054750714)
RN = radius of curvature of the prime vertical
b/a = 1 – f
e2 = 2f – f2

Since there were no official transformation parameters as at the time, the values adopted for Tx ,
Ty and Tz were applied at the origin point of Minna datum,

𝑇𝑥 = 𝑎𝛿𝜑0 𝑠𝑖𝑛𝜑0 𝑐𝑜𝑠𝜆0 + 𝑎𝛿𝜆0 𝑐𝑜𝑠𝜆0 𝑠𝑖𝑛𝜑0 − 𝑐𝑜𝑠𝜑0 𝑐𝑜𝑠𝜆0 (𝛿ℎ0 + 𝛿𝑎 + 𝑎𝛿𝑓𝑠𝑖𝑛2 𝜑0 ) (4.3b)
𝑇𝑦 = 𝑎𝛿𝜑0 𝑠𝑖𝑛𝜑0 𝑠𝑖𝑛𝜆0 + 𝑎𝛿𝜆0 𝑐𝑜𝑠𝜑0 𝑐𝑜𝑠𝜆0 − 𝑐𝑜𝑠𝜑0 𝑠𝑖𝑛𝜆0 (𝛿ℎ0 + 𝛿𝑎 + 𝑎𝛿𝑓𝑠𝑖𝑛2 𝜑0 ) (4.3c)
𝑇𝑧 = −𝑎𝛿𝜑0 𝑐𝑜𝑠𝜑0 − 𝑠𝑖𝑛𝜑0 (𝛿ℎ0 + 𝛿𝑎 + 𝑎𝛿𝑓𝑠𝑖𝑛2 𝜑0 ) + 2𝑎𝛿𝑓𝑠𝑖𝑛𝜑0 (19)

Where,
𝜑0 𝜆0 = latitude and longitude respectively of the origin point.
𝛿𝜑0 𝛿𝜆0 𝛿ℎ0 = differences between Minna datum and WGS84 datum latitudes, longitudes
and ellipsoidal heights respectively of the origin point.
a = semi-major radius of the reference ellipsoid used.

79

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
δa, δf = difference between the semi-major radius and flattening respectively of the Clarke 1880
ellipsoid (Minna datum) and the WGS84 ellipsoid. From the computations carried out, the
following values were derived for Tx, Ty and Tz: Tx = 93.708m; Ty = 92.626m; Tz = -
121.330m. Minna Datum ellipsoidal heights were gotten direct from the WGS84 ellipsoidal
heights.

Applying least squares adjustment

The least squares solution of the unknown parameters (seven parameters) including three
translational, three rotational and one scale factor or the estimate of the correction to approximate
parameter vector (x) is given by Ẋ = - (ATPA)-1ATPL.

1 0 0 0 −(𝑍 − 𝑍𝑏) (𝑌 − 𝑌𝑏) (𝑋 − 𝑋𝑏)

𝐴 = [0 1 0 (𝑍 − 𝑍𝑏) 0 −(𝑋 − 𝑋𝑏) (𝑌 − 𝑌𝑏) ]
0 0 1 −(𝑌 − 𝑌𝑏) (𝑋 − 𝑋𝑏) 0 (𝑍 − 𝑍𝑏)
(20)

Where the X,Y,Z values are the Cartesian coordinates ranging from point 1 to 100. The value for
Xb, Yb and Zb is derived from the mean of the X,Y, and Z parameters in the local ellipsoid
(CLARKE 1880). This was computed using Microsoft Excel and the following values were
derived, Xb = 6274890.52, Yb = 942160.5278, Zb = 640909.2331. In this project, the parameters
were assumed to be weighted equally. Therefore, the value of P = I; L was developed using the
formula below,

𝑋𝐺 − 𝑋𝑏 − 𝑋𝐿 − 𝑋𝑏
[ 𝑌𝐺 − 𝑌𝑏 − 𝑌𝐺 − 𝑌𝑏 ] (21)
𝑍𝐺 − 𝑍𝑏 − 𝑍𝐺 − 𝑍𝑏

Where,
XG , YG , ZG = The X, Y and Z data from the global system
XL , YL , ZL = The X, Y and Z data from the local system
Xb, Yb , Zb = Mean of the X,Y,Z parameters from the local system.
This was computed using MATLAB to derive the 7 datum transformation parameters.

RESULTS

The computed values of datum transformation parameters are presented in Table 2.

Table 2: Datum Transformation Parameters/Std. Deviations for Cross River derived using
Molodensky-Badekas model (100 stations)

Parameters Estimated Values Approx. Std Dev.

ΔX(m) 99.344640001658689 99.3446 ± 2.433105012
ΔY(m) 14.883071267146580 14.8831 ± 2.445608309
ΔZ(m) -60.279356996844790 -60.2794 ± 2.433084462
Rx (o) -0.000000651559345 -6.51×10-7 ±0.000004385
Ry (o) 0.000021346404411 2.13×10-5 ± 0.00003947894
Rz (o) 0.000035924462804 3.59×10-5 ± 0.000074430417
K (ppm) -0.000069847626647 6.98×10-5 ± 0.000029443607
80

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
The variance covariance matrix (Table 3) was computed using the formula given in equations
12, 13 and 14. The degree of freedom is given by n – m.
Where,
n = number of observations and m = no of parameters
Since n = 100, m = 7
Therefore, df = 93
After computations, it was observed that;
A-posteriori variance scale = 5.919972057953653e+004
Using the "format short" command, the covariance matrix is given as;

Table 3: Covariance matrix

-5.9200 -0.0000 0.0000 -0.0000 -0.0000 -0.0000 -0.0000

0.0038 -5.9810 0.0000 -0.0000 -0.0000 -0.0000 -0.0000
0.0000 0.0000 -5.9199 0.0000 0.0000 0.0000 0.0000
-0.0000 -0.0000 0.0000 -1.9×10-11 -0.0000 -0.0000 -0.0000
-0.0000 -0.0000 0.0000 -0.0000 -1.6×10-9 -0.0000 -0.0000
-0.0000 -0.0001 0.0000 -0.0000 -0.0000 -5.54×10-9 -0.0000
-0.0000 -0.0000 0.0000 -0.0000 -0.0000 -0.0000 -8.67×10-10

The above set of seven parameter optimal transformation parameters for Cross River State were
computed using Molodensky-Badekas Model. All computations were carried out using
MATLAB. 100 common points in the Clarke 1880 and WGS84 were used for the initial
determination of the 3D Datum Transformation parameters for Nigeria. The results of the
computation shown above showed an acceptable estimate of the datum shift parameters using
the available data (dX, dY, dZ, K, RX, RY, RZ) and standard deviation of the parameters. The
computed residuals were fairly good. The result of the test computation of the shift parameters
using the entire 107 points were however not significantly different from those obtained with the
100 points.

Validation of points using the training samples

There was need to validate the parameters obtained from this project by testing them against the
seven coordinates which were specifically set aside for confirmation in the Clarke 1880
Ellipsoid. These points were carefully selected in order to spread across the entire state (North,
South, West, East and Centre) and they include; xsw148, xsw117, xsw126, xsw97, xsw82,
xsw64, xsw155 (Figure 3).

81

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)

Figure 3: Validation points

The formula given below was used for this exercise,

𝑋 𝛥𝑋 𝑋𝑀 1 + 𝛥𝐿 𝑅𝑍 −𝑅𝑌 𝑋′ − 𝑋𝑀
[𝑌 ] = [𝛥𝑌 ] + [ 𝑌𝑀 ] + [ −𝑅𝑍 1 + 𝛥𝐿 𝑅𝑋 ] [𝑌′ − 𝑌𝑀 ] (22)
𝑍 𝛥𝑍 𝑍𝑀 𝑅𝑌 −𝑅𝑋 1 + 𝛥𝐿 𝑍′ − 𝑍𝑀

Where,
𝑋𝑀 = 1⁄𝑛 ∑𝑛𝑖=1 𝑋𝑖 ,
𝑌𝑀 = 1⁄𝑛 ∑𝑛𝑖=1 𝑌𝑖
𝑍𝑀 = 1⁄𝑛 ∑𝑛𝑖=1 𝑍𝑖

Where,
n = number of common points
XM ,YM , ZM = the mean of the cartesian coordinates of common points in the local datum
(Minna)
X,Y,Z = Cartesian coordinates in the global datum
ΔX, ΔY, ΔZ = the translational parameters
82

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
RX , RY , RZ = the rotation parameter
ΔL = the scale factor
X',Y',Z' = Cartesian coordinates in the local datum
Table 4 is the expected coordinates in WGS 84 while Table 5 is the computed coordinates in
WGS 84.

Table 4: Expected coordinates in WGS84

Stn ID X Y Z
xsw148 6258493.911 969338.772 754436.0547
xsw117 6254271.032 1033136.284 703269.6455
xsw126 6273891.225 941839.4031 655283.4031
xsw99 6285872.13 875352.4007 632557.009
xsw82 6272979.892 973430.1241 616972.012
xsw64 6286219.834 923121.6107 556858.7991
xsw155 6288045.065 944929.2052 497179.467

Table 5: Computed coordinates in WGS84

Stn ID X Y Z
xsw148 6258274.547 969076.065 754386.435
xsw117 6254055.828 1032870.413 703228.420

xsw126 6273671.465 941577.586 655250.262

xsw99 6285648.958 875093.640 632527.719

xsw82 6272762.468 973166.709 616945.185

xsw64 6286000.561 922860.385 556842.017

xsw155 6287828.141 944666.833 497172.553

Comparison of the observed and computed WGS84 coordinates for the seven validation
points

Tables 6 to 12 expressed the difference between observed and computed WGS84 coordinates.

83

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
Table 6: Difference between observed and computed WGS84 coordinates of Point xsw148 for
Point xsw148

Latitude (Degrees) Longitude (Degrees) Results

6.8387293 8.80422641 Expected
6.839305608 8.802181377 Computed
-0.000576308 +0.002045033 Difference

Table 7: Difference between observed and computed WGS84 coordinates of Point xsw117

Latitude (Degrees) Longitude (Degrees) Results

6.372982095 9.379922272 Expected
6.373563455 9.377868247 Computed
-0.00058136 +0.002054025 Difference

Table 8: Difference between observed and computed WGS84 coordinates of Point xsw126

Latitude (Degrees) Longitude (Degrees) Results

5.936555594 8.537515166 Expected
5.937145463 8.535471400 Computed
-0.000589869 +0.002043766 Difference

Table 9: Difference between observed and computed WGS84 coordinates of Point xsw99

Latitude (Degrees) Longitude (Degrees) Results

5.729963949 7.927860033 Expected
5.730558281 7.925824104 Computed
-0.000594332 +0.002035929 Difference

Table 10: Difference between observed and computed WGS84 coordinates of Point xsw82

Latitude (Degrees) Longitude (Degrees) Results

5.588355912 8.820706812 Expected
5.588950061 8.818658260 Computed
-0.000594149 +0.002048552 Difference

Table 11: Difference between observed and computed WGS84 coordinates of Point xsw64

Latitude (Degrees) Longitude (Degrees) Results

5.042420957 8.354087155 Expected
5.043024196 8.352043680 Computed
-0.000603239 +0.002043475 Difference

84

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
Table 12: Difference between observed and computed WGS84 coordinates of Point xsw155

Latitude (Degrees) Longitude (Degrees) Results

4.500839505 8.546114313 Expected
4.501450148 8.544066797 Computed
-0.000610643 +0.002047516 Difference

Judging from the results of the validation which was done by transferring coordinates in the
global system (WGS84) to the local system (CLARKE 1880), the transformation model gives a
rough estimate based on the available data of the validation points xsw148, xsw117, xsw126,
xsw97, xsw82, xsw64, xsw155 in both horizontal and vertical plane and can be accepted for the
purpose of this project.

CONCLUSION

This study has estimated the optimal transformation parameters for Cross River State between
Minna and WGS84 Datums using Molodensky-Badekas Models. One hundred (100) of the One
hundred and seven (107) common point coordinates in Minna and WGS84 Datums provided
through secondary means were used for the computations of the parameters. From the results of
the validation, the transformation parameter determined in this study defines roughly the spatial
locations of the validation points (xsw148, xsw117, xsw126, xsw97, xsw82, xsw64, xsw155) in
horizontal plane even though the ellipsoidal heights are in variation to some meters. However,
these height disparities are likely to reduce when a local and more precise geoid model is in place
for Cross River State; from which subsequent undulations would be derived as inputs for future
versions of geodetic datum Transformations. Also, considering the size of the State, more
common point's data would perhaps produce better results in the future. Achieving better results
in subsequent determination would require many more common points than the present. Also,
continuously quality assurances of all geodetic coordinates before and after computations are
imperative in order to ensure internal accuracy and homogeneity. Noting that, the control points
used for the 7-paramter transformation in this work are acceptable as it did aid the accuracy
achieved.

This research makes the flowing recommendations:

i. Efforts should be made by the Office of the Surveyor General of the Federation to ensure
that other states in the country develops a suitable model for transformation as this will
have increased accuracy than relying on the parameters developed for the entire country
ii. The determination of precise geoid model for Nigeria is imperative.
iii. It is pertinent for other researchers to validate the findings of this study by using other
models such as the Bursa Wolf or the Veiss model.

REFERENCES
Agajelu, S. I., & Moka, E. C. (1989). An estimate of the transformation parameters for the minna
datum of the nigerian geodetic network. Australian Surveyor, 34(5), 494-501.
Ghilani, C. D. (2000). The surveying profession and its educational challenge. Surveying and
Land Information Systems, 60(4), 225-230.
Heiskanen, W. A., & Moritz, H. (1967). Physical geodesy. Bulletin Géodésique (1946-
1975), 86(1), 491-492.
85

ISSN 2054-6351 (print), ISSN 2054-636X (online)

 British Journal of Environmental Sciences
Vol.7, No.2, pp. 70-86, May 2019
Published by European Centre for Research Training and Development UK (www.eajournals.org)
Laundal, K. M., & Richmond, A. D. (2017). Magnetic coordinate systems. Space Science
Reviews, 206(1-4), 27-59.
Mahboub, V. (2012). On weighted total least-squares for geodetic transformations. Journal of
geodesy, 86(5), 359-367.
Moritz, H. (1980). Geodetic reference system 1980. Journal of Geodesy, 54(3), 395-405.
Ojigi, M. L., & Dodo, J. D. (2013). 3D Datum transformation between Clarke 1880 and
WGS84 Ellipsoids for Nigeria. Workshop on Adoption of Transformation Parameters for
Nigeria held on 2-3 July, 2013: NICON Luxury, Abuja.
Okeke, F. I. (2013). Coordinate Transformation Parameters and Procedure for Nigeria.
Workshop on the Adoption of Transformation Parameters for Nigeria held on 2-3 July,
2013: Nicon Luxury, Abuja.
Rollins, C., & Avouac, J. P. (2019). A geodesy‐and seismicity‐based local earthquake likelihood
model for central Los Angeles. Geophysical Research Letters.
Sella, G. F., Dixon, T. H., & Mao, A. (2002). REVEL: A model for recent plate velocities from
space geodesy. Journal of Geophysical Research: Solid Earth, 107(B4), ETG-11.
Shen, Y. Z., Chen, Y., & Zheng, D. H. (2006). A quaternion-based geodetic datum
transformation algorithm. Journal of Geodesy, 80(5), 233-239.
Trojanowicz, M., Osada, E., & Karsznia, K. (2018). Precise local quasigeoid modelling using
GNSS/levelling height anomalies and gravity data. Survey Review, 1-8.
Uzodinma V. N. (2005). Improved Minna Datum Geoidal Heights and Origin Definition from a
combination of Astrogeodetic, Geopotential, and satellite Doppler Data. A paper presented
at 2nd International Conference of Nigeria Union of Radio & Planetary Science (NUPRS),
Port Harcourt.
Uzodinma, V. N. (2013). Computation of the Minna B Datum Ellipsoidal Heights. Final
Workshop on the Adoption of Transformation Parameters for Nigeria held on 2-3 July,
2013: Nicon Luxury, Abuja.

86

ISSN 2054-6351 (print), ISSN 2054-636X (online)