Journal of Marine Science and Technology, Vol. 19, No. 1, pp.

35-42 (2011)

35

COMPUTATION PROGRAMS OF THE

ASTRONOMICAL VESSEL POSITION
Chih-Li Chen* and Tsung-Hsuan Hsieh*

Key words: intercept method, observed altitude, astronomical vessel position.

ABSTRACT
In open sea sailing, as opposed to sailing with satellite
navigation systems, the astronomical vessel position (AVP) is
not limited by military codes. If programs can be developed to
solve the AVP directly, drawbacks of current methods for
performing marine operations can be greatly improved.
Complete AVP computations have two necessary points:
compute the observed altitude and compute the AVP for two
bodies. Thus, this paper utilizes Matlab programming language to develop the observed altitude (ObsAltPro) and the
AVP by using intercept method (IMPro-2) programs, respectively. Adopting the ObsAltPro can avoid the use of nautical
almanacs and directly obtain the observed altitudes of various
celestial bodies between 1986 and 2050. As for merits of
using the IMPro-2, they can skip limits of tabular methods and
directly use the dead reckoning (DR) as the input variable to
replace the assumed position (AP). To address the deficiencies of the intercept method (IM), iteration computation is
added into the latter program as a verification function. The
above two programs, through demonstrated examples, have
shown advantages of being simple, fast, and accurate.

I. INTRODUCTION

Other softwares, NavPac developed by a British research organization [7], Navigator [12] and STARPILOT PC [13] developed by private companies, use the intercept method (IM).
However, they do not have the verification function of iteration
computation. This disparity between expectations and reality
leads to the motivation of this article.
The task in developing the AVP computational programs
is to understand the process. Due to the complicated astronomical algorithms that are involved with nautical almanacs,
this paper will not discuss this topic in detail. Other than that,
as the IM is the main way of the sight reduction methods,
which requires more than two lines of position (LOPs) to
obtain AVP this paper will develop the programs aiming at
the observed altitude of various celestial bodies, namely,
ObsAltPro and the intercept method for solving AVP for two
bodies condition, namely, IMPro-2, respectively as shown in
Fig. 1.
Apart from this introduction, the following sections are arranged as follows. Sections II organizes the computation
procedures of ObsAltPro and IMPro-2. Based on this, sections III uses MATLAB to develop the programs. Sections
IV then verifies the self-developed programs through demonstrated examples. Finally, sections V provides some concrete
conclusions.

II. ORGANIZING COMPUTATION

PROCEDURES

For ocean navigation, the main approaches of positioning are

the classical method of astronomical vessel position (AVP) and
advanced satellite navigation systems. Due to the complicated,
tedious and time consuming nature of current AVP and the
accuracy limitations due to tabular methods, the majority of
navigators prefer using satellite navigation systems. However,
the AVP has the advantage of not being limited by military
codes. If assisted by computerized programs, the practical
operation process can be greatly improved. Thus, in searching
for computation software for the AVP, it was discovered that the
software developed by the U.S. Navy is not purchasable [5].

1. Computation Procedures of Observed Altitude from

Sextant Altitude
This paper uses the nautical almanac tabular formulae [6] to
establish the computation procedures in conjunction with the
semi-diameter (SD) of Sun and the horizontal parallaxes (HP)
of the Moon, Venus, and Mars obtained from the works of Jean
Meeus published in 1998 [2, 3, 10]. These formulae are only
suitable for use between 1986 and 2050. The organized computation procedures are as follows:

Paper submitted 08/05/09; revised 12/18/09; accepted 12/23/09. Author for

correspondence: Chih-Li Chen (e-mail: [email protected]).
*Department of Merchant Marine, National Taiwan Ocean University, Keelung, Taiwan, R.O.C.

Dip = 0.0293 he ,

Step 1: Calculate the dip of the horizon (Dip).

where he is the height of eye above the horizon in meters.

(1)

Journal of Marine Science and Technology, Vol. 19, No. 1 (2011)

36

observers
position

observed
time

hs

moving
reference
point

nautical
almanac

ObsAltPro
(altitude correction
procedure)

(Dec, GHA)

Ho

where s denotes the Sun radius vector in astronomical

units (AU), and the Sun SD in arc minutes.
Moon: Computing the distance between the Moon and the
Earth using the ELP2000/82 method can obtain the radius
vector of the Moon (m). Then, the Moon HP is calculated
in Eq. (6) [3], and substituting the HP into Eq. (7) yields the
Moon semi-diameter (SD),
6378.14
,
m

(6)

SD = 0.2724 HP.

(7)

sin HP =
IMPro-2
(included iterative function)

Where m denotes the Moon radius vector in kilometers,

the Moon HP in degrees, and the Moon SD in arc minutes.
Navigational stars and planets: the SD is ignored due to the
center of their observed bodies is easier to determine.

AVP
Fig. 1. Computational programs of the AVP.

Step 2: Calculate the apparent altitude (ha) of the observed

body.

Step 5: Calculate the parallax (P), from the HP and the ha, the
formula is as follows:

ha = hs I IC Dip,

P = HP cos(ha ).

(2)

where hs is the sextant altitude, I is the sextant instrument

correction, and IC is the sextant index correction.
Step 3: Calculate refraction (R).
The refraction formula at a standard temperature of 10
Celsius (C) and standard pressure of 1010 millibars (mb) is
as follows:
RO =

0.0167
,

7.32
tan ha +

ha + 4.32

(3)

0.28 B
R=
RO ,
T + 273

Sun: Substituting the Sun HP 0.0024 in Eq. (8) to calculate

the Sun P.
Moon: After obtaining the Moon HP through Eq. (6), substituting the HP into Eq. (8) yields the Moon P.
Venus and Mars: After computing the distances between
Venus and the Earth, Mars and the Earth, with the VSOP 87
method, respectively, one can obtain the radius vector of
Venus or Mars (st). Then, the HP is calculated in Eq. (9)
[3]. Substituting the HP into Eq. (8) yields the P of Venus
or Mars.
sin HP =

in which RO represents refraction for standard condition.

If the temperature TC and pressure B mb are known, then
the refraction formula is as follows:
(4)

(8)

sin 8.794
,
st

(9)

where st denotes the radius vector of Venus or Mars in

kilometers, and the Venus or Mars HP in degrees.
Jupiter, Saturn, and navigational stars: the P is ignored due
to their far distance from the Earth.
Step 6: Calculate the observed altitude (Ho).

in which R represents refraction for non-standard atmospheric conditions.

Step 4: Calculate semi-diameter (SD).
Sun: Computing the distance between the Sun and the Earth
using the VSOP 87 method, namely obtaining the radius
vector of the Sun (s), and then further calculating the Sun
semi-diameter in Eq. (5) [2].
SD =

959.63
,
s

(5)

Ho = ha R SD + P,

(10)

where Ho represents observed altitude.

2. Computation Procedures of AVP by Using the Intercept
Method for Two Bodies
The basic idea of the IM is to choose the assumed position
(AP) at the nearby the dead reckoning (DR) position, and take
it as the reference point to calculate the altitude and azimuth.
By comparing the computed altitude with observed altitude,

C.-L. Chen and T.-H. Hsieh: Computation Programs of the Astronomical Vessel Position

DR

DRL

observed
time
DR

tan Z =

hs

GMT
DRObsAltPro
(altitude correction
procedure)

nautical
almanac
GHA

Dec

LHA

navigational spherical
triangle
Z

Ho

Hc

Zn

LOP
Fig. 2. The flowchart for solving the LOP by using the intercept method.

the difference of two altitudes (called the intercept, a) can be

obtained. Therefore, once the AP, the computed azimuth angle
(Z) of the body, and the intercept are all determined, the LOP
can be plotted according to the three elements. Hence, the IM
is essentially a kind of trial-and-error method. If the computation method is adopted, the choice of the initial position, such
as the DR position in place of the AP, can be unconstrained.
The flowchart for solving the LOP by using the IM is shown in
Fig. 2. It should be note that an observed bodys position variables in celestial equator coordinate, such as the declination
(Dec) and the Greenwich hour angle (GHA), can be obtained
at observed time by using the nautical almanac or related
software. Besides, a celestial bodys local hour angle (LHA) is
the algebraic sum of its GHA and the observers longitude ().
Computing an observed bodys position variables in celestial horizon coordinate, such as the computed altitude (Hc) and
azimuth angle (Z), the IM is a navigational spherical triangle
problem. That is, when the two sides (co-latitude and polar
distance) and the included angle (LHA) of an oblique spherical
triangle are given, the problem is to solve the third side
(co-altitude) and the outer angle (Z). When the simplicity and
minimum error propagation are taken into account in the
computation, the optimal formulae are the side cosine and the
four-parts formula [8, 9]. The formulae are as follows:
sin Hc = sin( DRL) sin( Dec) + cos( DRL) cos( Dec) cos LHA,

(11)

sin LHA
,
[cos( DRL) tan( Dec)] [sin( DRL) cos LHA]

37

(12)

in which DRL is the DR latitude, Dec is declination, LHA is

local hour angle, Hc is computed altitude, and Z is azimuth
angle.
Some important issues should be noted here. The first is
that in the above equations if sign of Dec is the same as DRL, it
is treated as a positive quantity. On the contrary, it is treated as
a negative quantity. The second is that the side cosine formula can be reduced to Eq. (11) because sine functions are odd
and cosine functions are even. In addition, if the computation
method is used, the accuracy of the obtained AVP can be increased and verified by the iteration computation [4, 8].
Basically, the classical problem of celestial navigation, in
its simplest form, is the determination of an AVP; one intersection is resulted from the different LOPs of two celestial
bodies, observed at known GMT on a known date. However,
the most case is observing altitudes of the same or different
celestial bodies at different time [14]. Thus, the running fix
concept is usually adopted and after the course, speed, and
period of two observing times have been identified. To this
point, for complete computation procedure, the moving reference point, should be taken into consideration. This is
explicated below.
In order to move the circle of position (COP) or LOP to the
same time, the running fix concept to do so is to use the rhumb
line sailing in conjunction with the moving reference point, the
geographical position (GP) of the COP or the DR of the LOP,
is adopted [1, 5, 8]. The reason for obtaining a celestial fix is
that each would have to be advanced or retired to the desired
time for celestial fix, and making proper allowance for the
travel of the vessel during the intervening time. The computation procedures are as follows:
Step 1: Calculate the moving distance (d) and difference of
latitude (l) by using
d = s t,

(13)

l = d cos C ,

(14)

and

where s is speed, t is time interval, and C is course angle.

Step 2: Calculate the new reference point latitude (L2) by
using
L2 = L1 l ,

where L1 is the reference point latitude.

Step 3: Calculate the meridional difference (m) by using

(15)

38

Journal of Marine Science and Technology, Vol. 19, No. 1 (2011)

L
1
e
sin
L

1
1
ML1 = a ln tan 45 +
,
2 1 + e sin L1

Start

(16-a)
Date, Time, I, IC, he, hs, TB

L2 1 e sin L2 2

ML2 = a ln tan 45 +
, (16-b)
2 1 + e sin L2

ha = hs I IC (0.0293 he)

RO =

and
m = ML1 ML2 ,

0.0167

7.32
tan ha +
ha + 4.32

No

(16-c)

TB == []

0.28 B
R=
RO
T + 273

where a is the equators arc minute length, e is earth flattening,

ML1 is the meridional parts at L1, ML2 is the meridional parts
at L2.
Step 4: Calculate the new reference point longitude (2) by
using
DLo = m tan C ,

(17)

2 = 1 DLo,

(18)

where DLo is difference of longitude, and 1 is the reference

point longitude.

Yes

R = RO

Bodies
Sun

SD =

959.63

Venus and
Mars

Moon

sin HP =

6378.14

sin HP =

SD = 0.2724 cos HP
P = 0.0024 cos(ha) P = HP cos(ha)

However, as Eq. (17) can be seen, when the course angle is

90, tan C = , and the DLo can not be calculated. In this case,
the parallel sailing should be used for calculating the DLo.
The formula is as follows:

Jupiter, Saturn
and Stars

sin 8.794

SD = 0
P = 0

SD = 0
P = HP cos(ha)

Ho = ha R SD + P
Ho, ha

DLo = d sec L,

(19)
End

where L is the reference point latitude.

It is worth mentioning, that because of an irregular oblate
spheroid of the earth, parameters of meridional parts are taken
from the 1984 world geodetic system (WGS-84) [11], in which
21600
1
a=
, e = 2 f f 2 , and f =
,
2
298.257223563
The parameters of the Global Positioning System (GPS)
and Electronic Chart Display and Information System (ECDIS)
both adopt WGS-84; therefore the applications of developed
program in this paper are quite broad.

Fig. 3. The flowchart of the ObsAltPro program.

develop the IM program for two bodies, named IMPro-2, and

the iteration computation function is provided for the accuracy of the AVP shown in Fig. 4. Moreover, to ensure the
completion of the IMPro-2, moving reference point computation is included as shown in Fig. 5. For friendly use, the above
two computation programs were both developed using the
graphical user interface (GUI) of the MATLAB programming language.

III. DEVELOPING COMPUTATION PROGRAMS

As explained in Sections II, this paper uses computation
procedures of observed altitude from sextant altitude as the
basis in developing the observed altitude program of various
celestial bodies between 1986 and 2050, named ObsAltPro
and its flowchart is shown in Fig. 3. As for computation procedures of the AVP, the IM method is used as the basis to

IV. DEMONSTRATED EXAMPLES

Example 1: At 10-00-00 GMT (07-00-00 local time), June 16,
1994, the navigator obtains a sight of the Moons upper limb.
The sextant altitude (hs) is 206.7'. Height of eye is 18 feet;
there is no index error. Determine the observed altitude (Ho)
of the Moon [1].

C.-L. Chen and T.-H. Hsieh: Computation Programs of the Astronomical Vessel Position

Start

39

Start

A
Ho1, Ho2

Dec1, Dec2

GHA1, GHA2

L1, 1, Cn, s, t

DR1, DR2
DR1
DR2

Transform Cn
into C

DRL1
DRL2

LHA1, LHA2

Yes

Call the function to

compute Hc and Z

C = 90

No

L2 = L 1

d=st

d=st

l = d cosC

DLo = d sec L1

L2 = L1 l

2 = 1 DLo

L2 1 e sin L1 2

ML1 = a ln tan 45 +

2 1 + e sin L1

tab229_intercept
Hc1, Hc2, Z1, Z2

Call the function to

solve the intersection of
LOPs and bearing lines
Call the function to
determine the
boundary of the LOP

Hc1, Hc2

Z1, Z2

a1, a2

Zn1, Zn2

msplot

mapselect

Call a function to
determine the boundary
of Mercator chart

L2 1 e sin L2 2

ML2 = a ln tan 45 +

2 1 + e sin L2

msplot

Call the function to

determine the AVP
navfix
AVP
AVP

AVP ==DR

m = ML1 ML2
No

Iteration

DLo = m tanC

Yes
Iteration Fix

2 = 1 DLo

End

Fig. 4. The flowchart of the IMPro-2 program.

L2, 2
End

Required: The observed altitude can be determined in the

following two approaches for sight reduction.
Using the nautical almanac to solve the observed altitude.
Using developed program ObsAltPro to solve the observed
altitude directly.
Solution:
Extracted from American Practical Navigator (No. 9), result of the Moon observed altitude is 2637.1' [1].
Executing ObsAltPro shows the Moon observed altitude to
be 2637.1', as shown in Fig. 6.
Explanation: For the same results, this example is verified
that self-developed program is correct. Moreover, the ObsAltPro program, directly performs calculations using formulae,
adopts GUI expression to demonstrate the advantages of being
simple, fast and accurate. Besides, computerized program is
more versatile than tabular method due to its design. That is, the
ObsAltPro program does not require the use of nautical almanacs, and using this program can directly obtain the observed
altitude of various celestial bodies between 1986 and 2050.

Fig. 5. The flowchart of moving reference point function in IMPro-2.

Example 2: On June 16, 1994, at 05-15-23 local time, at DR

position L 30 N, 045 W, the navigator already took a sight of
the Suns upper limb. The ship is keeping on course 030 and
speed 10 knots. Again, at 07-00-00 local time, another sight of
the Moon was obtained, as explicated in Example 1. Thus, the
navigator wants to adopt the running fix concept to determine
0700 celestial fix, namely AVP. Therefore, the navigator records the needed information and further reduces it from the
nautical almanac for sight reduction as shown in Table 1 [1].
Required: The AVP can be determined by using self-developed program, IMPro-2, to solve the AVP.
Solution: First, use the moving reference point function of
IMPro-2 to obtain the post-moving reference position L
3015.1' N, 04450.0' W, as shown in the upper-left of Fig. 7.
Then, treat the obtained position as the DR position of the
panel, and enter the needed information, such as the Dec,

Journal of Marine Science and Technology, Vol. 19, No. 1 (2011)

40

Table 1. Extract of relevant information for sight reduction

in example 2.
Body

ZT

Ho

Dec

GHA

Sun

05-15-23

248.1'

2320.5' N

30342.1'

Moon
07-00-00
2637.1'
013.8' N
34615.0'
Source: [1] The American Practical Navigator, pp. 303-307 (2002).

Table 2. Extract of relevant information for sight reduction

in example 3.
Body

ZT

Ho

GP

Alkaid

20-03-06

7734.9

4925.7 N

00314.2

Capella

20-04-08

1519.3

4558.4 N

13124.8

Source: [5] Duttons Nautical Navigation, pp. 332-335 (2004).

Fig. 7. Using the IMPro-2 to solve the AVP in example 2.

Fig. 6. Using the ObsAltPro to solve the observed altitude in example 1.

GHA, and Ho, of the two celestial bodies (The Sun and the
Moon), as shown in the lower-left of Fig. 7. Clicking the
computing function can obtain an AVP of L 3048.4' N,
04458.0' W. Further, clicking the iteration function can yield
the accurate AVP of L 3048.5' N, 04458.0' W, as shown in
the lower-right of Fig. 7.
Explanation: This designed example is only demonstrated
how to use IMPro-2 to solve the AVP.
Example 3: The 2004 DR position of a vessel is L 4134.8' N,
01700.5' W. At 20-03-58, the star Capella is observed with
a sextant. At 20-02-56, shortly before the above observation,
another star, the Alkaid is spotted. The navigator records the
needed information and further reduces it from the nautical
almanac for sight reduction as shown in Table 2 [5].
Required: The AVP can be determined in the following three
approaches for sight reduction.
Using the IM in conjunction with the inspection tables (No.
229) to solve the AP, Zn and a, and plot the LOP. (Tabular
method)

Treating AP as the initial reference point, using developed

program IMPro-2 to solve the possible AVP and verify it.
(AP to run IMPro-2)
Treating DR as the initial reference point, using developed
program IMPro-2 to solve the possible AVP and verify it.
(DR to run IMPro-2)
Solution:
Extracted from Duttons Nautical Navigation, results of AP,
Zn and a are listed in Table 3 and the graphical AVP is
L 4138.6' N, 01708.1' W [5].
AP was treated to run IMPro-2, AVP is L 4139.0' N,
01707.6' W. Then, executing the iteration function yields
fixed AVP is L 4139.1' N, 01707.3' W, as shown in Fig. 8.
DR was treated to run IMPro-2, AVP is L 4139.1' N,
01707.3' W. Then, executing the iteration function yields
fixed AVP is L 4139.1' N, 01707.3' W, as shown in Fig. 9.
Discussion: In actuality, this example is validated and the
accurate AVP is pointed at L 4139.1' N, 01707.3' W from
the paper of Hsu, T.P. et al. [8]. Consequently, it is demonstrated that DR running IMPro-2 can obtain the true AVP,
more accurate than tabular method or AP running IMPro-2.
However, executing the iteration function showed that, regardless of whether AP running IMPro-2 or DR running

C.-L. Chen and T.-H. Hsieh: Computation Programs of the Astronomical Vessel Position

41

Table 3. Three elements for plotting the LOPs by using

the intercept method (the tabular method) in example 3.
Body

The three plotting elements of LOP

Alkaid

42 N
AP
01714.2W

Zn = 047.9
a = 10.4 Away

Capella

42 N
AP
01724.8W

Zn = 318.8
a = 24.2 Away

Source: [5] Duttons Nautical Navigation, pp. 332-335 (2004).

Fig. 9. Treating DR to run IMPro-2 for solving the AVP in example 3.

IMPro-2 not only directly uses the DR as the input variable

to replace the AP but also adds an iteration function to
verify the determined AVP.
For complete solving the classical problem of celestial
navigation, the moving reference point function should be
taken into account. The developed IMPro-2 adopted rhumb
line sailing in conjunction with parallel sailing to deal with
this problem. Again, for conventional user, the developed
IMPro-2 can also demonstrate the plotting figure to help
illustrate the result.
Fig. 8. Treating AP to run IMPro-2 for solving the AVP in example 3.

ACKNOWLEDGMENTS
IMPro-2, the post-iteration AVP always is L 4139.1' N,
01707.3' W, as shown in Figs. 8 and 9, further demonstrating
the accuracy of AVP. Besides, IMPro-2 can also demonstrate
the manual plotting process; the figure in the upper-right part
of this GUI, as shown in Figs. 8 and 9, can be expanded,
outputted, and saved.

V. CONCLUSIONS
This paper proposes two computerized programs, ObsAltPro and IMPro-2, allowing users to easily and quickly complete computations of the AVP. Actually, the computerized
solution is always more it avoids than tabular methods because
it is free of rounding error. The above two programs, through
demonstrated examples, have shown advantages of being
simple, fast, and accurate. In summary, the characteristics of
these two computation programs as are follows:
Running ObsAltPro does not require information of nautical almanac at all, thereby it avoids its limitations and errors.
The observed altitudes of various celestial bodies from
1986 to 2050 can be directly obtained.
Due to the inherent nature of the IM, a trial-and-error method, and improving drawbacks of the IM. The developed

The first author would like to thank the National Science

Council, Taiwan for financial support under contract number:
NSC 96-2628-E-019-022-MY3.

REFERENCES
1. Bowditch, N., The American Practical Navigator, 2002 Bicentennial Ed.,
National Imagery and Mapping Agency, Bethesda, Maryland, U.S.A.
(2002).
2. Bretagnon, P. and Francou, G., Planetray theoris in rectangular and
spherical variables. VSOP 87 solutions, Astronomy and Astrophysics,
Vol. 202, pp. 309-315 (1988).
3. Chapront-Touz, M. and Chapront, J., The lunar ephemeris ELP 2000,
Astronomy and Astrophysics, Vol. 124, pp. 50-62 (1983).
4. Clough-Smith, J. H., An Introduction to Spherical Trigonometry, Brown,
Son & Ferguson Ltd, Glasgow, Scotland, U.K. (1966).
5. Cutler, T. J., Duttons Nautical Navigation, 15th Ed., Naval Institute Press,
Annapolis, Maryland, U.S.A. (2004).
6. Her Majestys Nautical Almanac Office, Rutherford Appleton Laboratory
and the Nautical Almanac Office of the US Naval Observatory, 2008
Nautical Almanac, Commercial Ed., Paradise Cay Publications and Celestaire Inc., Taunton, Somerset, U.K. (2007).
7. Hohenkerk, C. Y. and Yallop, B. D., NavPac, Version 2.1.2, Council for
the Central Laboratory of the Research Councils, Taunton, Somerset, U.K.
(2004).
8. Hsu, T. P., Chen, C. L., and Chang, J. R., New computational methods
for solving problems of the astronomical vessel position, The Journal of

42

Journal of Marine Science and Technology, Vol. 19, No. 1 (2011)

Navigation, Vol. 58, No. 2, pp. 315-335 (2005).

9. Maritime Safety Information Center, Sight Reduction Tables for Marine
Navigation, Pub. No. 229, National Imagery and Mapping Agency, Bethesda, Maryland, U.S.A. (1981).
10. Meeus, J., Astronomical Algorithm, Willmann-Bell, Richmond, Virginia,
U.S.A. (1998).
11. National Geospatial-Intelligence Agency, WGS 84 Earth Gravitational
Model, (http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/

egm08_wgs84.html) Retrieved Jul. 9th, 2009.

12. Reis, O., Navigator, Version 4.0, Celestaire, Wichita, Kansas, U.S.A.
(2005).
13. STARPILOT, LLC, STARPILOT PC, Maryville, Tennessee, U.S.A.
(2002).
14. Van Allen, J. A., An analytical solution of the two star sight problem of
celestial navigation, NAVIGATION, Journal of Institute of Navigation,
Vol. 28, No. 1, pp. 40-43 (1981).