Tides, Tide Tables, Tidal Predictions Tides, Tide Tables, Tidal Predictions

Stormy Weather SoftWare

[Tidal Prediction Software]  [Stormy Weather Homepage]

TIDAL PREDICTION PROGRAMME NOTES

These notes are purely technical, will probably interest very few people, prove uninteligible to non - mathematicians and therefore might never be read. They are included for completeness of our Tidal Prediction SoftWare

Some Manufacturers' Trade Marks are included (Microsoft, OS/2, Apple etc) and should be recognized as such.

HISTORY.

Version 1 was a virtually unusable set of math routines, based on various source documents, mainly by Doodson. User friendliness was non-existant, but the results were usable, demonstrating the validity of "new" algorithms that we had developed specifically for computerised solutions after being prompted by Mike Harris (Pangolin Software) looking for simple but fast results from programmable calculators and very small computers.

The new algorithms were adapted for use on PC type computers running a minimum of DOS 3.1, 80286 CPU and .5 Mb hard disk space available.

Version 2 developed ease of use, after the decision was made to use sets of constituents containing a maximum of ten harmonics with the possibility of inferring six more. Seven nodal corrections were used. Version 2.1 followed a number of most useful contacts with the Canadian Hydrographic Service, and Version 2.2 (the latest fully usable version as of November 1995) responded to comments made by user/mariners to whom working copies had been made available.

Accuracy, compared to published tide tables, is very acceptable; for semi-diurnal and diurnal tides where little shoal water effect is felt errors rarely exceed 2% in height and 5 minutes in time. However mixed tides in estuaries can produce errors up to 8% in height and 15 minutes. Average errors are obviously lesser, and the majority of time errors occur at stands.

Speed was improved to the point that a complete month of High/Low predictions was tabulated, ready to print, and produced graphically on screen within 15 seconds on a 286, and nearly instantaneously on a 486 DX (the video driver being a major "slow" point!)


It has become evident that further development in accuracy is not possible without further harmonic input; also 32 bit CPU technology is developing so fast (OS/2, Windows 95, etc) that programming should now be orientated towards capability of compiling for DOS as well as 16 and 32 bit MS Windows. Minor variations for OS/2 are easily incorporated at compilation time. Current programming in QuickBasic 4.5, Basic 7.0 and VisualBasic 3.0 is adaptable to these ends. Investigations into a Macintosh/Apple version continue. If a C++ version becomes necessary, it will demand a complete rewrite. Fortran has been considered for the math routines, but the interface with other languages poses considerable difficulty.

VERSION 3.x

(References in this section are, alphabetically
AD A.T. Doodson
BA British Admiralty
CH Canadian Hydrographic Service
DC D.E. Cartwright and R.J. Tayler
GG Gabriel Godin, as quoted by MF, his works being out of print
MF Mike Foreman, Institute of Ocean Sciences, Sidney, BC
SH Service Hydrographique et Océanographique de la Marine)


As planned, and partially developed already, the next version will aim for increased accuracy and ease of use in formats compatible with DOS, OS/2 Win 3.x and 95. Speed may decrease somewhat on older machines, and total programme size (*.exe plus data files) will increase to approx .7 Mb, but users, so far, have not made adverse comments on these matters.

Major changes include:



Data File Structure:

To include IHO reference system and 37 constituents (see below) in binary format. A fixed length record of 184 Bytes is required for each station. This is just about equal to v2.x ASCII records containing only 10 constituents. This is operational, together with utilities to convert other formats (Canada, US, UK, France) automatically.

Astronomical Data:

A complete new utility has been written based J2000. Its accuracy, after comparison with the Astronomical Almanach, is far superior to tidal requirements.
Using a fixed reference date (1 Jan 1980 for v2.1 or 1 Jan 2000 for v2.2) with daily increments can, due to variations in orbital phenomena, plus Delta T to translate from TBT to UT, lead to errors well in excess of 1°. It seems a shame to use harmonic data correct to 0.1° under these circustances, even if the original analysis contained the same degree of uncertainty.
Therefore it is proposed that daily increments from 0 hours, 16th day of each month be used. An added bonus is that single precision (7 sig figs) can be used in lieu of double (15 sig figs).
Either a data file containing say 1995 to ? could be used, or the Tide programme could access the astro utility directly.

Constituents:

A total of 37 constituents can be used if available. See Appendix 1. The choice of 4 third diurnals and 3 sixth diurnals represents a major departure from v2.x. MF reviewed the list and suggested the inclusion of M3. Eighth diurnals and further have been neglected, therefore double highs (ex Southampton) may still not be quite accurate. 16 satellites (nodal corrections) have been retained at this point, pertaining to the main constituents. Selection was based on the Cartwright coefficient being greater than 0.1. Third order harmonics have been avoided where possible (quarter phase shifts and corrections for latitude), negative amplitudes can be corrected by a half phase shift.
All Doodson numbers (s, h, p, N and p') are corrected by adding {Species x ( -s +h )} (Species being 0 for long term, 1 for diurnals, .... 6 for sixth diurnals); this facilitates conversion from Tau (lunar day) to UTC in later calculations.

Formulae:


(1) ASTRO Rebased 1/1/2000 (Brackets 1980) where NumDay is the number of days (+ or -) from reference date and the function Modulus 360 maintains orbital data in range 0° to 360°

s = FNMod360#(211.72826# + 13.17639673# * NumDay&) 'Long moon (78.15194)
h = FNMod360#(279.86807# + .98564734# * NumDay&) 'Long Sun (279.72234)
p = FNMod360#(83.29754# + .11140408# * NumDay&) 'Perigee (349.49439)
N = FNMod360#(234.98196# + .05295392# * NumDay&) 'RA Moon (208.15463)
p' = FNMod360#(282.93732# + .000047069# * NumDay&) 'Perihelion (282.59344)

the result being for Time = 0 hours on NumDay (T0), partial times if required are found by linear interpolation; T24 = T0 on NumDay + 1

(2) SUM (Sigma) OF CONSTITUENTS by species where A = Amplitude, g = total phase lag at T0, repeated for T24.

a) for each species (0, 1 .....6), starting SigmaX = 0, using major constituents and satellites

SigmaCosine = SigmaCosine + A * Cosine(g)
SigmaSine = SigmaSine + A * Sine(g)

b) reduce to average harmonic for species where Ex is amplitude, Phi is phase lag

Ex = SquareRoot ( SigmaSineý + SigmaCosineý )
[If Ex = 0 OR Ex = SigmaCosine then Ex = Ex + 10 ^ -10 to avoid division by 0]
Phi = ArcCosine(SigmaCosine / Ex) [* -1 if SigmaSine is negative]

We now have a series of "Average" amplitudes aand phases for each species at Time = 0 and Time = 24 for the day: Ex0 and Ex24, Phi0 and Phi24 BY SPECIES

3) HEIGHT OF TIDE at any Time (TX)

Let suffixes 0, X and 24 represent Time in hours on any date, and maintaining -180° < (Phi24 - Phi0) < 180°, then

  • ExX = Ex0 + [TX/24 * (Ex24 - Ex0)]
  • PhiX = Phi0 + [TX/24 * {(Species * 360°) + (Phi24 - Phi0)}]
  • HeightX = Zo + Sum of [ExX * Cosine (PhiX)] for species 0 ... 6
Note that this approach gives Error = 0 at 0 hours and Error = Maximum at 24 hours, so that if TX > 12 it would appear prudent to work back from Ex24 and Phi24; no noticeable improvement has been obtained.

4) EXTREMA found by using derivative for slope = 0 of type Sine/Cosine. The value of the derivative multiplied by K = -12.42 or -24.84 / 2 * Pi can be used to approach the absolute value of the Timestep (see 5 below)

Using Formula Height in 3) as FHtCos, and rewriting with Sine in lieu of
Cosine as ............... FHtSin

d = K * { FHtSin(Species1) + 2*FHtSin(Species2) ....... 6*FHtSin(Species6)}
_____________________________________________________________________
{ FHtCos(Species1) + 4*FHtCos(Species2) ....... 36*FHtCos(Species6)}

Note that integers have been used, but could be refined using average angular speed.

5) TIME STEPS for extrema are currently 12.42, 6.21 and 1.03 hours respectively for diurnal, semidiurnal and mixed tides, and works well as a "DeltaTime", to or from which the derivative above is added or subtracted as refinement in a recursive loop. The number of recursions averages 3.7 for a final "DeltaTime" < 1 minute. This will be modified to respect GG, and thus to ensure compatibility with CHS tables.

6) Further precision can be obtained (at the expense of computer time) by using times of extrema from 4), and returning for new values in steps 1) through 4) of each extremum. This does not appear to be necessary for pure diurnal or semidiurnal tides, but does show an improvement for mixed tides, particularly those that within a calendar month produce a mix of two or four extrema per day. However further investigation is necesary, above all in assessing the influence of shoal water effects.

7) Stands and double tides. These are currently assessed on the basis of height alone, and not recorded as extrema although they appear in the graphical daily representation (heights calculated for every 0.25 hours). The programme has now been modified so that 6 extrema per day can be recorded and will be refined, again along the lines suggested by GG, so that time as well as height (proposed as 0.01 * Z0) will be considered. Preliminary trials using Trois Rivières looks promising.

Appendix 1: Constituents



Each file record contains the following:
  • PortName: alpha name from catalogue (STRING * 22)
  • PortNumber: IHO format (LONG integer) format:-
  • 1 2345 67890 1 2
  • Q LATD LONGD A N where Q = quadrant, NE= 1, NW=2, SW=3, SE=4
  • A = Analyisis# for SAME station
  • N = DIFFERENT Station same one mile square
  • TimeZone: West - Negative. Because of half hours, TZ = input * 2
  • For .SGx files, change sign: TZ = input * -2
  • Data Structure for Constituents: Ax = Amplitude in mm
  • gx = degrees * 10

Doodson Numbers given as programmed (CHS, MF equivalents in brackets difference being {Species * (s - h)}
Angular speed in degrees/solar hour (°/hr) to 7 sig figs

Z0 - Mean level - Constant

Long Period, total = 5, Species 0:

Sa h -p' h -p' 0.04106664 see NOTE 1
Ssa 2h 2h 0.08213728
Mf 2s 2s 1.090833
Mm s -p s -p 0.5443747
MSf 2s -2h 2s -2h 1.0159

Diurnals, total = 12, Species 1:

Sig1 4s +3h -90° -3s +2h -90° -12.9272
Q1 -3s +h +p -90° -2s +p -90° 13.39867
J1 s +h -p -270° 2s -p -270° 15.58544 AD => s +h -p +90°
OO1 2s +h -270° 3s -270° 16.1391
NO1 -s +h +p -270° p -270° 14.4967
2Q1 -4s +h +2p -90° -3s +2p -90° 12.85421
Rho1 -3s +3h -p -90° -2s +2h -90° 13.47151
P1 -h -90° s -2h -90° 14.95893 SH -h +90° Amplit * -1
S1 p' -270° s -h +p' -270° 15.0
Pi1 -2h +p' -90° s -3h +p' -90° 14.91786
O1 -2s +h -90° -s -90° 13.94304
K1 h -270° s -270° 15.04107 was +90°
Not included: Chi1 Psi1 Phi1 The1 Tau1 SO1 Alp1 Bet1 Ups1 H1 M1(if NO1 is inc)
Repeats: O1=MK1 K1=MO1=SP1 SIG1=SIGMA1 Q1=NK1 J1=MQ1 M1=NO1 P1=SK1 MP1=TAU1 KQ1=UPS1 M(SK)2=H1
? Series1 ? M(SK)2=H1

Semi Diurnals, total = 9, Species 2:

N2 -3s +2h +p -s +p 28.43973
Mu2 -4s +4h -2s +2h 27.96821
L2 -s +2h -p -180° s -p -180° 29.52848 SH -s +2h -p Amplit * -1
T2 -h +p' 2s -3h +p' 29.95893
2N2 -4s +2h +2p -2s +2p 27.89535
Nu2 -3s +4h -p -s +2h -p 28.51258
K2 2h 2s 30.08214
M2 -2s +2h none 28.9841
S2 none 2s - 2h 30.0
Not included : Lda2 OQ2 R2 H2 Gam2 Eps2
P2 MKS2 MSN2 Eta2 2SM2 OP2 2SM2
Repeats: M2=KO2 S2=KP2 N2=KQ2 L2=2MN2=L2 2MN2(Fr) 2N2=2MS2
KJ2=ETA2 MNS2=UPS2 M(KS)2=H2
? Series2 ? 2MN2S2=ST36 M(SK)2=H1

Third Diurnals, total = 4, Species 3:

M3 -3s +3h -180° -180° 43.47616 AD M1+M2 = same +90° NOTE 2
MO3 -4s +3h -90° -s -90° 42.92714
SO3 -2s +h -90° s -2h -90° 43.94304
MK3 -2s +3h -270° s -270° 44.02517 AD M2 + K1 = same +90°
Not included: SK3
Repeats: MO3=2MK3 MQ3=NO3

Quarter Diurnals, total = 4, Species 4:

M4 -4s +4h none 57.96821
MS4 -2s +2h 2s -2h 58.9841
MN4 -5s +4h +p -s +p 57.42383
MK4 -2s +4h 2s 59.06624
Not included SN4 S4 SK4

Sixth Diurnals, total = 3, Species 6:

M6 -6s +6h none 86.95231
2MS6 -4s +4h 2s -2h 87.96821
2MK6 -4s +6h 2s 88.05035
Not included: MSN6 2SM6 MSK6 S6 2MN6
No others e.g. M8
Repeats:? Series6 ? 2MN2S2=ST36

Satellites:

Coeff > 0.1 included 11/95
(Doodson Numbers for p, N, p' to be added to astro argument of the major constituent, Cartwright coefficient applied to Amplitude of major constituent)
2Q1 0 -1 0 0.1885
Sig1 0 -1 0 0.1884
Q1 0 -1 0 0.1884
RHO1 0 -1 0 0.1882
O1 0 -1 0 0.1885
NO1 0 1 0 0.2004
NO1 -2 0 0 0.3596
S1 0 0 -2 0.3534
K1 0 1 0 0.1356
J1 0 1 0 0.1356
OO1 -2 0 0 0.1496
OO1 0 2 0 0.1342
OO1 0 1 0 0.6398
L2 2 1 0 +180 0.1102
L2 2 0 0 +180 0.2505
K2 0 1 0 0.2980

Satellites coeff > 0.035 or Latitude dependant NOT included 11/95
NO1 -1 0 0 +90° 0.2227 Latitude dependant
2Q1 -1 0 0 +270 0.0607 " "
Rho1 2 0 0 +180 0.0576
J1 1 0 0 +90 0.0816 " "
2N2 0 -1 0 +180 0.0374
2N2 -1 0 0 +90 0.0678 " "
Mu2 0 -1 0 +180 0.0375
N2 0 -1 0 +180 0.0373
Nu2 0 -1 0 +180 0.0373
L2 0 -1 0 +180 0.0366
M2 0 -1 0 +180 0.0373
S2 0 +1 0 0.0811 SH; includes +2h!

++++++ Summary as used in .EXE DataBase ++++++

Note: Only upper case (capitals) used to facilitate programming
Note: SGx numbers appear to end at 61
Format: [DATA] SGx# or 99etc , Constit Name , etc

USED CONSTITUENTS: Total 37
DATA 1,O1,1,MK1,2,K1,2,MO1,2,SP1,3,M2,3,KO2,4,S2,4,KP2,5,M4,6,MS4,7,MF
DATA 8,N2,8,KQ2,9,SA,10,MM,11,SIG1,11,SIGMA1,12,Q1,12,NK1,13,J1,13,MQ1
DATA 14,OO1,15,MU2,16,L2,16,2MN2,16,L2 2MN2,17,NO1,17,M1,18,MN4,20,SSA
DATA 21,2Q1,22,RHO1
DATA 24,PI1,25,P1,25,SK1,26,S1,31,2N2,31,2MS2,32,NU2,34,T2,36,K2,37,M3
DATA 46,MO3,46,2MK3,47,SO3,48,MK3,51,MK4,55,M6,57,2MS6,58,2MK6,61,MSF

UNUSED CONSTITUENTS: Total 108 4/12/95
99 or SGx used for known unused constits
98 used for total unknowns UK
97 used for total unknowns France

DATA 19,M8,23,CHI1,27,PSI1,28,PHI1,29,THE1,30,OQ1,33,LDA2,35,R2,38,TAU1
DATA 39,SO1,40,EPS2,41,OP2,42,MKS2,43,MSN2,44,ETA2,45,2SM2,49,SK3,50,SN4
DATA 52,S4,53,SK4,54,2MN6,56,MSN6,59,2SM6,60,MSK6
end SGx numbers Total = 24 unused + 37 used = 61

DATA 98,MNUM,99,MSM
Tot Long = 2

DATA 99,ALP1,99,UPS1,99,KQ1,99,BET1,99,MS1,99,MP1,98,TH1,97,KHI1,99,THETA1
Tot Diur = 9

DATA 99,OQ2,99,2MK2,99,SNK2,98,NA2,98,MA2,98,MB2,99,LAM2,99,2SK2,98,MSNU2
DATA 99,KJ2,99,SKM2,99,2MN2S2,99,2NS2,99,3M2S2,99,MNS2,99,MNUS2,99,M(SK)2
DATA 99,M(KS)2,99,LAMBDA2,99,NKM2
Tot Semis = 20

DATA 99,MQ3,98,2MS3,98,2MP3,99,MS3,98,2MQ3,99,SP3,99,S3,98,MP3,99,K3
Tot Thirds = 9

DATA 99,3MK4,99,3MS4,98,MSNK4,98,MNU4,99,2MSK4,98,MA4,99,2MKS4,99,3MN4
DATA 98,M2SK4,98,MR4,99,2MSN4,98,MT4,99,2NMS4,99,2MMUS4,99,2MNS4,99,2MNUS4
DATA 99,N4,97,3MN4ML4,97,NK4,99,2SMK4,99,2SNM4,99,2MKN4
Tot Quarters = 22

DATA 99,S6,99,4MK6,98,M2N6,99,4MS6,98,2MSNK6,98,2MNU6,98,3MSK6,98,MA6
DATA 98,3MKS6,98,4MN6,99,MNK6,98,2(MS)K6,98,2MT6,99,3MNS6,99,3MNUS6,99,2NM6
DATA 99,2SN6,99,3MSN6,99,3MKN6,99,3MNK6,97,4MN.2ML6,97,4MN.2ML
Tot Sixths = 22

+++++++++++++++++++++++
NOTE 1 Sa :
SH does not include -p'; AD (in Admiralty Manual) does not include -p' in Tables 7.1 or 15.1, but manages to reintroduce it in 7.17. MF appears to include it. Research problem. 20/12: include it.
NOTE 2 M3 :
Phase correction -180° or negative amplutude? Also AD implies only a quarter phase shift by addition of M1 and M2. 20/12: -180° in its own right. If interaction M1/M2, then should be found seperately.
NOTE 3
In general Cosine x° = Cosine x+360° = Cosine x-360° (same holds for Sines). Therefore, is a phase shift of -90° equal to one of +270° ? Also is a phase shift of +/- 180° the equivalent of Amplitude * -1 ? I have a sneaking feeling at the back of my mind that phase shifts of 360° should not be admissable, but cannot prove it mathematically. 20/12: confirmed no diff.
NOTE 4
The only Latitude dependant satellite with coeff > 0.1 is third for NO1. Ask MF for details. 20/12: MF might be able to supply formulae.
NOTE 5
Astro: Cartwright appears to be in error with h (Lon Sun). Example: Table 3, p56, Geophys J R astr Soc 1971 gives Phi r line 3, col 2 as 60.11923°. Should be 61.38645° ???
NOTE 6
Stands: Is there a "standard definition" either in terms of time or height? Between well defined highs and lows, frequently diurnals, and as derivative (slope) approaches 0, with or without change of sign, should time be the only criterion?

[Top of this page]  [Tidal Prediction Software]  [Stormy Weather Homepage]

Revision Date 19 July 1997 E-mail to     E-Mail us Copyright © 1997