MONTHLY NOTICES

OF THE

ROYAL ASTRONOMICAL SOCIETY

GEOPHYSICAL SUPPLEMENT
Vol. 6

No. 7

1953 June

T H E HARMONIC ANALYSIS OF THE E A R T H S MAGNETIC FIELD, FOR EPOCH 1942

Downloaded from http://gsmnras.oxfordjournals.org/ by guest on October 7, 2013

S i r Harold Spencer Jones, F.R.S. and P. J. Melotte

(Received 1952 May I)

Summary
The results are given of the harmonic analysis of the Admiralty magnetic charts of declination, horizontal intensity and inclination for the epoch ' 1942.5. Within the limits of observational error, the Earth's magnetic field appears to be entirely of internal origin. There is no evidence of a dipole field of external origin greater than 0-1 per cent of the field of internal origin. The intensity of the dipole field is at present decreasing at a rate of about 5 per cent per century. The geomagnetic poles have a westerly drift at a rate of 4"'s per century; the north magnetic dip pole is moving in a direction a little to the west of north, but the south magnetic dip pole appears to be practically stationary. In consequence of the dearth of magnetic data over the oceans since 1929, magnetic charts are becoming less accurate and there is a great need for airborne magnetic surveys of ocean areas.

I. In a paper entitled "The Earth's Magnetic Potential", by Dyson and Furner", a harmonic analysis of the Earth's magnetic field was given, based on data from the Admiralty Magnetic Charts for 1922. These charts, for magnetic declination, horizontal intensity and inclination, were compiled at the Royal Observatory. 2. Declination charts have been prepared at five-year intervals, while charts for other components of the magnetic field have been prepared at less frequent intervals. Charts for horizontal intensity and inclination, in addition to declination, were last prepared for the epoch 1942.5. In future, in accordance with a recommendation of the Association of Terrestrial Magnetism of the International Union for Geodesy and Geophysics, charts for declination will be produced at five-year intervals, 1955,'60, '65, etc., and charts for horizontal intensity, vertical intensity, total intensity and inclination at ten-year intervals, 1955, '65, '75, etc. Preparations for the compilation at the Observatory of the complete series of charts for the epoch 1955.0 have accordingly been commenced.

* F. Dyson and H. Furner, M.N., Geophys. Suppl., I,76, 1923.

G 30

410

H . Spencer Jones and P. J . Melotte, The harmonic

3. Since the unfortunate loss of the non-magnetic ship, the Carnegie, in 1929, practically no magnetic data for the sea areas have been obtained. Both the values of the elements of the field at epoch and their secular change have consequently become increasingly uncertain over the sea areas with lapse of time. It therefore seemed desirable to make a harmonic analysis of the Earth's magnetic field, as depicted by the Admiralty Magnetic Chartsfor 1942,in the expectationthat it would provide some indication of the regions in which the charted data were seriously in error and that comparisonbetween the analysis for the epochs 1922and 1942would provide information about the secular change of the field between those epochs in much greater detail than could be obtained from the scanty observations alone. The analysis was completed in 1945,but pressure of work has hitherto prevented .the results from being written up. 4 . Values of the declination, horizontal intensity and inclination were read off from the charts at points of a grid spaced at 10"intervals in longitude and latitude between 80" N. and 80" S. latitudes. From the values so tabulated, the northerly, easterly and vertical components of the field were calculated ; the values of these components are tabulated in Tables 1-111. If the magnetic potential is assumed to arise solely from forces situated inside the Earth, it can be expressed in the usual form :

V =a

n-0 m-0

e)n+l

{Hnm(X)(gnm cos m+ + hnmsin 4)),

where a denotes the radius of the Earth, r the distance from the centre, andgnm and hnmare numerical coefficients, (n - m)(n- m - I) Hnm(h)= cosmh {p"-- z(zn- I) Pn-m-2

4 is the longitude, A the latitude, and p stands for sink The nomenclature due to
C. F. Gauss and used by J. C. Adams (Collected Papers, Vol. 11, "The Theory of Terrestrial Magnetism") is here followed. The symbols g and h are known as the Gauss coefficients. From this expression, if X,Y, 2 denote the northerly, easterly and vertical (downwards) components, it follows that at the surface (where r = a ) : dHnm I aV X==C r ah dh {g," cos m+ + hnm sin m+},

y=--

rcosr\

av = XmHnmsec h { -gnm sin m+ +hnmcos M),

a+

Z = - av =Z(n + r)Hnm{gnm cos m+ + hnmsin 4). ar

The first six harmonics were included in the solution. 5. Though the functions Hmm have been extensively used in harmonic analysis by Laplace, Gauss, Adams and others, Adolf Schmidt introduced new orthogonal functions, Prim (0), defined by when m=o, and

TABLE I

X,Northerly Force. Unit 0.001gauss

4 0 '
30
f28I
321
O0

+80" +
zoo

7 0 '
6 0 '
zoo 10 f3Io $313

60"

50'

loo -k 225

oo E.
3ZI

+70
331 23 1 238

fro6

+ I88 +236

+272
244
253
I21

+I58
II2
102 I15 105

10

71 243 249 253

255

20

72
303

30 40

7 1

II 0 II 2 I10

60
322

50

105 I00

I 9 3 I97 198 I99 290 295 299

85

7 0 ' - 80' +167 +162 I45 155 152 I39 I46 I 4 5 I33 133 I37 I28 126 123
I11

91

82

200 201 201

70 80 90
288 352 351

67 62 54 4 5 36 29 263 273 289 306 340 257 262 267 275 283 307 314 324 336 346
52

202

75 72
291 289
285 325
311

2 0 4

+
22
20

96 80 56 26 83 70 60 53 60 34 26

27

7 1

207 214

- 26
16 16 57 61 74
85
I01
10

100 II 0 I20

220

130

26 28 33 282 276 268

351 3 . 5 0
258
280

140
300
321

37

73 76 81 94 349 342 33 1 317 506 293 357 360 359 356 353 290
313
285

225

334 342 350 358 370 381 392 394 389 374 357 340

150

I06

229 23 1

283 291 299 308 317 333 35 1 365 377 385 387 386 380 371 361
35 27 17
19

I60

I I4

232

33 I 337 345 353 364 381 393 4 400 400 392 376 361 345 332

170 E .
346 340 338
332

II 8

227 249 240 234

228
311 310 312

47 64 81 88 91 91

I80

220

88

170 W.
327 324
323
323
321

39 41 41 38 33 346 341 334 328

1 ' 5 107

212

I60

28

201

2
78 77 33
1.51

150
225 222

21

114 I 26 I34 142

.!!

140 314 318

320
321

+ 9 217
208
320

-10

+
161 I73 186 I97
213

12 15

I3

94 79 63 46 29
319
313

54
82
Y

130 I20 II 0 I00

I2

191 I81 I 68 I57 I 4 3

3 18 315

90
222

I1 I0 a1

5 9

131

24 37 40 70 98 123 148
173

23

I2

270 263 26I 262 266 269 269 265 259 247 234 308
300

I94
209

80 70
304 294
28 I

w
218
252

60 216

- I3

+
+

I2

r 20 116 II 9 124 272 270 274

284 288 293 299 303 305 303 296 284 268 257 290
280

217
222

50

29

132

I8 35 48 64

I99 185 176 1 7 1 1 7 1 I77 I88

226 237 240 238 226

210

221

109 129 I45 160 I74 I81 I 8 1

213

40 30

4 1 51

77

319 317 314 309 305 299 290 279 268 254 242 272 264 26I 23 1 +264 +226
,

180

20

58

90

I0

w.

f63

.. I 0 0

146 158 171 I81

191
,
,

201

177

226 254 202 24I 264 281 256 217 285 . 228 +269 f298 $300

+ '53

+I72

+I83

+I69

e
n

H . Spencer Jones and P. J . Melotte, The harmonic

TABLE I11
2,Vertical Force. Umt 0.001gauss
40"
0"

50"
10"

50"

0"

E.
-103

+
o
0

+ 10" + 18
- 300

10

518

20

519

129 I34
131

30

522

423 425 430

40
112
310

530

445

50

464 483
325 30 32

30 262 258 2 6I 27I 284 296

+ 17 5
126 98 107
122 132 132 121

501

60 70 80 90 340 348 349 340

325

5I 4 530

2 9 40 40 34

100

544

II 0 I20

130

140
252

150

548 547 539 51 0 475 306 2 8I 260 249 I09 98 87 7 3 59 39

20

I60

442

362 364 3 7 1 385 403 412 429 447 460 469 473 457 429 401 366 342
330 331

39 43 47 48 49 57 7 1 90

170 E. 180 170 W. I60

105

+
2

261 273 292 1 1 9

130

I4
20

150

313

140
150 520

I39 1 6 9 186 209

230 255
131

% Y
Y

130

I20 II 0 I00

339 359 390 423 456 490 2 9 37 46 60 8 1 107 149

153

90 80 70 60

418 412 417 439 463 494 53 1 557 584 601 602 586 566 266 265 249
201

id

R
I39
I00

50

544 508

40
335
302

3 4 1 370 393 419 453 483 481 472 445 407 3 7 1 I47
95 50

30

+ 50
-

10

20 I0

w.
276

540 549 558 564 569 573 577 578 578 577 573 569 565 560 565 572 579 586 592 597 598 595 591 585 578 566 553 539 53 1 525 519 556 562 549 530 504 467 435 406 386 3 7 1 60 -139

60" f 470 467 499 469 502 477 505 486 511 496 520 511 542 529 555 547 566 559 576 569 586 577 595 580 603 602 578 5 7 1 593 563 584 547 574 526 563 506 552 495 551 497 558 521 572 552 586 596 581 592 605 605 615 620 615 610 610 598 598 578 584 564 569 541 552 534 522 499 519 488 51 0 503 $80

484 462 441 429

2 0 ' 30" 40" -219 -238 -260 286 27I 249 269 287 301 2 8I 294 3 I3 282 300 328 282 313 351 328 390 289 348 425 303 318 367 441 395 449 336 429 470 369 465 502 385 524 491 387 384 495 542 370 491 560 563 351 484 559 330 468 536 310 442 292 419 501 386 264 480 243 362 459 226 338 428 209 316 399 293 190 377 268 354 172 240 152 32 6 210 126 294 178 257 96 142 214 64 184 115 46 104 I64 38 I06 I59 42 62 I 67 II 9 183 I44 104 208 I 7 3 I 4 3 180 204 236

60" 70" - 80" ,428 -513 -354 315 441 519 372 330 392 457 526 418 473 3 4 5 535 3 7 1 489 543 4 39 409 505 455 552 441 472 522 562 459 536 491 572 474 548 512 583 482 558 592 525 480 568 605 524 621 475 577 525 479 589 639 527 610 506 535 650 660 647 542 555 600 690 566 665 668 685 578 647 682 662 666 579 675 574 657 665 660 553 650 658 618 650 650 526 498 637 645 564 628 587 474 525 607 455 507 555 492 538 586 436 410 474 521 568 377 446 505 554 338 543 421 488 293 384 467 534 266 448 353 525 237 431 326 518 418 512 227 313 228 310 410 509 236 312 408 507 323 256 507 4 11 276 338 417 509

414

H . Spencer Jones and P . J. Melotte, The harmonic

where P , denotes the associated Legendre function definedby

= cos e, where 6 being the co-latitude, i.e. (n/z-A). It is readily shown that P , is related to Hnmby the relation

(zn- I ) !! Hnm. (n-m)! Hnmis identical with the function which Schmidt denotes by Papm. It follows that

(n - m) ! 112 (zn- I ) !! p n , ,n = { z m } (n-m)! =5 p n . m (SaY). The values of 4 = Pnrn/Pnnm for n = I to 6, m = o to 6 have been tabulated by Schmidt. * The use of the functions Pnmin geophysical investigations in preference to the functions Hnm has been recommended by S. Chapman. Their properties are described in Geomagnetism, Vol. 11, Chap. XVII, by Chapman and Bartels. Schmidt (Zoc. cit.) has tabulated the values of the function Pnm(cos 0 ) and the associated functions Xnm = dPnm(COS e)/ndo; Ynm = mPnm(cos e)/nsin 0

for values of m and n up to 6 and for values of 0 (the co-latitude) at intervals of 5" from 0' to g o ' . His tables can therefore be used to obtain the values of dHnm/dA, mHnmsec A and (n + I) Hnmin the expressions for X,Y, 2 above. 6. Vestinet carried out a harmonic analysis of the U.S. magnetic charts for epoch 1945. He expresses the magnetic potential from internal forces in the form :

V =aCX

( ) :

*+I1

- Pnm(Anm cos m+ + Bnm sin m+) n

and, at the surface, X = CXnm{Anm cos m+ + Bnm sin 4}, Y =X Ynm{ - Anm sin 4 + Bnm cos m+},

+Bnm sin m+), n where Xnm, Ynm, Pnmare as tabulated by Schmidt. In the Gaussian notation
Z =C -Pnn& {Anm cos m+
n+I

Xn"

- n dHnm/dA ;

'

Y n m= n (mH,m sec A) ;

Prim =(H,".

* A . Schmidt, Tafeln der Normierten Kugelfunctionen, Gotha, 1935, p. 20. t E. H. Vestine and others, The Geomagnetic Field, Its Description and Analysis, Camegie Institution, Washington. Publication 580, 1947.

415 The values of the Gaussian coefficients gnm, hnmare therefore related to the coefficients A,,", Bnm used by Vestine by the formulae

analysis o f the Earth's rnagneticJield, for epoch 1942

These formulae can be used for comparing the results of the analysis by Vestine of the American charts for epoch 1945 with those of the present analysis of the British charts for epoch 1942. It should be noted that Vestine, in his Table IX, tabulates the first eight Gaussian coefficients of the Earth's magnetic potential from a series of analyses going back to the earliest, by Gauss, 1835. These coefficients are gnm,hnmup to the value of 2 for m and n. The tabulated values are not, however, those of the coefficients as defined by Gauss, but those of the coefficients Anm,Bnm. It is advisable that a distinction should be made in the nomenclature for, and the designation of, the coefficients according to whether the expressions used by Gauss or by Schmidt are adopted. 7. The values of X,Y, 2, given in Tables I, I1 and 111, were analysed to give the g and h coefficients, the different latitude zones being combined together with the relative weights used by Dyson and Furner, viz. :Lat.
oo
&1O0

Wt.
I0

I0

Lat. f30" f40"

Wt. 8

Lat.
&60

Wt. 3
2 I

+zoo

f50"

7 5

f70" *soo

These preliminary solutions having been made, the weighted residuals for each latitude zone for each of the solutions were tabulated, and the mean residuals obtained for X,Y, 2 separately. The mean residuals corresponding to unit weight for each latitude zone and for each component were then obtained. For X and Y the zonal values ran smoothly, but in Zthe scatter was large, particularly in southern latitudes. For each component the mean residual was plotted against latitude and a smooth curve was drawn through the plotted points. The values read off from this curve were adopted as the mean residual for each zone of latitude and for each component. From these values, relative weights were obtained, which represent more accurately the uncertainties in the chart values than the provisional arbitrary weights originally used. The relative weights so derived, which were used in new solutions to obtain the definitive values of theg and h coefficients,are given in Table IV. 8. It will be seen that in general the weight for each component for a southern latitude belt is appreciably less than the weight for the corresponding northern latitude belt. This is because the land areas are more extensive in the northern hemisphere than in the southern ; over many of the land areas the magnetic data are reasonably well known from land magnetic surveys, whereas there have been very few magnetic observations in the sea areas since 1929. An additional reason is that there are many more permanent magnetic observatories and secular change repeat stations in the northern hemisphere than in the southern, in consequence of which the secular change data are more reliable for the northern hemisphere than for the southern. The weights in Table N have been formed on a uniform basis for each component. It will be noted that the weights inXand Yare of the same order, the

416

H. Spencer Jones and P.J. Melotte, The harmotzic

weights in Y being mostly greater than those in X in northern latitudes, but somewhat smaller in southern latitudes. The X and Y values are formed from the observed values of declination and of horizontal intensity ; for most of the region of the Earth's surface between 60" N. and 60" S. latitude, the declination does not exceed 30". A given error in H will therefore produce a larger error in X than in Y, while a given error in D will produce a larger error in Y than in X. I n the southern Indian Ocean the declination is known to be very uncertain, as in this region the secular change in declination is large, while the rate of change of declination with latitude is also large.
TABLE IV
Relative Weights for each Latitude Belt and for each Component ( X , Y , Z) Lat. +80 70 60 50 40
30

X
Ia 6 1
2 -63 4'35 . 6.95 8.62

Y
2a 0 6

3 '05 5 '00 7-58 I I '40

15'2

2 0.35 0.46 0-63 0.87 I '25

1.51
1 '54

10'2

20

II' 9

19-2
20.8

I0 0 I0 20

13.2
I3 ' 9

19.2
15'2

1-39 I '24
1-01

13.5
11'1

30 4 0

-80

50 60 70

8'20 6.17 3' 7 9 2-28 0' 7 4 0.36

10.9 7'25 5 '00 3 '47 1' 7 9 o .80

0-42

0.76 0.53 0'35 0.18

0.15

0.13 0.13

The weights of the vertical intensity are relatively low. This component is formed from the charted values of horizontal intensity and inclination. It seems probable that there are appreciable errors in the inclination: over much of the surface of the globe, moreover, the inclination exceeds 45",so that the errors in inclination are much enhanced in the vertical intensity. 9. If a portion of the Earth's magnetic field is of external origin, the magnetic potential will be expressible in the form

+ uX (i)" [H,"

(g-nmcosm$

+ hWnm sin m$)].

The analysis, in the form explained above, by which the coefficients of the terms depending on the longitude are derived will determine for X,Y g,m+g-,m, n g,"for z

---g-,m.

The coefficients determined from the separate analyses of the northerly and easterly components of the magnetic field should therefore agree, apart from the effect of accidental errors. The values of the coefficients obtained from the

analysis of the Earths magneticfield, for epoch 1942

417

analysis of these two components, with the weights first assumed, were in fact in close agreement. A new solution was accordingly made by combining the data for these two components and using the definitive weights given in Table IV. The derived values of the coefficients are given in the second column of Table V. The vertical component was analysed separately, using the definitive weights given in Table IV, and the derived values of the coefficients are given in the third column of Table V. 10. If there is an external component of the total field the difference Y)and 2 is given by between the values of g,, derived separately from (X+ ( z n + ~ ) g - , m / ( n + ~ ) In . comparing the values given in columns z and 3 of Table V it should be noted that the values in column z have a weight that is seven or eight times the weight of the values in column 3, the combined weight being given in column 5 , and that the probable error corresponding to unit weight is about +o.0035. From an examination of columns z and 3, in conjunction with the relative weights and probable errors, no evidence of the existence of an external component of the field is found. I n relation to the probable errors of the various quantities, there are no significant differences between the corresponding values in columns z and 3 . The main portion of the field is represented by the dipole field of a uniformly magnetized sphere (Section 16). The intensity of this field, Ho, the co-latitude, do, and the longitude, rJo, of the northern pole of this field, as derived separately from the north and east and from the vertical components of the field, are as follows :From X and Y From 2

0.309 6 + o*ooo4 0.310 4 + 0-0010

H O

60
IIO-I

40

IIo.O

- 68O.4 - 69O.4

It can be concluded from these results that there is no evidence of a dipole field of external origin which exceeds one-tenth of one per cent of the field of internal origin. It has therefore been assumed that there is no portion of the Earths magnetic field that is of external origin. The equations for the (X+ Y) and the 2 components were therefore combined together with appropriate weights and solved to give the values listed in column 4 of Table V ,the relative weights being given in column 5. 11. It is of some interest to compare the results of the present analysis with the results obtained by Vestine. The charts from which the data for the two analyses were read off differ in epoch by only three years. Though the British and American charts were prepared quite independently, they are based essentially on the same data. But over many areas recent observations are completely lacking and both series of charts consequently have been based in some areas on extrapolation, involving the assessment of the secular change of the Earths magnetic field and on the rate of change of this secular change. The comparison between the Gaussian coefficients and the first four harmonics from (X+ Y), and from 2, as derived by Vestine and in the present investigation, is given in Table VI. The agreement between the coefficients derived from the north and east components is satisfactory. The discordances between those derived from the vertical component are appreciably greater, which is to be expected from the

H . Spencer Jones and P. J. Melotte, The harmonic

Tmm V
Values of Coefickts
From Xand Y f.303 8 +a017 8 - -025 5 - '039 7 '030 3 From From From

r, Yand
(spherica Earth) 9
I

Relative weights 97 59 28
12

r, Yand;
spheroidd Earth) +-3022 +.or76 - '0292 - -0385 +a0182 - -0142 +a0218 - -0507

'3039 + '304 7 + +a0176

+'015 a 0 2 4

- '039 7
+.OZI 9 '020 8 - -047 6 +'044 3 - -032 9 - '042 6 --011 7
--021

+'0293
-a0211

- *0255 - *0398
+.0218 - '0509

5
2

--021 0
+.OZI

g
2

+ 4 - -040 5
'052

--051

+'0515

- '030 2

- '007 5

- '0397 - '0329 - '0073

+.0190 - '0139 '0057 - *0026 --or35 - .0236 --0238 --or30 - *0007

I93 94 38 I5 6
2

- -0384

+-0508

-'0342
- '0054
--0553 + '0259 +'OIg4

- '055 4 --055 4 - '0555 + -026 6 +'022 4 + .0260

+ '019 4 - .or4
I I

+.or6

--010

2 I
0

I93 94 38
15

- '002 3 -*013 3 -so23 9 - -023 6 - -013 5

--oar
0

+a006

--001 2

-*003
--015

+ '0053 --001g
- -0235 - -0234
--or35

-'Or43

- a003 4 - *0189 - .007 5

+a008 5 - '002 4

--004

9 --or9 4 - -027 9 -Om7 3 5 - -006 8

532
115

32
I0

+.-

-moo 4
-*OOI

+ .007 o
- -036 5
+-5
+.OIZ
--001

- -0044 - '0033

+ -0076
--0018

3 532 1 I5 32
I0

--or23
'0002

+ -0076
- -0206
--m17

- '0044 - '0033

+.om 7
+ ~ O 6 I
+'OOO

+ -008 9 + '002 3
7

- .006 6
3 9 7
+.or1 I
+*005 0

- .0204 - '0074

+ so087 f '003
+.OZIO

3 983 I39 31 8 983 139 31 8 1523 168 30 1523 168 30 2189 196 2189 196 2835 2835

+ ,0087 + *ozog '0033 +

- '0001

- '0074

'000 0

- 'OOOI
+*001g '0009 +a0018 --0018 -0034 '0017

+moo 8

+ +

'000 7 '010 0

+*0019 'ooog +'0017

- '002 0

+ -003 3
+a02 I

+a000

+.OOI 0

+ *0034 +moo 8
+*om 5
4 --'000 5 '000 5 +-OOO 6
--OOO

+.003 9 -moo3 8 +'000 9 +.oo1 9 +.003 I

+-OOO

+ + + +'0032
--om4

--0018

+ '0033 +.0018
+.OoIO

'0010

'000 0

+ *-9 + -0005 - *om4

+. o o q + a006 +'OOO2

+'0032 + .ooog + -0005 - '0004

- '0004

--'ooo4 - '000 6 '000 3

+ .om5

+.000 2

+.000 2

+ -0006 + '0002

419 relatively low weight of 2 from the British charts, as shown by Table IV. The agreement between the coefficients derived from the elements ( X + Y ) and 2 is better from the American charts than from the British ; the American charts of vertical intensity are the better. TABLE VI
Gaussian coefickts derived by Vestine and by Jones and Melotte

analysis of the Earth's magnetic jield, for epoch 1942

x+ Y
Vestine
'305 7 +*019 o - -028 7 --wo 3
+.02I
0

z
Difference Vestine Jones and Melotte Difference

Jones and Melotte

+*303 8 +.017 8 --025 5 - '039 7
f.021

+ '304 7 + '305 7 + + +.or5 9

+'Or7 9

'0010
20
22

- so26 7 - '043 7
--0g10

--ow5

- '039 7

40

--051 2

+'051 9

-*w3 --058

+ 4 - -0405
'052

--051 2

+ '022 7 + -020 8 + - *0476 +-053 8 +w4 3 + - -032 9 --'041 5

19 34 95 86
25

'055 4 + -026 6 + -028 7 I

+.or5 9
--007 8
--014 I - .023 4
--022

+*OI9 4 --014 I

--057 9 --'055 4 + 4 + + *0297 +-016 + 3 --.006 4

'022

'014

2 --010 I

- .or3 3 - a023 9
--oz3 6
2

--014 4 - '023 3 -a023 2

-0004 I --004 0 +*or1 4

- '015 9

- -019 4 - a027 9
-*006 8 --'ooo 4 --OOI 6 -so06 6 3
-a01

--w3 5 +.OIO 8

--* 6 -a04 - .003 4

+a008 5

+ + +

73 I9 37
15 39 47

+
+ + -

+ -

27 36 130

- .006 9
f *007 9
4 +-0017
--OOO

- -007 5
+*008 9
'000 0

--.007 3 + -008 5
'OOO 0

7 32
I7 28 30
0

7
0

f-0016
--'002 0

f.002

+.005
+.OOO +.OOO

- '002 2
+*OOO

--'002 9

+.oo1

+.OOo 9

Vestine has not given the weights of his coefficients; he used the weights that were adopted by Dyson and Furner for combining the data for different latitudes, which give the southern latitudes too great a weight relative to the northern. On the basis of the assumption, which is not strictly correct, that the weights of each coefficient deduced from the X and Y charts are equal in the two investigations, it was found, from a comparison of all the coefficients for the first six harmonics, that the probable error corresponding to unit weight was & 0.0054. The value obtained in the present investigation was k 0.0035. The concordance of these two values is sufficiently close, in view of the fact

420

H . Spencer Jones and P. J. Melotte, The harmonic

that the basic assumption about the weights is not strictly accurate, to suggest that the weights of the two determinations are not greatly different and that the discordances between the individual coefficients are mainly accidental. 12. With the adopted values of the coefficients, given in column 4 of Table V, the values of X,Y, 2 at each point of the 10" network were computed and compared with the values derived from the charts. The differences between the chart data and the computed values are given for the three elements respectively in Tables VII, VIII and IX. The quantities in these tables necessarily run on the whole smoothly, because the chart data are smoothed to eliminate as far as possible the effects of local magnetic anomalies, while the computed values necessarily run smoothly. Comparison between these tables and the corresponding tables in Dyson and Furner's paper (loc. cit. pp. 85-87) is of interest. In Tables VII and VIII there are several regions in which the residuals are systematic and of appreciable magnitude. The residuals in these tables are larger on the average than those based on the 1922 charts, an indication that the 1942charts are of lower accuracy than the 1922 charts, as would be expected from the paucity of ocean data since 1929. There is, moreover, little correlation between the corresponding residuals from the 1922and 1942 charts ; if the charts were of high accuracy and the same regions stood out in the same direction at different epochs, it would indicate that the general magnetic field was not represented with adequate accuracy by the six harmonics. The comparison between the 1922 and I942 residuals suggests that there must be appreciable errors in the secular change field that was adopted for bringing forward observations to the epoch 1942.5 of the later charts, while the large systematic runs of residuals suggest that there are areas where the charts are appreciably in error. Comparison between the tables clearly shows that the vertical force data are of much lower reliability than those for the north and east components. The largest residuals are found in regions where the errors in horizontal intensity and in inclination are additive and where, in addition, the tangent of the dip has a large value. In middle latitudes (between 50"N. and 50" S.) the residuals tend to be larger than those obtained from the 1922 charts, but in the polar caps the very large residuals obtained from the 1922 charts have been reduced. 13. The charts on which this analysis was based were charts of declination, horizontal intensity and inclination. From the values of X,Y and 2 computed from the coefficients listed in Table V,the values of these three components were derived. Tables X, XI, XI1 give the differences between the chart values and the computed values for the three elements declination, horizontal intensity and inclination. Table X is of special interest because of the use of the charts of declination for navigational purposes. Too much regard need not be paid to the large residuals in high north and south latitudes, because of the proximity to the magnetic and geographical poles, which are singular points in the system of isogonals. In latitudes 50" S. and 60" S. there are some considerable residuals in the southern Indian Ocean, which suggest that in places in this area the charts m,ay be at least 5" in error. The southern Indian Ocean has been recognized for many years as a difficult region, for not only are the isogonals more closely spaced there than elsewhere, but in addition it is a region where the secular change of declination is both large and ill-determined. It is clear from Table X that the charts in declination are not of the accuracy that is needed for certain modern developments in navigational aids.

TABLE VII
40"
30"
0 0"

X , Northerly Force. Chart minus Computed. Unit 0.001 gauss

+80"
- 10
60'
7ob -80b 40' +4
soo - 5
0 -2

60"
30
+ 5
0

50"
+I

+ 1o0
-3
-1

2oa

oo E.
+2

-1

20 +3

-2 +2
-1

+2

+7 +3
+ 3
+ 2
- 2

-6 -3
0
0 0
0

I0 20

- 6 - 8 - 8
+I 0 -2

+3
+3
+I
+I

+2 +2

30
-2

- 7
-2 +2
-1 -1

-1

+2

-I 0 - 14 -I 3
+21

-2
+2

+2 -2

+3 +4 +3 - 3
- 2 - 2

40 50

-1

-4
+I

- I 9 -25 - 15 -22 -25 - I 3 -20 -23 - 6 -20 -21 - 2 -20 - I9 + 6 - I9 -15 + I5

60
+3
+2
-1

+I

-2

+I $2

-6 -3 -4 -4 -4 -5 -3

-8
-2

+5
0

- 8 - 5 - 3

$2

+4 $4 +4
+5
0
-1 0

+27
-21
-11

+4
0

7 6 6 8 9
+2
-1

70 80 90 $3
+5
+I

-I 0
-7

-8 -7 +9 +7 -4 -6 -3
+ I + 2

70" 6 6 5 5 5 4 5 5 4 - 9 -I 0 - I5

$2

t5
+4

-18 -15 - 18 I5

+ I
-1

+3
$2

-5
-3

+ 5

+26 + 18 I2

100 II0

- 8 - 7

--I

+4 +3 +3 -3 -3 -3

+ 8
+5
10

-I 2 -I 0

I20

- 5
+2
+2 +I
-1

+ I - 1
0

-1

+6 +6 +4 +3 +3
$2 0 -2
+I +I +I +I +2
0

-1

-3 -3
$2

+
+ + + + $
+ 2

- 6
- 2

+ 9
+I1

130
-2
0 -2 -1
0

140
+2 +2

-2
-2
-1
--I

- 2

- 5

-3 +4
+5
$2
-2

$3 $4 +3 +3 +3 +3 +5 +5 +3
0 0

- 2
+2
+2

-2

- 3
+ I

12 + I

150

- 2

-2

-2 -2

-3 -3
0 0

- 2

5 8 7 8 8 6 6 3 4
+6 +5 +7 +7 +5 +5 +3

+ + 3 +
+2
+2

21

- 1

7 6 5 5 4

+ 5

+3
+2
+2
-1

-3 -4 -5
-2
-1

I 60 170 E. I 80 170 W.

-4 -5 -6 -8 $4 $3 +3
-2 -2

+ 2 + 2
0

+ 5

+ 5
+ 2

160
+I
-1

- 3 - 6 - 9
+2 -2
-1 0

-9

-4 -4 -3 -4 -3 +3
+I +I +2
-1 -1

-I 0
+2 +I +I
0

-2

+2

+ I

+3
+ 2 + 2 + 2

+ 4 - 3 - 9 - 19 -24 -28 -33 -39 -40 - 53 -26

- 2

150 140
+I +2

+6 $7 +7 $6 +4 +4 +3
+2
0 -2

-10

+ 4 16 -26 -34 -34 -46

+ 3 + 5

130
+2
+I

-7 -6 -6
+I
$2
-1

- 3
- 2
- 1

-23 - I7 -32 -22 -16

- 3
-11

-2

I20 II0 I00

+3 +3
+3
0

+I +2 0 -1

+3

+ 5

-I 0 - 12 - I7 - 16
+3
-2
-1

+I +I

+2

3 4 5 7 6 5 8 - I9 - I4 - 7 - 3 15 +3 +4 +5 $4 +4
+2

+2
+I
-1

-I 0
+2 +2 +I +I -2
-1 0
+I 0

+2
+I

- 5
0

-21

-I 0
+2 +I

+I -1

+ 9 + 6
$ 2 + 2

+I

+ + + + + -3 -4 -3 -4 -3
+I 0 0 -2 -2 0

3 4 4 4 3
+ 2

90 80 70 60
-3
-2 -2 +2
-1
-1

- 18
+2 0 -2

-11

-1

+3

+4 +6 +6 +4

- 2 - 2

4-3
-1

- 3
-2
+2

$2

+4 +6 +6 +6 +6 +4 +3 +3
+ 2

+ + + + + + +

5 5 4 6 7 6 5

- 2

50

-2

40
+I +I +I

+ 2 + 2

- 6 - 8 - 6
0 0

-3 -4 -3 +5 +6

-1

-1

+I

+ 2

9
+ 2

+4 -4
+5

+ 2 + I + I --2

- 5
-11

30

- 7

-2

-3

- 2

4 5 5 9

20

-2

-2

- 5 - 9 I4
-3 -6 -7

-5
+ 4 + 5 $4

-I 2
- 4

- I4 - 15

I0

w.

- 3 - 7

- 6 - 6

+2 +2 +I +I

-19

-20

- I7

R n

P
0.001 gauss

TABLE VIII Y , Easterly Force. Chart minus Computed. Unit

40'
-10'

N N

+80'

70"
30'
-4
-1 0

60"
-5
0 0
+I --I

soo
-2

zoo

+roo
oo

zoo

50'

70"
-80"
- 1

0"

E .
I

+ 6
-3
+I
- 2
-1 -1

40' - 3

60" - 3

o
+ 2
0

+ I I

-1 -1

-4 -3

- 4

+ 6
+ 9
+I2

I0 20
-2

+ + 5 I
I
-1 +I -1
-1

- 0

-1

-1 -2

30. -3 -3 -3

+
+ I
0

- 2 + I I

30
0
--I

- 2

-5
-3
- 9
-2
+2
0 +2
+

-4 -4 -4 -4

-6
- 6
- 2

- 2 0
-1

- 2

-4
-2 -1 -2

-2

-3 -3

-3 -3

- 3 - 5

+I3
+ Z

40 50

I
-11

3
0 0
--I

-8
-14
- 2
+ 2

-5

-3
-5
0
-11

-3 -5 -3

-21
-12
0

+I3
+ 4
$15

60

+ 4

- 2

-5

-5

-3
-2 -2

-6

-6
-2

+ 3
+ I

+I2

+
-5
-2
-1

2
0

-1

-4 -4

-6 -7 -6
-4 -3

-7
+ 4

+ 7
+IZ
2

70 80 90
-2

+ 8 5 + 3
+I

+3
+I

I
0

-4 -4 -3 -3

--I

-3 -4 -4
0
-1 +I 0
0

-4 -5 -3 -3
o

100

+ I
+2

+I

+ 4

+ 8

+I9
+ 9

+ 6 + +I4
$18
+I4

110

+I

+4

44
+5

+3 +3 + 3
+ 2
0 2

+ 7
+ 4
+12

I20

+3 +3 +6 +3
+I
-2
+2 $2

--I

+3
+I
0

+3
+I
+2

+3
+2

+3 +3
+I 0
+2

+ 8
+-I0

- 3

130
0

+I0
+2 -1-1 +Z +2

+2
+2

-5

140
0

.-

-4
+2 +2

+I -1 -2 -2

+4
$4 +4 +5
+S +3

+
2

I
0 0

+ 7 + 7
+ 2

- 2
0

150

I I

-2

+3

+6
+4 +4
+I

+z
+2

P
1

I60

+13 +I3 +I3

+7
+5
+4

+3

+I4

9
+3 +z

170 E.
+z +z +z
+2
+2
+2

+I1

+I

+s
+6 +4.
+4
0

I 80

+3 +3 t 4
+5
0 0

t 3 +6 +5

+3 +3 +6 +z
+2

+
I
0

- 7 - 4 -4 - 4 7

170 W. I 60

+9 +7
-t5
+4 $5 $4 $4
+I

+ +3 +3 + +
+z
+I
+I 0 +I +I +I 0
+2 +2 +2 +I
0

+3 +3

+2

+ + +3 +4 +4 +
0 -2
I I

+ +
-2

150

2 I + I 0

$2 0

I I 2 + 2

+4
+ I
- 1

+3
- 3
- 5

-2
+16 +I4
2

140

+I

+I.

+33 +36 +36 +38 +35 +37

Y
+IS

130
+z
+2
0

+3 +3
+2
+I

+I

+3 +4
0 -2
0

+4 +3 +3
-3
0
0
0

+7 +6 +3 +
-3
-1
0

I20 II0

+ 6 + 6 7 + 6 z
+3
+Z

+ + 4 +8
+I
0 0 0 $2

+4

+ + +
+I +2

+I1

+3
+I
0
-1

+z +4
+I +2
+I

+3 +3 +5 +5 +5 +4 +5 +3 +3 +3 +3 +3 -3
0
+I
-1

- 3 - 4 - 3
--I
+I

- 5
- 2

- 2 - 3
--I

+26 $15

5
$ 8
+ I

+I3 + 5
- 5

I00

+8 +8
+I

-I

+Z

90
0

+I2

5
-1
-1

-4

+ 3
0

+ z -3
-2 0 -2

- 3
- 2
+I +I -2

- 4

+2 +2
0 0

6
+I 0 0

-2

-4
-2

-10 -12

- 5
0

-5 -4

-4 -3
-2 -2
-2 -2

- 2

- 4

-16 -4
-3 -3
-1 -2
0

80 70 60
-3
-2
0

+13 + 5 4

-4 -4 -3
-2

-3 -3

0 I 2

+ I I - 2

- 3
- - I
- 1

- 4

-21

--I

+I

- 3
+3
-5 -5
+I +I
+Z

- 1

- 3 - 4

-21

50 40
-2

+ + 4 +7 + + 6
-4 -4
-3

+ 7

+I

-17

30

+ 9

-2

20

+9 +

+ 5 3

-2

+
-3

I0

w. +

-3 -3

-3

-8 -6 -4

-4 -5

-3

0 I 0

I 0

+ I + 2

- 9 - 4 - 3 - 3
- 1
- 2

Tmm I X
OWOI
20"

+80" +
- 3

oo E.

6 0 ' 70" + 3 + 9

40" soo 60" 70 + 4 -14 -16 - 6

-15

-80"

+
+ 5

I0

I0

+ 2

+ 6

+ +
-18
- 2 2

20

+ I

30
+ I

-18 - 18 - 18 - 18 -6 -9 -45
-15 -23

-23 -31

+ 7
+ 3

- 2

+ 4 + 4

40

- 15

- 5

+ I

gauss 30" I5 I 7 + 6 I0 + 6 + 4 0 + 4 + 3

50

-I 0
+ 8

-9 + 6
-35
-32
22
+21

-6
- I
8

f 2 9
+I30
+ I I ~
-10

-I 0

-I 2 - 18 -22 -24 -23 -20

-11

60 70

-7 -5

0 - 2

+ 4 +6 + 6
+ 8
3-2

80

- I3

34 - 5 0 53 - 46 5 7 - 37 3 7 - 24

+ 9

90

- 5 - 5

+
+40 +60
+58
+52
+21

+
22 I0
+ 2

+ + + +
+I00
3-11

f 14

100

-6

- 5 -7 -6

II 0
+I1
0

- 5
+ 3
10

- 15 -20 -20

- 14

2,Vertical Force. Chart minus Computed. Unit O0 - 10" zoo +loo 3 0 ' soo 40" 0 0 6 + 8 + 6 + 4 + 5 + 6 - 3 - 3 + 3 + 9 + 4 - 3 - 2 - 2 + 2 + 4 - 3 + 3 - 1 + I + 5 -4 + 3 - 2 + 6 + I - 3 -4 + 5 - a + 2 + I + I + z + 8 + 4 - 2 I2 + 2 -4 I 9 + 7 + 3 - 2 + 8 I2 28 - 3 - 5 + 7 - 12 - 5 +I1 + 9 + 3 I3 34 I2 1 6 + S 40 -13 I3 -4 0 I2 18 2 6 + 5 14 -7 + I + 6 1 0 + 14 19 3.21

+
+38 +16
-10
12

+ + + + +
+ 73 + 50 +III + 79
+lo8

I20

+ 8
+2I

+I1

+
+I4
- 8
+I1

+ + I3
-7 -4 -9

130
I2
+ I
+ 2

-4 + 8

+ 9
+ 2

+ I + I

+ 8

+ 8

+ 19
+ 3

140
+I1 I0

-6 -6 -7

150

+I1 10 I0

+
+ 6 + 3 + 7 + 9

20

27 26

+
+
+ 7
+ 3

+ +
+ 6 + 3
+ I + I

20

+ +42 +46 +58 +61 +45 +60 +33 + 3-42 + + 9 -5 + 3 -4

+ 3
+ I
- 2 0
+ I
0

+I7 +I3 +I 2 +3I +43 + 7 +49 +43 - 4 + I5 - 7 22

-7
+I5 3

I 60

+ 8

+33 +35

+ + +
- 3
+

10

+26
+ 3
0 0

0
0

+ 4

+I1

- 2

- 2

- 3

+ 8 + 8 + 5

+ 82 + 96 4 5 3 +53 + 30 + 3
+I0

-I 0 - I3 . - 16 170E I80 -20 170W. - 15

- 5 - 5

+ 4 -4 - 5

-3

- 5

I60

-8

- 2

-11

150

- 2

+ 6
+ I

-4
+ 5
+ I + I
+ 2 + 2

+ I 0

+ 3

+ I

+ I

+ Z

+ 9
+ 8 + 5

- 3 + 3

- I5 -13 -I 0
-19 -24 7 -28 +I7 - 9 27 30

+ I
52

- 8 - 5 - 5

+ 6 + 4
+ I

+ Z

+I

+ I

140 + 6 + 4
+ 8 + 8 + 5
0

+ 4 + 6 + 6
+ I
+ 2

+I 2
+ 4 + 4 + + + +
+ 3 + 3
0 0

+ 8 + 5

+ 2

130

+
+2I
10
22

10

+ 13
+ 7 + 9
+ I

+16
+I1

+ 9 + 7

+ 6 + 7 + 6 + 6
3-9
10

k
Y

+ 3

+ 4 + 7 + 4 + 6
I0 I2
1 . 5

+18 +17 +I4

+I1

20

I20 110 I00

+ 15 + I7

+I5 +ZI

+ 9 + 9

I2

7 7 4 3

+I2

11

90

12 I0

$ 9

+ + +37 + 9 + 48 + 5 I + + 37 3-48 +30 + 24 +42 +36

+IS
+I3
+I1
11

4 + + + 7 + 16 +23 + +9 + 14

+ 8 + 5 + 7
10

+ + + 4- 31 + 27 + 18 + + 6 + + +

+33

+37

+24

+35
20

\o

80

13 +I1

+ 7 + 6
+ 8
+I3

+I4 +
+ 9
+ 8
I2

+ + I3 +
+ 7
+ZI +2I

+13 + 12 +15 + 14
+
+ 4 14 18
+I1

+ +
I2 I5

+32 +29
+ 9 + 7

70 60

50
+I1

+16 +25 + + 19 + + 15 . + + I4 + 8 + +4 + + 5 + 8
I-10
I0

12 10

- 3

+I1 +I1

+I1

+ 5

+ 7
13 15

+ + 8

+ +
3-9 + +15 + I 1 + 8 +17 + I7 +I1 + 15 + I3 + I4 + + 17 +I9
13

+26
I0

+23

40

-11

+ 4

+14

+I1

+ + 16

+ + 16 +I4 +I 2 +
+ 7
+ 4

30

- 14
+ 6

+ I

+
+ + +I 2

10

13

20

+ 2

+ +I4 + 17 + I 1 + I9 + 2 I + I4 +I3 4-15

+I1

+ + + + +
+ 7

9 6 5 7 7 + 4
+ 8

+ + + -

I0

w.

- I5 - 18

+ 2

+ I3 + 14

+ 9 + 7

+r9

+2I

+5

4 9 7 +I2 4 + 9 1 + 4 I + I 5 - 5 7 -11

+I9 +I7 + 14

I2

R w

TABLE X

-fs
P

if
soo

60"
-32 31

+ +
0 ' 70" -8 - 27 I - 9 -16 -13 -27 -35 -38 -42 2 8
-55

+
- 35 - 35 - 31 - 15

+
-29
-10

-32 -31 -33 -35

+ I7 + 26 + 47 + 70
+ +

+ 97

% R
-

t ,
rn

+ + + + +

+5
+18
+21

-61 -58 -48 -42 -48 -26

+ +

+ +
+IS
+IO

3 x
2
R

+ +

R.

+ +

+ + +

+ +

+6 +7 +7

w. +

+4

+ 6

+ + +

+5 o
-15

24 +I17 49 + I 1 2 80 8 3 +I0 33 63 +27 28 2 1 +35 90 - 4 $46 +216 - I 0 $63 +219 - 4 1 +19 - 39 + 3 - I 2 0 61 -10 -20 140 87 186 -10 196 182 + 3 +213 +I4 +221 +255

+ +
4
-10

+ + + +

5 2

+IS

+201

+I11

+
0

+
+ I
0

+ + +
0 0

+ +
+- 4
I - 2

+ +
o
- 5

+I1

I04 + 7 + I47 + + 78 - 3 + 86 + 18 - 6 -5
+223

5
-16 -8
0 0

West). Chart minus Computed. Unit o O . 1 +IO" oo 1 0 ' 20' 3 0 ' 40" + 3 -4 -9 -13 - 1 0 - 5 + 2 + I -10 - 5 - 5 -6 - I -3 -10 -7 + Z -10 -6 -12 -7 - 2 -4 + 3 - I -9 -6 - 2 - 5 -4 -6 - 3 -6 + 3 + Z 6 -10 -12 -13 5 -22 - 6 - 1 0 -13 -14 -16 -36 -9 -9 I -17 -12 - 5 -4 - 2 - 5 - 3 + 6 +IO + 2 0 0 0 +I0 +I4 + 4 + 2 o + I O . +16 $6 + 4 4 +I0 + 4 + I 0 + 2 0 + 3 + 2 0 + 2 + 3 + 6 + 3 +6 + 4 + 4 + 3 + 6 + 9 + 7 + 4 + 4 + 6 + 8 + 6 + 4 + 2 + 3 + 8 + 6 + 3 + 3 + 6 + s +12 +IO + 6 + 4 + 3 + 5 + 2 + 2 + 4 + I + 2 + 7 + z 0 + 5 + z 0 + 7 + 3 - I + 7 + 8 + 3 + I + 6 + 6 $ 3 + I + Z -4 - 2 -8 + Z - I + 5 1-5 $ 3 -7 -12 - 1 - 5 + 4 + I o -6 -8 -13 -9 - I - 2 -3 - 5 - 3 -6
+ Z
- 1

+ 52
+ I
- 2
- 2

23
0

-4 o
- 2
-11

32
- 3
- 5

- 5

- 5

-9
- 5
- 1

-6 -4
-15

4 6 6 7

- 6 - 8 - 6

+
-16
-11

-6 -7 -8

- 5

-6

- 3

+ro
+I0
-11

6 7 7 4
2

-8
-12

+ 3 -6 -7 -3 -4 6

+ 4
5
+ 2

D, Declination (+we. if East; -ve. 80" 60" 50' 4 0 ' 30" 20' 70" oo E. 0 -10 7 I -10 30 -4 I0 -8 22 - 5 - 3 -6 -4 - 3 20 0 - 2 - 2 - 2 -3 + 5 - 3 + 2 4 - 1 - I -4 + I 30 - 3 0 - 2 -4 40 - 7 - 2 -1 0 - 7 0 - 2 50 - 15 - 7 - 18 - 1 0 - 3 60 0 - 2 + 3 - 2 55 - 1.5 -15 -7 -6 I07 70 15 - 15 -10 - 5 80 I02 + 6 -4 -4 -8 -7 - 13 101 -I 2 - 2 - 2 - 4 - 5 90 - 3 I00 + I + 3 + 5 + z + I 87 + 3 II 0 I02 + 3 + 4 + 6 + 7 23 + 5 120 I24 - I + Z + 5 + 6 25 -7 130 I57 + 4 - I9 - 5 - 3 + 4 + 4 181 - 5 140 - 17 - 5 + I + 4 + 3 1 . 5 0 o + 6 + 6 + 3 I79 + 7 -7 I60 167 16 - 3 7 +16 +IO 4 . I70 E I42 +I0 +I1 9 9 1 4 + 4 I80 I0 + 8 + 7 + 9 I35 I 4 +XI I70 134 + s + 6 + 9 + 9 14 + 7 I 60 +I15 + 4 + 7 +I0 15 I 7 + 4 +I10 150 I0 + 3 + 3 + 5 + 8 2 9 140 I97 52 + 2 + 4 + 4 + 7 15 82 I30 2 6 + 3 427 + I + 4 + 7 I20 146 22 +448 + 5 + 3 + 5 + 5 II 0 160 +318 + 7 + 5 + z 1 9 + 9 I00 - 1 + 5 0 -495 573 - I 0 + 5 2 8 + 3 + 4 + 3 + I -529 -267 90 80 -8 - I + 5 + 3 -464 - 158 -44 + 3 + 3 + z 70 -275 - 130 -24 - 5 60 + I -27 - 6 + 6 + I 25 - 40 - 16 - 2 + 6 0 0 50 5 9 - 25 o + 3 -6 -8 40 5 6 - 3 -9 0 -I 0 - 3 o + I -8 30 54 20 45 - 4 - 17 - 8 - 6 - 3 - 3 0 10 1 7 - I3 - 9 -10 - 8 - 2

+ + + w.

- +3 -I +2
-15

-7

+ 8

+4 +9
+ 6

-7

-17 -24

1 8

38 3 6 43 60 58 46 2 6 16 I7

analysis of the Earth's magnetic field, for epoch 1942

425

"0

v)

em: 2 d m ~ mN O n n - W e m m n I l l l l++++l l l l l l l l l
~ n u m n ~ ~ ~ ~ ~ n v )

0 0

N m m N N O W t-W

v)

v)m

2k
~ ~ ~ m ~ ~

++++++++++++I
v ) ) t t m v , ~

l l l
o

"0

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
~

a
O +

+I
m-

I I I
o o
n

.........................
m+mn
N n n
n n

8\

n cI

N N

n N N

m u mmmrcl*N

+ I + +

IIII++++++++++++++++++++

+++ o +++++
I
V N N n n

n u

I I

IN

a o o o o o n n I n I - ~ - v ) e t - w em m o I l l I I I I I I I I I I I I I

+++++ o
n n N N N

I I I I

n n

-& m
50'

+lo"

70"

60" 60"
-33

50"

-h0
-9

0"

E .

+6

+8

+ 3

+
+
-20

10

+6
+XI
I -12
-10

- 2

30" 4 0 ' +ro +27

20
I

+5

30

$5 +5 +5

+6

+ Z + I + I

-3

+
-27
-28 -29 -33 -30

- 5

+ 4 -13

-38 -38 -26

-23

7 0 ' -28 -31

40

+6

+5

+
-5
-15
-11

3
-29
-22

+
-37 -36 -31

50

+5

+3

+ I

- 2

+2

- I

+I +2

+ I I

- 2

+5
o
+XI
+20
I

-17

8
9

-2
-16
-13

+ 15
+27 +34 +33

60 70 80 90

+3 +3 +3 3-3

-5

+ +
+ZO +ZI
4-21

+ +
+ 4 -8 +13 -7
+I7
+IS
+XI
+22 +I5 +IO

-27 -27 -24 -16 -8

-8
-11

-9 -7 -6 -3 +I +2 +5 +4
-9
-10

4
+ 2

2 Y

I00 II 0 I20

+2

4-1 +I +2

- 2

+
+16
+XI

+I -I

+2

-6 -6 -4
0

+4 + 8 +
- 4
- 1
- 1

+ + + + +
+6
+ 4
+ 2
0

130

+3
- 5

+ 9
+I0
- 1

-3 -8
I

+ 3 +5 +7
+IO +13

+2

+7
+ 4
I
-10
- 1

140 150
-

d-3
I

-3 -4 -4 -3
- 5

+I3

+I0 +I1 +I0

TABLE XI1 I, Inclination. Chart m i n u s Canrguted. Unit O".I 40" 3 0 ' 20' +IO' 0" -10' 2 0 . + 4 - 8 - 4 -15 -14 o +IS o - 7 -10 -15 -26 -3 9 - I - 7 - 8 - 6 -10 - 8 + 5 - I + I - 8 - 8 - 3 -11 0 -2 I -13 -10 -11 -12 -3 o + I -10 - 5 -8 -7 -3 -6 + 2 - 2 - 2 + 9 +I3 + 4 - 5 + I -I 2 +I3 +26 +I5 -5 I 4 5 +I3 +27 +27 -9 2 5 4 +I3 +28 +32 - 8 + I + 7 + 8 + 4 +20 +25 -2 -5 3 7 +I4 +I3 +I9 I -2 -2 4 +IO +12 +14 0 I - 5 + 7 + 6 +XI + 3 0 + 4 -4 + 4 + 4 + 7 + 4
+ 4
- 3

+3
+ 8
+I2
- 2
- 1

+16
- 5
I I

+ 8

+ 6
+ 5
+ 2
0

+3 +9 +5 + 8
+ 2

+17
+ 3
I

-2

+I

+14
- 4

-7 -6
- 3 - 4 - 3
0

f 7

-1

+I

+I1

+7 +2 -4 +3 -4 +IS
+ 2

+s +3 +
+ 2

+14 + 3 +I7
0 0

+I9 +19
+21

I 60 170E . I80 170 W. I60

-I

+5
- 2
I I

+ 8
- 5 - 5
- 2 + I

+ Z

-I -1

+ 3

+ + + +

5 3 4 4

+I9
- 1
0

+9 +9

+ 6 + 7

+ I + I
+ 2
+ I
+ I

+ 7
+ 3 + 9 + 8
+I1

+ I

+ 5

+ 3

+IS

150 140

o
+ 3 + 5

+7 + 7
+ S
+ 2

+ I

+ 4 + 7
+I0

+ 9 + 5 +lo

+6
+4 +7
+I2 -1-15

+IZ

+XI

+I7 +I7 +I7 +16

+I4
+I5

-1

-1

+ 8 +4 + 4 + z + 3
o
- 2
+ I + 2 - I

+ I

+ 7 + 6
+ 4

+ 3 + 4 + 6
+ 9

+I4 +I4
4-13

4-20 +20

130 120 II 0 100

-2

- 2

+9
+I7
+22 +20

+6
+ 3 + 3
+ 8

+14 +I3

+IO

+14 +17
+I9
+21

- I

0 + I

+ 3 + 4

+ 4
0 0

90
o
- I

+I +I

+2

+2 0

o
+ 2

+ I
+ 2
+ 2

+ s
+ 6
+ 2

+IS +I2

+IS +17 +IS +IS +17 +14

I

+ 6 + 4
+ 3 + 5
o
+IZ
+22

+8 + 7 +6
+IS
+20
+I1

+IO

80 70
+ 3 4

- 3

- z - a

-60

+3 +4

o +3
+XI

+ I

+ 2

+4 +26
+IS
+XI
+IO

+6

+4 +3 +6 +7
+zz
+ 6
+31 +30 +16 + 3

i-22

+I5
+I1

+I4 +I4 +I7 +I7 +IS

+I20

+5 +7 +6 +4 +6 +I
- 2

50 40

30
+ 8
- 2 + Z

+5 +5

+7

+IS +27 +27 +34 +28 +30

+21

+16
+25

+XI +6 + 8 + 6 + 7 +7 + +4 +I3 +7 +12 + 6 + 5 +3

- 2
- I

+ 7 + 3
+I4 +35 +35

20 I0

w.

+6 +6

+7 +7 +7

+ 6

+4 +6 +7 +8

+3 + +5 +9 + 5 + 5 + 7 +16 + 4 + 8 +19 +3 +ro

+ a

+IS +14 +30 +so +27

+ 9 +IS +19
+IS

-10

- 2

-6

-14

-18 -28

-10

-20

-4 6 7 8

analysis of the Earth's magnetic field, for epoch 1942

427

14. It seemed desirable to redetermine the coefficients, taking into account the spheroidal figure of the Earth. The formulae required for this analysis are given by J. C. Adams in Section VI, Part I1 of his Collected Papers, Vol. 11. They can be summarized as follows :Hnm = sinm8'Gnm,

Xmm =an+2[(n-m)sinm+lB'G,m+l-msinm-l8' co~e'G,m]/r"+~,

y,m

=an+2mSinm-1 B'G'nm/rn+2,

Z,m = an+2(n

+ I) sinm 8' G,m/rn+2.

The values of ln(an/rn) for values of n from I to 12; of lncos8', of In sinm8 ' for values of m from I to 10; and the values of G," are tabulated in Section 11, Part 11. r denotes the radius of the Earth and 8' the geocentric latitude. From the values of Xnm,Ynm, Znmare computed the primed quantities defined by Xnm =Xnm cos y5 + ZNm sin y5,

Y,"

=Y , " ,

sin y5 + Z,m cos 4, y5 denotes the angle of the vertical; the values of Incosy5 and lnsiny5 are tabulated by Adams. Unprimed quantities relate to axes along and at right angles to the radius vector; primed quantities to the normal and tangential components. The equations I;X,mg,m=x',,
Z',m = -X,m

r, yrnmgnm =y',,
I;z,mg,m
= z',,

with similar equations in h,m, are then used to determine the coefficients, the righthand members being derived from the tabulated chart data for the north, east and vertical components, as in the solutions for the spherical Earth. The combining weights for the different latitude zones, given in Table IV, were again used, and a single solution combining the data from all three components of the field was made. The values of the coefficients from the solution for the spheroidal Earth are given in the last column of Table V. The first six coefficients (gn0)are changed slightly, but the remaining coefficients are substantially unaltered. The solution does not add anything of special interest to the solution for the spherical Earth and it does not appear that the additional labour involved in the solution for the spheroidal Earth is justified in future harmonic analyses. 15. Comparison between the principal coefficients and those determined from other spherical harmonic analyses, including two determinations of slightly later epoch, are given in Table XIII. The coefficients glo, g20,g22 and h,l appear to change linearly with the time, h21 having the most rapid rate of change. The change in gll appears not to be linear and is slower now than a century ago. The values of g21 and hll show a considerable scatter but no appreciable change with time is apparent.
G 31*

428 H . Spencer Jones and P. J. Melotte, The harmonic The values of the other coefficients have been compared with their values as found by Adams for the epochs 1845 and 1880, by Dyson and Furner for 1922, and by Vestine and others for 1945. The coefficients of a number of the higher harmonic terms show appreciable changes in the course of a century; gso,g31, g41, hg1,g42, gS2, gS2, he2 are those that show the largest and most systematic changes. The fact that many of the coefficients change so rapidly is an indication of the complicated structure of the secular change field.
TABLE XI11 Values o f principal coefichts
& o l !

Erman-Petersen Gauss Adams Adams Schmidt Fritsche Dyson&Furner Jones&Melotte Afanasieva Vestine&Lange

1829 I835 I845 1880 1885 1885 1922

1942

+3 201

+ 12
-

gao

g1l

gal

+284
311

-445
505

I945 1945

3235 3 219 3 168 3 168 3 164 3 095 3 039 3 032 +3057

- 76
I3

+ 12
-I-

gas

+ 73

75 53 133 176 187 +190

491 514 222 481 7 . 4 1 495 226 518 218 509 229 498 +210 -512

278 243

4 53 56 59
125

hl' -601 625 578 603 595 591 592

555

- 127
136 116 129 129
123

haa

135
130 -141

73 44
42

590 -581

- 46

16. The first-order harmonic, which corresponds to the dipole field, is given
by Writing

glosin h + (g: cos C $ + hll sin 4) cos A. glo= H, cos 0, ; g,l= H, sin 0, cos do; hI1= H, sin 0, sin 4,)

H , gives the intensity of the dipole field ; 0,, doare the colatitude and the longitude of the northern pole of the dipole field, or the northern geomagnetic pole. The southern geomagnetic pole is the antipodal point to the northern : these poles are t o be distinguished from the magnetic dip poles. The values of H,,0, and do are given in Table XIV. The positions of the two poles from the present analysis are 78O.9 N., 68O.5 W., and 78".g S., 111O.5 E.
TABLE XIV

T h e Dipole Field Epoch H O

Erman-Petersen Gauss Adams Adams Schmidt Fritsche Dyson & Furner Jones & Melotte Afanasieva Vestine & Lange
I 829

00
0

d
a

I835 I845 I880

I885

'326 9 '330 9 .3282 '323 4

'323 1
-3228

11.7
12'1 11'2

11.6
11'3 II ' 4

I885 1922 I942

I945

I945

-315 9 '309 7 '309 7 '311 9

11.6
11'1

-64.7 63' 5 64' 3 68.0 69' 5 67-8 69.1 68-5

68-8

1 1 . 1 11.4

70-0

can be represented by the formulae: Ho =0.318 7 - 0.017 o (T-I~OO), e O = ~ ~ o ~ (T-I~oo), ~ - o o ~ ~ -do =67O.8 +4"-5 (T- goo), where T is measured in centuries.

The values of

H,,, 0,)

GEOPHYSICAL SUPPLEMENT

VOL.6, PLATE 4

Computed isogonals for north polar regions

sir Harold Spencer ]ones and P. I. Melotie, The harmonic analysis of the Earih's magnetic field. far epoch 1942

GEOPHYSICAL SUPPLEMENT

VOL.6, PLATES

Computed isogonals for south polar region

Sir Harold Spencer /ones and P. j . Melotte, The harmonic analysis of t h e Earths magnetic field, for epoch 1942

analysis of the Earth's magnetic field, for epoch 1942

429

It will be seen that during the period covered by these analyses the intensity of the dipole field has progressively decreased, the rate being about 5 per cent per century. The latitudes of the geomagnetic poles have not changed appreciably, but the northern pole shows a progressive movement westwards, and the southern pole a corresponding movement, of 4"'s of longitude per century. 17. The geomagnetic poles, on the dipole axis, are not the same as the magnetic dip poles, the two points on the Earth's surface at which the horizontal intensity vanishes and the inclination or dip has the value of go". I n the construction of the charts, the positions adopted in the 1942 charts for the north and south magnetic poles were 71"N., 96"W.; 714"S . , 151"E. respectively. The position of the north magnetic pole is practically the position determined by Amundsen from observations in the Boothia Peninsula between 1903 November and 1905 May and is within one degree of the position assigned by Ross in 1831, the agreement suggesting that the movement of the pole is slight. I n the course of the preparation of the Admiralty Magnetic Charts for x922, a complete revision of the north polar area was made by one of us, by constructing first the projected lines of magnetic force.* The following comment was made : " It will be seen that the revision has considerably reduced the residuals which, however, are still systematic in their nature. It does not seem possible to reduce them further whilst adhering to the position which has been adopted for the magnetic pole. A better fit could have been obtained if a position about 2" further north had been adopted." Nevertheless, as there had been no Observations in the vicinity of the north magnetic pole, the Amundsen position continued to be used, not only in the British charts but in those of other countries. The south magnetic pole has never been reached and there has been no detailed survey round it. The position assigned for the 1922 and I942 charts was based on the observations in the vicinity of the magnetic pole obtained by Shackleton's expedition of 1907-09, and ,Mawson's expedition of 1911-14. 18. Charts for the polar areas of the different elements were constructed by computing the values of the elements at intervals of 24O of latitude and of 10" of longitude, from latitude 67+"to the pole; from these charts the positions of the magnetic poles were derived both for a spherical and for a spheroidal Earth. Positions were assigned by Dyson and Furner from their analysis of the 1922 charts, but no information is given as to how they were derived. The comparison between the assumed and the computed positions is as follows :S. Mag. Pole N. Mag. Pole 7 1 g S., 151"E. Assumed (observations c. 1910) 71"N., 96"W. 71"S., 151"E. 75"N., 100" W. Dyson and Furner (1922) Jones and Melotte (1942) Spherical Earth 77"N., 103+"W. 71' S.,1504" E. 76"N., 102"W. 70' S.,150"E. Spheroidal Earth

The differences between the assumed and computed position of the north magnetic pole seemed to be appreciably larger than would be expected from the general uncertainty of the chart data, while the comparison with the determination by Dyson and Furner suggested that this pole had moved in a direction

* H.Spencer Jones,GeogruphicalJournal,61,419, 197.3.

430

The harmonic analysis of the Earth's magneticJield, for epoch 1942

slightly to the west of north since its position was determined by Amundsen. For the south magnetic pole the assumed and computed positions were in as good agreement as could be expected from the uncertainties attaching to both of them. 19. When plans for the north polar flights of the Lancastrian aircraft Aries from the Empire Air Navigation School, Shawbury, were under consideration in 1945, it was therefore suggested that a flight should be made over both the Amundsen position of the north magnetic pole and the computed position. The data provided by two flights in May 1945 indicated that the true position of the pole was probably between these two positions but nearer to the computed position ; the inference was that the position of the north magnetic pole in 1945 was approximately 74"N., IOOOW. The flights provided the first definite information that the pole had moved considerably northwards since Amundsen's observations in 1904. Since that date, independent confirmatory evidence has been provided by ground observations in the Canadian Eastern Arctic, arranged by the Dominion Observatory, Ottawa. In a report made in 1947 May to the Royal Astronomical Society of Canada, R. Glen Madill stated that the Dominion Observatory had for many years been fully aware that the magnetic pole was travelling in a northerly direction, this conclusion being based on the convergence of the isogonals based on observations made periodically at a number of repeat stations extending from Newfoundland to Alaska. During recent years the network of stations occupied by observers from the Dominion Observatory has been extended considerably northwards. I n 1946 observations were made in Denmark Bay, Victoria Island, and Fort Ross, Somerset Island ; during the summer of 1947 a full-scale airborne expedition made extensive observations at ten stations throughout the Northwest Territories, six of which were on the islands in the vicinity of the north magnetic pole. The mean pbsition of the magnetic pole derived from these observations is 73i0N., IOOOW. This position will be adopted in the Admiralty magnetic charts for 1955 which are now being prepared at the Royal Greenwich Observatory. 20. I n each of the polar regions the system of isogonals possesses two singular points, the geogfaphical and magnetic dip poles. The structure of the isogonals is of some complexity and very few observational data have been obtained in recent years. The isogonals can best be constructed by computation from the harmonic analysis. I n the two plates the isogonals derived in this way are depicted for the north and south polar areas. The positions adopted for the magnetic dip poles are the computed positions ; though for the north polar area the computed position and the observed position are somewhat discordant, the charts can serve as a guide in future magnetic chart construction, being adjusted where necessary as observational data become available.
Royal Greenwich Observatory, Herstmonceux Castle,
Sussex : I952 A @ d 29.