This action might not be possible to undo. Are you sure you want to continue?

# UNCLASSIFIED

AD 277 479

ARMED SERVICES TECHNICAL INFORMATION AGENCY ARLINGTON HALL STATION ARLINGTON 12, VIRGINIA

UNCLASSIFIED

NOTICE: When government or other drawings, specifications or other data are used for any purpose

**other than in connection with a definitely related
**

government procurement operation, the U. S. Government thereby incurs no responsibility, nor any obligation whatsoever; and the fact that the Government may have formilated, furnished, or in anY way supplied the said drawings, specifications, or other data is not to be regarded by implication or otherwise as in any manner licensing the holder or any other person or corporation, or conveying any rights or permission to manufacture, use or sell any patented invention that may in any way be related

thereto.

o

@

277 479

AFCRL-62-341

P

R

PART I

--

**STUDIES IN IONOSPHERIC PROPAGATION
**

The Exact Earth-Flattening Procedure in Ionospheric Propagation Problems by M. Katzin PART II and B. Y.-C. Koo

--

VLF Signal Enhancements and HF Fadeouts During Sudden Ionospheric Disturbances by M. Katzin

Final Report on Contract AF19(604)-7233 Project 5631 Task 563109 Prepared for ELECTRONICS RESEARCH DIRECTORATE AIR FORCE CAMBRIDGE RESEARCH LABORATORIES OFFICE OF AEROSPACE RESEARCH UNITED STATES AIR FORCE

BEDFORD, MASSACHUSETTS

**Report No. CRC-7233-1 15 April 1962
**

ELECTROMAGNETIC RESEARCH CORPORATION

500 COLLEGE AVENUE

S2-

COLLEGE PARK, MD.

I

AFCRL-62-341

0

STUDIES IN IONOSPHERIC PROPAGATION

PART I -The Exact Earth-Flattening Procedure in Ionospheric Propagation Problems

by

M. Katzin and B. Y.-C. Koo

PART II

--

**VLF Signal Enhancements and HF Fadeouts .During
**

Sudden Ionospheric Disturbances

by

M. Katzin

Final Report on

Contract AF19(604)-7233 Project 5631 Task 563109 Prepared for ELECTRONICS RESEARCH DIRECTORATE AIR FORCE CAMBRIDGE RESEARCH IABORATORIES OFFICE OF AEROSPACE RESEARCH UNITED STATES AIR FORCE BEDFORD, MASSA1CHUSETTS

ELECTROMAGNETIC RESEARCH CORPORATION

5001 COLLEGE AVENUE

COLLEGE PARK, MO.

Report No. ORC-7233-1 15 April 1962

Armed Services Technical Information Agency Arlington Hall Station Arlington 12...S. and other Government agencies should be directed to the. D..C.. Department of Comerce Office of Technical Services Washington 25. . All other persons and organization should apply to th U. their contractors.Requests foT additional copies by Agencies of the Department of Defense... Virginia Department of Defense contractors must be established for ASTIA service or have their "need-to-know" certified by the cognizant military agency of their project or contract.

2&. da i prz4 spherical Han..al fuan.in t.-e ..he -s .s of Bess-Fel ftcnctaoni.'. 1The~ C.± cr. a1 stratifica4 J cvi. tt . frm which may be derived either a rlay-optical series or a normal rlodE series.tI DI as an -Mae t~jse c.functi-on of D.nyb bne pc Atdu.funec!-io± ard ts.blem is formulated as an integral representation. A nph-3-ctdal geoiLetry is Jinvestigated as a ba.eloped fo-r an isotropic spherically-stratified atmosphere. In the latter case. these functions is derived w'hich is niot of asyrartoic clhaacter.derfoimid by. of sJhe:CIC3L cas "lie -..L f~ -U'cat: T nce-::: tions being infliaiite d~stance to radius. Ani ir4Tpro-ved method o.NG ROCEDURE IN IONOSPHRC PROPAGATION. the normal modes involve the normnalized spherical Hankel Punction and its deivative.THE MXAT EARTH-FLATTNI.-ion as a o h ~m A outcn. for dealina witb pr oms Solationo for tin anglar ftinr_. is extendnd to the c:ase of a srpherical earth and atmosphere eniveloped by a sharp1y bounded ionorsphere.. i~ m --n tezz- -f ox..nite sex3. the ri~ i.r t-e J-ti) c-f ICIZI-fooall f bci .. ThOBDiMS The exact earth-f-L-atterdng procedure previous ly dev.CiV-C. The general solution of' the pro.a-_e f:cnsJ.relcpad 7or 'he .3pherica. sui of +be aci'. of ixon-spheri.e tha zeroz7 of -ile -cad.C.alloed S :in-. rla r.e t it s-as '1e. obtaining thi-e zercs oZ..

appears that an increase in the shapness of the lower boundary of required during the onset of a solar flare The mechanism by the D-region is which this takes place needs to be determined 0 iv . This reduced height causes reflection to take place at a the effecConse- level of higher collision frequency. while the v-l-f enhancement -s occasioned only by the changes at the lower level. absorption is Since the Increase in h-f the svm of the increases in the two r-egions. an adequate explanation of the mechanism of the v-i-f enhancement is not available on the basis of present knowledge° Phase measurements show that a definite decrease in height of the lower boundaiy of the D-region is caused by the flare..I PART II VLF ENHANCEWNTS AND HF FAIDEOUTS DURING SUDDFN IONOSPHERIC DISTURBANCES ABSTRACT Simultaneous observations Of short-wave fade-outs of a 13o5-Mo/s signal and sudden signal enhancements of a 31o15-kc/s signal over substantially the same transatlantic path of approximately 5400 km show no evident correlation between the magnitudes of the two effects of the SID. no correlation should result between the two effects.tted during a flare. which can be expected to vary from flare to flare. is understandable on the basis of a two-laye' This absence of correlation D-regiono The relative intensifications of the two D-regions will depend on the spectral distribution of hard X-rays in the 1-10 A range emf. On the other hand. it the ionization gradient remains the same. which should result in a decrease in tive conductivity of the layer if quently.

2. NON-SPHERICALLY STRATIFIED IONOSPHERE 3.1 4.2 H-f Effects V-l-f Effects 4. 4. 2.3 2°4 2 5 Formulation of the Problem The Angular Fmction T The Radial Function U Evaluation of the Integral Representation The Complax Zeros of u ( 2 ) z) 1 2 3 6 9 13 16 3.1 2.2 Formulation of the Problem The Angular Function T The Radial Function U 22 22 23 24 3.1 Short Distance Characteristics 33 36 37 IV .PART I iii ABSTRACT - PART II PART I iv 1. SUMARY REFERECES PART 11 28 29 1 2o INTRODUCTION DESCRIPTION OF MEASUREMENTS 30 31 3. INTRODUCTION SPHERICALLY-STRATIFIED IONOSPHERE 2.3 4.1 3. RESULTS DISCUSSION 31 32 4.Table of Contents Page ABSTRACT .2 2.

3 4. 6. Eclipse Effects !4 3 D-Layez Friduction and Structure 42.4 The Dio-Layer Model Bracewell's Exhaustion Region Ionizution Mechanisms 43 44 4 46 Comparison With SID Results 4.1 Absence of Correlation Between Magpitudes of SWF and SSE 4.4.71 vi .2 MechamrLsms Associated With SSE 47 47 5.23 4 'oyag Distance Characteristics SID ]Pfects 39 41 42 43 ~4..Eate i4.2 Z3.'IGURES 1 -- 50 52 26 (PART II) 58 .. OCOVIUSIONS RIELI RAPHY .1 647.2 4.4.

an exact earth-flattening procedure was given for propagation in an inhomogeneous atmosphere over a spherical earth. but should not be of great consequence in problems of tropospheric propagationIn the case of ionospheric p~opagation. involving complex layer distributions. In particular it was shown that the differential equation for the height-gain function in the usual earth-flattening approximation was equivalent to a small.3.t of the contract. The subject of ionospheric propagation..y is considezedo This is n'to account lateral variations of the reFor this purpose a spheroidal atLifcation). ThMs formulation led to the realization of the physical nature of the approximations introduced by the usual earth-flattening procedure. INTRODUCTION In an earlier paper Cl]*. for never will be capable of a complete self-contatae: *Numbers in brackets refer to the corresponding numbers in the References on po 29° . encompassea many ramifications which probably t. arbitraX7 angles to the earth's magnetic field.--eatmento Consequently. change in the refractive index veiation with height. it appeared desirable to investigate whether the exact earth-f-lattsaing procedure could improve ionospheric propagation analysis.ably g: .: ied out in Sec. In other words. coupling between modes.ter Conaeqaexutly.be important heights involved (in wavelengths) may be conside. the physical problem is The amount of this change changed somewhat by the earth-flattening app. and is accomplished in See.PART X THEXACT EARTH-FLTTENING PRO0DURE IN IONOSPHAIC PROPAGATION PROBLEMS 1. ca. atc.tWive index (non-horizortal geomet-.-oximationo or deviation increases with height. the extension of this theory to take frao. magneto-ionic splitting and propagation at. t. This is one ob ective of the research conducted under this 2 Ln additional objective is pa:i.

since this is the only effective form of radiator at these frequencies.purposes of the present study we shall adopt an often-used idealization of the ionosphere in order to confine attention to the specific objectives stated above. for example. it is a form adapted to direct introduction of the earth-flattening possible to formulate separatecorresponding the ly the cases of vertical electric and vertical magnetic dipole sources. In the isotropic case treated by Friedman. Actually this Hertz vector (within an appropriate mltiplying factor) is nothing more th&w the radial component of the electric (magnetic) field in the case of the radial electric (magnetic) dipole source. For this purpose the ionosphere will be considered to be sharply bounded and of uiform electrical properties. however. so that the results will be of chief interest in this frequency range. respectively. SPH IGALLY-STRATIFIED IONOSPHME A rigorous formulation of the field due to a vertical electric or magnetic dipole in an inhomogeneous isotropic atmosphere over a spherical earth was given by Friedman [2]. 2 . Scheikunoff [4]). anisotropic case. In each case. to vertically and horizontally polarized fields. various field components are derivable from a Hertz vector whose direction is radial. 2. It is then logical to consider only a vertical dipole source. will be given in procedure. a rigorous formulation of the spherical This formulation problem (with a shorp ionosphere boundary) will be sketched here. electric and magnetic modes are coupled in the ionosphere. since all other components are In the derivable from the radial components (see. This assumption is the one usually made in study- ing v-i-f ionospheric propagation. For plane geometry. For completeness. this was extended by Wait [3] to include the essential mixed polarization effects due to the anisotropy of a sharply bounded ionosphere.

1. excited by dipole source. 2.Geometr7 of spherical earth. I . - - Earth Fig. the boundaries of the earth and ionosphere. the boundaries of Ionosphere Air Vertical Dipole-. The ionosphere will introduce magneto-ionic splitting. in turn. so that new 3 . A vertical dipole of located at R = b. will give rise to reflected components at. Consideration of the physics of the problem will assist a proper formulation. earth and ionosphere being at R = a and R = h. Thus. respectively.1 Formulation of the Problem The geometry of the problem is (infinitesimal) length Z and current I is shown in Fig.so that the problem must be formulated in terms of mixed components from the outset. with concentric sharplybounded spherical ionosphere. the primary field due to the source will give rise to a field which has a polarization determined by the direction of the source current. This primary field.

6) and P is a solution of the homogeneous equation VaP + kP 3 = 0.. so that we must represent the radial Hertz vector by a column matrix of the form where JR is the unit radial vector. respectively. Then the corresponding fields are derivable fro the equations E. Consider first the electric component primary and a secondary field Ie =nj 11e and write it as the sum of a + n.polarization components will arise there. in general. and the subscripts e and m refer to electric and magnetic modes. (2. 3 (2. will differ in amplitude and phase. is a solution of the inhomogeneous reduced wave equation V7-2P + kp ' k W (2.-k R(PI +P) + ' grad [~R(P +P~ (2-4) providing that P. The two components.7) 4 .2) Now put I1I is stimulated by the vertical source current. while 11 arises from reflection 3 at the boundaries. From these facts it is clear that a combination of electric and magnetic Hertz vectors must be used for derivation of the fields.

9) T(e) U.9) are characterized by different %alu.ror t. e the right-hand side of (2.! . the solutions of (2. and V are functions only of 8: R. Hence we can separate P.8) Sin. (R)V() (2.12) SRz ~ d42X + Yr2'I in 2.. (2. T .to Its an integer of (2.11) + e Ati R ) 2 u 0 (2.tely will be fixed by the bounda-j .First consider (2. U.6).= 0. (2.0. The current density Ji -maybe related' to the dipole moment I by integrating over the source region: ll so that dv Jjj7JRksiflf i 8 ded~fdR._ 0. possibly.6) can. complex values 4 take m.rsn..rtiie of (2. phtcn I. and The various solutions of ultim. respectively. T=O. including. . stw-ts. -rvE .13) in the form .+ k P. which as yet are arbitrary. solutions of the corresponding homogeneous equation V 2 P.2-periodiclty in cf. separates into the equations 2 d2'T + cot (2 9) then JW +(.13) s and ah.onditio-se (2. in v.ch x are the separation c.10) where T.8).-s of s end m. be assembled from. is zero everywhere outside the point (b. and write thb. Sj $) (R-b). in the form RF = (2.6).0). and p.

~A )TOV. f sTU.18) It may be shown that a lower bound for the absolute convergence of (2.s).19) 6 .15) where the amplitude function A(s) and the path C iu the complex s-plane are as yet unspecified.16) where T is a solution of (2. o= 'sO so that this covers a sector greater than ± m/2. A(s) and C y be determined by integrating (2.11). (R)V. and Vm is given by (2. we must choose the cylinder function to be the BAssel function J.18) is lei = 2. (2.. (2.Consequently all solutions of (2. V. Consequently the required solution of (2.% nuo (2. = 1.11) for m = 0 may be written as T= Za. Introducing the new independent variable x = se (2.9) with 2ic-periodicity in 9 may be obtained from the representation RP. so that R .17) where n-1 in which Z (sS) is a cylinder function. We now extend this type of solution to the case m j O.6) around an infinitesimal region enclosing the dipole source. and a.2 The Annalar Function T in [1] it was shown that a solution of (2.17) have the property T(O) = 1 as required. ds.14). C will extend over an infinite range. (2. U1 is a solution of (2. It can be shown that A(s) = s and f= a goC f . 2.11) for m = 0 is TT. Xn order that (2. 1 4q)As.12). seP"3Tn(. In general.(sO). provided that T(O) = 1.

where Zm is any cylinder function.21) (2..1".22) the E being the Bernouilli numbers. n (X)i (2o26) (2. -j 's.] By introducing the function Cm. The second equation then becomes L(y 2 ) =a(xZ+m 2 Zi) = a. (Z.o).Nx tj'+ m2 t. x)] = Zn CM.20) in the form L-(-) + -XL ( - ( V- - o a where o-L = '* BF/(Zp. A solution of the first equation is 7o = ZS(x).. Cx).) = o. (2.23) we obtain I.11) becomes We write (2.27) 7 ..n (X) = X"Z. (2.[i(m=+l)Z-x~z=. we obtain the system of equations L(. Assuming a solution of (2.. L (Is=) = ai. IA: By equating coefficients of like powers of a..21) of the form % =.-) %1(f~ (2. (2.and denoting the dependent variable by y.25) which has the property L [C.

ao v (2029) Hence the solution of (2..=.() = ZAn (S6Znm+n ..n a:'e eavily obtained from the recursion formulas for the cylinder functions: X =. j 2.kaO A10 (Se (2.28) By induction.+ P.32) If we substitute (2.. Cm/. ]o Now using the property (2.34) Consequently.23).27). (58). x .)Cm. we infer that =. T= #.25) becomes L(y2) = xj=(m1l)C=. we obtain the recursion formula De CPL~l.30) into (2. the required solution of (2.) X21. (59)". we obtain ( +m 0 ItI ( D p where Ao = I.n.11) is T ZA. iszO "2. where C-) (m-p)!M+n+) p != (2033) = C. and use (2.AP C.21) should be expressible in the form Lj = ZA.30) The following recursion formulas for Cm.= (mt. (2-31) (2. D (Zp +l)m m+Zp. the solution of this equation is seen to be ..I4IM~ padJ (2.(2. By equating coefficients of like orders of the function Cmn on the two sides of tWis equation.r (2... Cnn.31) and (2-32) to eliminate powers of x on the right-hand side.o-C .35) S .

11) and (2. see _. 9 .35) the solution of (2. 56d-. =A.) U. +. The integral along the positive real s-axis in (2. is that a more accurate calculation is possible than by the use of the asymptotic expansion for the latter functions. s d This form. integral corresponding to Hz ( 6e-) make the substitution s'= se " In the r .3 The Radial lumctit U With T as given by (2. (2-36) where Ao = 1 and An is given by the (2.36) (2. instead of the standard expression in terns of the associated Legendre functions.) d5.The advantage of using an expansion for T in terms of Bessel functions.Y. and note from (2.. For this purpose both the electric and magnetic components of L] will be required.= P. Ha) (s8) cos (ni +-Y. (50? . is adaptable to evaluation by residues o: by stationary phase.12) that T and U are even functions of s. is to be fixed by the boundary conditions. (60) 3 co (. These require that the tangential electric and magnetic fields be continuous at R = a and R = h.. 2. Z M~O ' n 00(-O . whereupon the integral for that term becomes 1 1 ".9) is R[P.34). depending on whether a normal mode representation. U. The function U.37) (A2 3 in view of the fact that the integrand is an even function of so becomes ® RP.. IA.36) may be transformed into an integral along the entire real axis in the following way: Write JTn 4"s (69J] = J" (W-mn . or a representation in terms of rays is desired. s sds Then (2.

O. 2 10 . we expect a mixture of upgoing and downgoing waves in the region a<R<h. We then pick the two independent solutions of (2. vI = V = VS.12) to correspond to upgoing and downgoing waves.Hn = (2.1). Then P3 satisfies the homogeneous equation (2.(se".44) n = TnUnVn.38) VaP + kzPszO. being independent of 8 and q). A similar choice is made The total field in the various regions then can be derived from a radial P function which has the matrix form R[P] in which R R P (2.)Usds. f 88'A.39) ourt ( P5 ). and denote these by U. (2. The corresponding fields than are derivable from the equations Em .. .. n = 1. (Sef Rw(e4) The constants c and I are to be determined by the boundary conditions at R = a Corresponding to the pysical picture of reflection at the boundaries. respectively.4+3) ftI± TA.9) and (2.39) may be written in a form similar to (2. in (2. f.42) (2.(2) and U1 (') for U2 and U3 .) maO . and write 2 ~=kPAp 3 .Hence we now consider the magnetic component n.. lead directly to the statements TI = T2 = T .40) (2.3. (s. The boundary conditions. coS(mT47'mn)U 5ds.2. 40 cos(mtY..41) Solutions of (2. (2.36) as follows: RP. RP 3 and R = h. 'r"PP+ grad a (Rz)].

12). (2045) u = U3") + US(S)$ s and introduce the reflection coefficient at the ground f where 1. If we denote the two independent solutions of this equation by u(1) and U (2) . R I Rub-& . At the ionosphere the reflection coefficient is a tensor [e co that f. (2 .49c) .46) el and tpare the reflection coefficients for vertical and horizontal Then polarization.LZ= K9 Zirkt K while U1 itself is continuous at R = b. i. 2 us = u2() + u ( ).. L. ..ell [ (h) i'X ".)] + eLsiU (h).. e .48) U3(h). Us satisfy the same type of differential equation as U1.e. at R = b we have the discontinuity condition for the first derivative of U1 in terms of the dipole moment L2] dU. (2.=elz (&). J U21'Ch = [ h +%).. respectively. respectively. + P. (2. U1.49a) (2.49b) U2 * UO 'ba o*R<h. then we may write in the various height regions = U. (2. 1%)(h). where u (I) represents a downgoing wave and U( 2) an upgoing wave.. Stu)+ C b4I'h. The radial functions U2 ...47) ) US'. oXR( b.Now we put U1 = UL( 1 ) + UJ(2). (2. U21hjj Finally. C2 (2.

.51a) I =.54 6.S. t 0"a *€=). (2. a K/r 0i.e..'d 14 - M . (2-51g) where "(h)(5 I t/Cu)(b) (2.51b) . .' MMA. OxiWb)=4i'(b) + (2.I -. 12 . Uol 0<R<h.50d) . 46.50g) The seven equations (2o50a-g) are sufficient to determine the seven constants '91. £[p( (2-500) 6at. ~b + Sg u'~ '(b) + K.. (2.50a) eU UIOW/uCt)().52d) M = p .Ozu -{ (h) (2.t) - .51d) . They are given by V. 4/1= . (2.) e..50f) (20. LIL' + g.US = f. (2.el *.- UW'M=)(b)I J1K u)b)..49d) The boundary conditions then yield 62/A ae.) 4 ea ... 14. . (2.

which are ratios that are functions of a. By the method of residues. type of solution in order to bring out the fact that the approximations usually made actually change the physical problem from that of a homogeneous atmosphere to that of a slightly inhomogeneous atmosphere. and (2. (2.43) give the values for RP in the space aR<_h.40). (2.primes denoting R-derivatives evaluated at the argument. the expressions (2.4]. the result may be expressed as a sum of rays reflected alternately a number of times from the ionosphere and the ground.42). (2°55) (2+ 4 )IM (2.+ (k = - SANti O (2°3 The solutions of this equaticm corresponding to downward and upward waves are the normalized spherical Hankel functions [5J 1. or waveguide-type waves. (2.54 u respectivey. 2.37). By the method of stationary phase.). the result is obtained as We shall investigate the latter a sum of normal modes. the integrand has poles at zeros of ths denominator in these ratios.4).56) With those functions inserted in (2. We now evaluate the form of the radial functions u(I) and u42 ) 0 These are solutions of U.4 Evaluation of the InteR niMpresentation Two different methods are available for evaluating the integral expressions for RP. Since the coefficients in the integrand (?h - ) involve the y-functions defined above.c) (kR) "( H. = ' s) '€(kR).49).k . on the other hand.e. where P = kk. Consequently. if we deform the integrand from the original contour along the real s-axis into the appropriate 13 . from which the fields may be evaluated by (2.5). (2. (2.

14 . the integral may be evaluated in terms of the singularities of the integrand in that half-planeo In addition to the poles just mentioned. terest in determining the normal modes are those of M. The matrix A[P] in (2.. the other hand. since e. similar to the way in which Bremer [6) treated the tropoand u (N. andem= a . p. on integral and has shown that it is negligible in practical cases. is zero. Consequently the integration path is deformed into the contour shown in Fig.44) has an integral representation vdich can be assembled frmi (2°37). This has branch ±i The integrand vanishes at infinite values of s in the lower half-plane. there is also a branch point where the order of the spherical Hankel functions. points at s = This can be seen from (2.43) by using the U-functions given in (2. in effect. poles of pi j. and thus is easentially a part of the ground-wave field. This integral represents the effect of the currents which penetrate into the ground. plus an integral around a branch cut along the negative imaginary axis from -i/2. Poles The principal poles of inThe investigation of these The poles is a separate problem in its own right which we shall not go into here. but his procedure. In the case of a perfectly-conducting ground the integral vanishes altogether. (2. of the integrand are those of the functions e.56). can be expressed in terms of the two limiting cases spheric case. is equivalent to neglecting this integral.42) and (2. ultimately can be expressed in terms of y-functions and the properties of the reflecting medium. tively.49). respec- These can be determined from the zeros of t(h) Thus we consider the method used for the determination of these zeros. attempts to avoid the branch-cut integral by making a double traverse in the lower half-plane.II half of the complex plane.%h and M. 2. The integral then is the negative sum of the residues at the poles in the lower half-plane. Friedman [2] has discussed the importance of the branch-cut Wait [3j.

-ation Ccnto-u: in s-plane .Plcne Fig: 2 -Ink eg.s.

W.. upon putting w-w = u. shifted to the right by x/2o of variables t A simple change (zp) - (258) 16 .5 The Complex Zeros of u (2) (z) The zeros of u (2 ) and U( 2 ) # are the same as those of Hp( 2 ) and Hp() These are found by the Debye method of steepest descent.). 0 and draw the contour W so as to pass through the two points (stationary points) 2 at which FI(w) = 0.±r (2..2.O.. Since F"(w. wo =/29 and F(wo ) = 0. (2.. By truncating the Taylor's series expansion of F(w) at the third derivative term. and are usually expressed in terms of Airy functions.57) 0.) + (Tw") F.5?) becomes o where the contour U2 is merely W. or Hankel functions of order one-third.du expand the exponent F(w) in a Tay1or's series about the point where F'(w) e W& a pw1/)dwW.) . we obtain F(w) 0 F(we) + (w-w. F'(wo ) = -i(z-p) F" (wo ) = iz° Consequently. procedure is to write The Nm)(. F'(w.-zcos we have .

plane LI Fig.55) - Contour for Mdified Hankel Functions (2061) Z (.results in In/" '' 6 4ee ] d"ta: (2. Using the notation for the latter. It was pointed out in [1] that the approximation (2.jr h2 (2. in first approximation. 3 Then from (2. 17 . while h2 (r) is a soluticn of Stoke ' O. This is immediately evident from the fact that equation (2 62) bp(2 )(z) it a solution of (2.ed in [7]) give. we obtain 12z) Hz "- e-/1 I. 3. the zeros of RpC2 )(z) and of hp(2 )(z).12). The integral in (2. e. or Mdified Henkel Functions of order one-third [7].59) where the contour L2 in the t-plune is shown in Fig.59) may be expressed in terms of the Airy Functions.- l a() Consequently the zeros of hg(i) (tabulat..61) is equivalent to a change in the physical problem.60) t.24)7q-7rVA.

we must have k~ze = ' / 11 = k&(1 +j). if we put 52 crLik. It is evident from (2. whereby the radial equation becomes d2J iu + k2. In order to arrive at this solution. we obtain. the radial equation becomes 2 a4U. Stokes2 equation d z. and q are constants.(R) V(V). however.64) 2/a in order to where k. k must be s function of Y which . + a solution of which -'s . + _L~ 22/ sU Next.63) that q - (2. satisfy the equation for small In this case.9) by writing P1 in the form P. putting Uo = uo e we obtain d2U'--"O . finally.& Ua 1 r/ ' + (ke ')- (263) dq2 where To reduce this to Stokes" equation.61) may be found as follows: We first siparate (2.5 Now by introducing the transformation q=a log(RIC. = T (e)U.).I The physical problem corresponding to (2.

67) has a radius of convergence which is limited by the next zero of F (w). corresponding to Stokes' equation.+_0..II satisfies (2. except that Ao = 1.65) If we put Rt' then we have a+ H = (i + .+h'(4 IB (2. this requires that k have the form 0_0_______(I+ ______(R __toq* (2. decreases monotonically with increasing height H above the ground level a.57) becomes precisely the exponent in the integral of (2. and Ursell [9]. t+ '" Then (2.67) Z POZO and integrating (2.= --(-W) in _-Z (Z-$IIIW-_ F(w) = . in general. whereas the original problem dealt with a constant refractive index. 2 To obtain a higher order approximation for the zeros of E (s (z).64). one may follow a procedure due to Olver [8] and Chester. which occurs at 19 . In terms of the variable R.65).68) n which the coefficients A. is. The above procedure is asymptotic because the series expansion (2. an asymptotic expansion is obtained in terms of h 2 () and h. Friedman.57) becomes By expanding R in a double series of the form Tr 9 qmtta~c'".66) termwise.' f()tA. dw 0 +tz 7j. whereby a change of variable is introduced so that F(w) in (2.(V) 3 Hp (z)'. inverse fractional powers of .(2.59): t le. Thus the refractive index (2. and Em involve.

The integral (2. We write This does not appear to have been reported previously F(w) = t * t s+at.-s ' (p/z).68).a Termwise integration then yields an expression of identically the same form as (2. for given z.) and B(.) being polynomials in x of degree m/2 or less.) I z =C i* V&'/6Z' T/.69) that the zeros of HW (z).)} (2. differ 20 .6 f d.57) then becomes I V &.69) I+ f MRS ~ m z) Am(. An alternative evaluation of (2-57) which is not of asymptotic character may be developed. while the interval of integration extends to infinity. e eI and expand e in the absolutely convergent series M Mao M1 wn 90 =r"t-Lm PWt .(2. h 2 (g) + 3(g) h )-.I + . Then hp (Z) " p (.fI{() + a-. and f£ (iz/gr A24 F(")w n! t. Next we write e 1~w = e.*+t/3e . bin V".70) It is evident from (2. (Z) where Z . where i and t are given by (2 0 58).I w -. however.d.

72) and (2.69) we then find (2.+r.3 h (). then may be This is a series in EV whose coefficients are known. corresponds to a value of r. . obtained by successive approximations.) {.which we denote by =4Z * So "1" -0 so that V.58).) t . Then by (2. 21 .. where g is related to z and p by (2-58)o In order to find the values of p for which Hp" (z) is zero.. ) end a. we can proceed as follows: Denote the zeros of 'N() bylgo.73). H(2 (z)= 0.o We now expand o.&0. and make use of (2. by. which as yet is unknown..ha'==(. .69) (2. is small compared to go..)h.71): (i"+ + + . h in Taylor's series about to. Then from (2. ioeo. so that h2 (go) = 0. From (2.slightly from the zeros of b2 (g).(r. (2. and put p = - qo (2.71) H(a (z) - '6 so h" i 0.73) H 'z)= =. (C') We denote this zero 2) The value of p corresponding to go is near a zero of H( (z)..72) T. -r "} Consequently we obtain 11 +q The value of 1.

so that the above type of analysis is an idealization which should be considered as only a first approximation to the true state of affairs. these are 22 . In order to introduce a form of non-spherical stratification which may be applicable to such situations. (z) then are p= o-q* (2. For the oblate spheroid.l Formulation of the Proble The reduced wave equation VZP "t P may be separated in spheroidal coordinates into radial and angular differential equations as in the spherical case. The detailed results obtainable by this procedure will be reserved for a later investigation. 3. NON-SPHERICALLY STRATIFIED IONOSPHERE In the treatment in Sec. For example. the earth-ionosphere region was assumed to be spherically symetrical (so-called "horizontally-stratified" medium). where the earth and ionosphere are coordinate surfaces of a family of spheroids.70). 3. we consider the case of a spheroidal geometry. either oblate or prolate in form. in general.we obtain I The required zeros of e Irt.74) These likewise are the zeros of the modified spherical Hankel function N2 (%) given by (2. there are situations of practical interest where the reflecting layers are tilted with respect to the horizontal. We give below the extension of the exact earthilattening procedure to this non-spherical goometry. 2. 4. This situation is not strictly true.

& -l'ftr Is. -sz + '-\T -o(31) (+ coordinates where f is the semi-focal distance. C05 V = f Cosh go (P lh The corresponding equations of the prolate spheroid are d2(U ATtcote IM + F(T k2f5a - 2~.5CzA) I C'CIXXI+Of r 5 23 .+ Moii 6) + so-2 = 3.~r vs)] LL sa C. 2 dT ot& dOT .(ki+ 2 -f + U =0. Introducing the new indepen- dent variable x = s&.2 32 for which a typical space point has the rectangular coordinates K -f ih F.se = f e_ We shall treat the oblate case in detail.l) and (3. 2.2) can be effected by simple transformations.(k' fsire r)1V= 0.2 The Angular Function T We consider first the angular function T.2. as in Sec. the second equation of (3.2) shows that a change from (3o1) to (3.coss9C0p. and a typical space point has the rectangular x .ie e.cp j5 . I'm 9 W4 I Y af 51VI Z h4 51.1) becomes dTT -t Is aa ay Te 2 This may be written as L(~T4jT' L(7 . since a comparison of the corresponding equations of (3.+coshF. 3.LT+ "' a ri r lct(.

y applicable to (3..21).iution of (3. Z a. integral representation of the fo -mgiven in the normal mode solution. and differs fr~xn it only in the presence right-hand side of the iditiona. fields will be obtaiiable in terms of an (2.35) is direct.3. Our aim will Then Ia to obtain a solution of this equation similar h.im e d i a t e ly pr e c ed i ng it i in (3 . These residues. 9'&/ !(305) (3. are n4d An'P.4) as T. Consequently.1). we shal IL seek solutions of the radial equation similar 124 . and its derivative.eJfore the form of solution gLve which thus has the solution in (2. F#4bP1KA .3). which Satisfies the first equation of (3.6) on te (34) is similar in form i o (2. Con s equ en t l y £l. (3.8) OL"II +- T. (L a . term bThis term has th 4 s avre p owe r o f x a s t he ( w+ we can immediately write the . wave. (310 303 The Radial Function U We now consider the risdial function U.4 ven by the recursion foiriula I. l). which i{ representation. ultimately may be based first on the zeros of the function U wbl ch represents an upgoing:. 2. b)I' p gl-.4 ) . We !4hall be interested in to that found in Sec. A * )'L+4 (30 9) The:.b . where the coefficients . . obtainable from the re:didues of the integral in the spherical case.36)...

(3. To eliminate the first-derivative term. and then will investigate their complex zeros as a function of order. 4 &L '3.12) Then (3.~h4ALe+3 j We seek to cast this into a form which resembles the spherical equation. The radial equation in question is a+ ta. we put .14) a = kf.+ .17) in the form S-+ 1/ ZN.16) VOY+ -s9 * -L ) L(0-)-") + SL e + 3 Cae-e)e 0.a L -Z! -+ I Ou/ - ) 0. 4+z Mr. (U11 We first note that the transition to the spherical ease is effected by allowing fcoshC -*.ZZ(3. (3o17) We now rearrange (3.+ (e ) x.ed to introduce the change of independent variable f cosh C and the now dependent variable u= RU. (3o15) primes denoting derivatives with respect to s.54) and (2.t 25 .to those given in (2. whereupon (3.R as f-*Oo Hence we are . R.16) then is replaced by (3.13) becomes U.55) for the spherical ease. we put z = kR9 (3.11) becomes dI9L + + [kR -1 + 3-3 Next.

18).e. Z.54) and (2. Hence we seek a similar solution to (3. The form of (3. and *- (3. reducing by means of the differential . + coefficients A.18).6 isZa soBtyno . but we shall find it more convenient to deduce directly a special form which is suitable for the normal mode problem.20) into (3.18) were zero (i. a = 0)..is a solution of the normalized spherical Bessel equation L(*) = 0 and Substituting (3.20). for example.0V.4- it 10( ]A. solutions of the equation would be the normalized spherical Hankel functions given by (2.=_Y (3. [10]).18) suggests a series solution in a/zo formed in the form Such a solution may be f.21) (3. and By in (3. then the first few coeffi- A 40 39 M4 4aL 26 . [ I __ 1 + -VA.. These are two simultaneous equations which comprise recurrence relations for the If we choose Ao = 1.= _ a). and equating coefficients of (Z) on both sides of the equation. Solutions of the radial equation as a series of solutions of the spherical Bessel differential equation are available in the literature (see.I or L 0-) where L~ t- (3.[%r*B where Bo = 00 equation for /j to terms containing only /k and like powers of 1.. cients are Zv(2V+1 (?..22) 4V52.18) -'[+(3o19) If the right-band side of (3. +0 a .20) and 40 wheres.. we obtain the two equations 1 .55).

(3. 27 . Introducing (3.7) (3.~F~~ - f7~' PI~'~J a~s + o(6) +(3-28) where use has been made of (2.26) We can reduce this problem to cue of' exactly the same kind as solved in From (2. ()hpN + 26(m) h m.28) into (3.29) where a~'~ (3. we have found a solution of (3.).Thus combining (3.26).27) h~(z - 24)~ZI.16).(.62) to eliminate h2 (.29) now is of the same form as (2. the only change required being the replacement of c((*) and P~(r.15) of the foz u(z) .29). we obtain (1) Upi (m) a (24) i. we am replace h( (z) by a suitable sum of h.30) Consequently the procedure by which the zeros of (2. we choose the function tto be the normalized spherical Hsakel function 4" or he Then in finding the normal mode solution for the spheroidal problem we are led to a determination of the zeros of the function cz Sec.) by o (() and 6.respectively.( ZV&e4 5'/ fcqhsC) + 16(aohs 1(6)].70) were found may be applied directly to (3.).27) and (3.) and z [. as follows: hp From this.wclk4. 2.(3.* Al er50 B" £1 (9)h('.23) 6(a) =(3.20) and (3.25) In Order to conform to the type of integral representation given for the spherical case in See.5.70).) * A )Wx )03 (3. 2.(). (24)-% it'/& (3.ca) + s where w al~) VV a) o)jcx.

the coefficients of these functions being infinite series in terms of powers of the ratio of semi-focal distance to radius. The radial function is expressed as a sum of the solution of the normalized spherical Bessel equation and its derivative. the normal modes involve the normal- ized spherical Hankel function and its derivative. An improved method of obtain- ing the oeros of these functions is derived which is not of asymptotic character. from which may be derived either a ray-optical series or a normal mode series.4. developed in Li] for an isotropic spherically-stratified atmosphere. and carried out in detail for the oblate spheroid. In order to deal with problems of non-spherical stratification. a spheroidal geometry is investigated. The developments for the spheroidal case are pursued in a way similar to that for the spherical geometry. may be found by the same procedure that was developed for the spherical case. It is shown that the zeros of the radial function as a function of order. In the latter case. SUMMIARY In this report we have shown how the exact earth-flattening procedure. Solutions for the angular function are found in the form of an infinite series of Bessel functions of the same type as tound for the spherical case. . The general solution of the problem is formulated as an integral representation. may be extended to the case of a spherical earth and atmosphere enveloped by a sharply bounded ionosphere. which are required in the normal mode solution.

New York. Friedman and F.. 317-350. "An Extension of the Method of Steepest Descents". 247. Olver. No. 153-204. [2] B. Koo and M. Friedman. "Propagation in a Non-homogeneous Atmosphere". pp. 61-64. Vol. March-April 1960o [4] J. Inc. NBS . Stanford. Mass. on Pure and App. "Tables of the Modified Hankel Functions of Order One-Third and of Their Derivatives". Soco. Slater. t7] The Staff of the Computation Laboratory. 53. "Terrestrial Propagation of Very-Low-Frequency Radio Waves A Theoretical Investigation".. B. Res. Radio Propagation. "Microwave Transmission". Phil. Flamer. 1949. R. Vol. 599-611. 1. Jan. Res. Ursell. 2/3. p. Harvard Univ.REFERENCES Li] B. Wait. Trans.. [3] J. Inc. Soc. Schelkunoff. Comm. NBS . C. [10] C.. Bremer. 0 [9] C. "Spheroidal Wave Functions". Vol. Katzin. 1.. 64D. 2. Press. Inc. 1957. J. Jour. 328-367. pp. "Advanced Antenna Theory". 1945. 64D. John Wiley & Sons. [6] H. 1960. 29 . No. Roy. 1952. New York. Calif. Stanford Univ. Camb. Vol. Press. pp. "Terrestrial Radio Waves". 8. McGraw-Hill Book Co°. W. Elsevier Publishing Coo. Vol0 IV. Cambridge.. Jour.-C. [5] So A.. 1951. 197-199. 1942. Series A. pp. pp..D. 1957.-Feb. No. New York. Phil.Do Radio Propagation. PrOo. Dec 1954. Math. "The Asymptotic Expansion of Bessel Functions of Large Order". [8] F. pp. Chester. "An Exact Earth-Flattening Procedure in Propagation Around a Sphere".

that frequencies from 27 to 40 kc/a showed the sudden increase. Bureau and Mairs [3] reported that abrupt short-wave fade-outs (denoted by SWF hereafter) usually were accompanied by simultaneous sudden increases in the strength of atmospherics received on very low frequencies (vlf). cation circuits. 52.I PART II VLF ENHANCBeGNTS AND HF FADEOUTS DURING SUDDEN IONOSPHERIC DISTURBANCES One of the spectacular phenomena of the ionosphere is the sudden ionospheric disturbance (SID). this effect being most evident as a decrease in reflection height They did not observe "any clear indication of a change in reflected Bureau [5] then pointed wave amplitude at the time of the phase anomalies" (SPA). Budden and Ratcliffe [4] reported that measurements at Cambridge component of the downcoming They concluded of the phase of the abnormal (horizontally-polarized) waves from GER on 16 kc/s showed an anomaly at times of h-f fade-out. but on 12 kc/s the effect was rarely observed. of the waves". and it study. Later. *Numbers in brackets refer to the corresponding references in the Bibliography on p. associated with the SID and concluded that the disturbance must be caused by solar ultraviolet radiation. investigated in which drastically affects high-frequency communireported by Mogel [l]* and later Dellinger sunnarized the various phenomena This phenomenon was first detail by Dellinger [2). quencies. They reported is One of the associated phenomena occurs on very low fre- this phenomenon that forms the subject matter of the present that atmospherics from all directions were reinforced simultaneously. In 1936. that an SID "has a marked effect at the level of reflection of the low-frequency waves (70 kin). 30 .

53 Mc/s) was selected. 3. also produced similar enhancement of v-i-f radio signals.I out that his observations on the sudden enhancament of atmospherics (SEA) ing SID showed that such increases were not obsWerved below about 17 kc/so An investigation was undertaken in 1938 to determine whether SID. the equipment was set up Some of the subsequent S1F were accompanied by sudden Consequently. RESULTS Sample records of a simultaneous SWF and SSE are reproduced in Figs. and a preliminary report of the results wae presented in 1947 (6]. extend- signal enhancements (SSE) of GLO. Inc. 1940. The purpose of this report is to present the essential results obtained. and continuous The recording of this signal was being carried out at Riverhead fon other purpoes . if so. great circle path length was about 5400 km Botb the GIA and the GLE equipments were calibrated at least once each day br means of standard signal generators. not been published. observations were continued. and respect to ionospheric layer to discuss the implications of these results witk structure and the modifications produced therein by the SID mechanism. ing over the period 31 October 1938 to 25 June :940. whether any quantitative correlation existed between the v-i-f and h-f The experimental phase of the investigation was completed in but has effects of the SID. 1 31. and. which had accompany- been shown to produce SEA. to record GLC (31o15 kc/s). For comparison of the v-i-f SSE withi SWF) the signal eeceived from GLH (13.e results. After several months' observations of the signal from SAQ (17o2 kc/s). 2o DESCRIPTION OF MEASUREMENTS The measurements reported here were zaCe at the Riverhead transcontinental receiving station of RCA Conunications. . since this signal traversed appro:imately the same path. with negati.

and that no evident correlation between the magnitudes of the two effects exists. change. (57. the characteristic behavior was a rather sharp initial Invariably. The largest GLC increase (14. was Conversely.3 db) on GLC. followed by a trough (or crest). the recovery was more rapid for the h-f signal. Fig. solar 32 . for example. 3 shows histograms of the number of coincidences between SWF of GLH and SSE of GLC during the period of the observations.5 db) was accompanied by only a small increase (2. although the magnitudes of the signal change varied rather widely from one event to the next. the deepest fade of GT.H associated with only a moderate fade on GLH. This shows a high degree of correlation. These records are rather typical of the data obtained. a considerable body of information has accumulated concerning SID effects. respectively. Examination of Fig. 4. and then a gradual recovery. 5 shows that in no case was the GLC increase as great as that of the GLH decrease. so that the proba- bility of a v-i-f enhancement is very high if the h-f effect is pronoucedo Fig. 5 represents a test to determine whether any correlation exists between the amplitude ranges of the v-i-f and h-f signals during an SID.1 db). The points are plotted with the increase in GLC signal (in decibels) as abscissa and the corresponding decrease (in decibels) of the GLH signal as ordinate. Fig 4 shows similar histograms of the number of GLH fades of intensity classi0 fied as fimoderate" or greater and GLC enhancements which occurred during the same period of observation. and of GLH SIF over a longer period 0 Coincidences were observed only during the daylight hours when the h-f fades were more numerous. In general.and 2. DISCUSION In the years since the observations described above were completed. Points with an upward arrow attached correspond to complete fade-out of the GLH signal.

examine the results obtained in the light of present-day knowledge. V-i-f waves. therefore. They showed that the bulk of the absorption is of the non-deviative type. takes place mainly in the intermediate D-region hand. Since an enhancement of D-region ionization should increase the "wall" conductivity. the attenuation depending on the conductivity of the guide "walls". have not been published previously. and on the layer structure and responsive mechanisms in the upper atmosphere. this will reduce the attenuation of v-i-f waves.1 H-f Effects Appleton and Piggott (7i have made a comprehez sive study of h-f absorption at vertical incidence during a period extending over a sunspot cycle 0 They found that absorption was definitely under solar control. it is necessary to examine the absorption and reflection processes. absorption. and that it must take place in a layer below the reflecting 33 . to In particular. as well as the changes in ionospheric layer characteristics. Observations of the type presented above. on the other however. it will appear that a sharpening of the lower boundary of the D-region must result from the flare. In particular. undergo a waveguide type of propagation between the conducting earth and the conducting D-region. It is of interest. A plausible qualitative explanatio at an early date: for the h-f and v-i-f effects was advanced The h-f waves are reflected by the E.and/or F-layers. it appears that these results have important implications on the type of solar event which causes the SID. and ionospheric structure. In order to bring this out. which take place as a result of a solar flare. however. since it varied in a regular manner with solar zenith angle.I phenomena. but will give rise to increased absorption of h-f waves passing through the D-regiono It will be shown below that the above qualitative explanation must be modified and made more precise in order to fit the observations. 4.

under conditions where the quasilongitudinal approximation holds. so that it is the ordinary wave which then is measured. have substantially the same dependence on the solar zenith angle.I level of the E-region. and the + sign is for the ordinary wave. the absorption is the same for reflection from either layer (apart from the period when the frequency is in the neighborhood of fE when additional deviative absorption takes place). they showed that the absorbing region cannot be merely the lower portion of the E-region. (2) For a frequency whose reflection level shifts during the day from the F-layer to the E-layer.. For non-deviative absorption (ioe.sign for the extraordinary wave. The absorption of the ordinary wave is appreciably less than that of the extraordinary wave when w/wL is not too large. depending on whether 34 . It can be seen that the dependence of r on the collision frequency v tends to a proportionality to either v or l/v. or to a sporadic 3-layer. in a region of ionization density N and collision frequency v. but must be an independent ionized region. the . which they identify with the D-regiono The evidence which led Appleton and Piggott to the above conclusions was obtained from three types of behavior: (1) The diurnal variations of absorption for two different frequencies. Furthermore. in a region where the refractive index is substantially unity). Appleton [8] gave for the absorption coefficient r. (3) The variation of absorption with frequency can be explained only on the assumption that the same medium is responsible for absorption over the entire frequency range. where wL is the magnitude of the longitudinal component of the angular gyro frequency. one of which is reflected by the E-layer and the other by the F-layer.

the absorption in winter being distinctly higher than for the same zenith angle at other seasons. the integrated absorption at vertical incidence for a wave which penetrates the absorbing region and is reflected (with negligible deviative absorption) at a higher level then is given by an expression of the form t o 4ds a A (w + wo F) (2) where A is a constant and F(X) is a function of the solar zenith angle.1. Appleton and Piggott [7] found a winter anomaly. Appleton and Piggott thus placed an upper limit for v of 2o107 /sec in the absorbing D-region. attachment). n = 1. Furthermore. the theoretical relation bhows that the effective reflection coefficient p depends on X in a relation of the form 11109 P1 4c (C01Xr.4 to 1. (3) For a Chapman layer (constant scale height the recombination coefficient is where n depends on the ionosphere model. n = 1. In the former case. In particular. recombination.g.0 Nicolet [9] showed that a region of mounting temperature with height would have a lower value of n than one of constant temperatureo The experimental values cf n determined by Appletcn and Piggott range from about 0. 6. 35 . This is shown by Fig.5.I v2 is small or large compared with (w + wL) 2 ..30. Thus ft follows that V2<<(w + wL)2 throughout the absorbing region. Taylor F10] found values from 0o7 to 1. Information regarding the electron production and removal processes in the absorbing region can be derived from a atudy of the dependence of absorption on the solar zenith angle X. and recombination coefficient). which depends on the rate and process by which free electrons disappear (e. X. Appleton and Piggott showed that the frequency dependence of in the total absorption (as measured by an effective reflection coefficient) is very good agreement with (2). while if proportional to the ambient pressure.

the types of variation described hold substantially for an oblique path of constant length. numerical procedures have been introduced to handle more general situations. 4. but results are available only for a limited number of combinations of parameters. More recently. definitely show that the absorbing layer is not of the Chapman type (for . 36 . which is made complicated by the anisotropy of the medium. since variations in the properties of the important regions of the ionosphere take place in a distance comparable with a wavelength. The requisite theory is much more complicated. The above studies of ionospheric absorption have been concerned chiefly with vertically incident waves. this is not the case for v-l-f waves. Our present knowledge of D-region structure has been promoted by studies of the propagation characteristics of v-i-f waves.which n = 1. and suggest that the region has a positive temperature gradient. and that the height region wherein the additional absorption during SID occurs cannot be localized from their measurements. This necessitates full wave An analytical theory. theory has been worked out only for special variations of electron density and critical frequency with height. It should be pointed out that Appleton and Piggott's findings relate to normal h-f absorption. although not completely understandable on the basis of present theoretical knowledge.2 V-i-f Effects Although the main features of h-f absorption are fairly well understood.The experimental values. Since the path length through the absorbing region increases as the secant of the angle of incidence on the absorbing layer.5). These characteristics will be sumarised here in order to provide a background for the subsequent discussion of D-region mechanisms. and then only for the case of a vertical magnetic field or of vertical propagation.

Measurements at a frequency of 16 kc/s. apparent height of reflection is deduced from observation of the amplitude pattern versus distance produced by interference between the ground and sky waves. The phase The depends on the length of the transmission path and the height of reflection. amplitude. These measurements have been made at various distances extending out to about 1000 km. Observations of the change in phase of the sky wave are especially useful in testing solar control of the reflecting medium. Variations in reifection height with time can be deduced from measureFor this ntenna ments of the phase variation of the sky wave at a given receiving point.1 Short Distance Characteristics The measurements at short distances may be sunarized as follows: (a) Relection Hight Typical results of the phase lag of the sky wave relative to 37 . Helliwell [12]. corresponding to an angle of incidence on the ionosphere of about 650. and then the modifications observed during SID.Although some measurements of layer height have been made at very low frequencies with pulse techniques (Brown and Watts [1]. and polarization. principally by English workers [14-22]. show that a distinct change in the character of the sky wave takes place in the neighborhood of 400 kin. The printipal characteristics of the ionospherically propagated wave (the so-called "sky wave") are its phase. the most extensive and detailed studies have been carried out on c-w transmissions. and also by measuring the phase difference between ground and sky waves for different frequencies.2. 4. for example. purpose the sky wave is isolated from the ground wave by means of a special arrangement. Consequently it will be convenient to discuss the short and long distance measurements separately. the Pennsylvania State University group [13]).

.1 km. the slope A(t) of the Fig. the components p22 and P12 of the tensor reflection coefficient [see Part I. with greater variability at the higher frequencies. = (4) An average value of ho is where ho is the value corresponding to X 0.I I the ground wave are shoun in Fig. a mean value is 5.1 kim. 4. (b) Polarization For short distances of 100-300 km. 8 shows curves of The apparent heights at Values of A(t) run I I noon and night near Cambridge. Iheight vs. and Ch(X) is the Chapman function. however. which is a reasonable value for the scale height onsequently this result was used for some time to infer that the reflecting layer was of the Chapman type.8 t 0. is not explainable on the basis of a Chapman layer.5 + 0. iand at 70 kc/s a-ound 3 kim. po 11] are approximately equal. ho and A(t) at 16 kc/s through the course of the year If reflection took place from a Chapman layer. The diurnal variation of the component p125 called the "conversion coefficient". It should be noted that the descent from the night-time height starts at a time very close to ground sunrise at the midpath point. is shown in Fig. 7. The polarization remains the (c) 4pJ4 e In view of the approximately circular polarization of the sky wave. in accordance with the relation h a h. This variation of A(t). The height of reflection shows marked solar control during the day. t A. however. nearly all observations show that the sky wave on all frequencies from 16-150 kc/s is approximately circularly polarized with a left-handed sense of rotation same through an SID.t og [C. and its seasonal variation in I8 . 90 around 6 kim. on 43 kc/s. I which reduces to sac X for X less than about 85o 73 t 2 km. 10. On 30 kc/s. England are shown in Fig.h (01. log [Ch(x)] curve would be the scale height.

respectively. (3) phase variations at 16 kc/s and lower frequencies in connection with basic studies of navigation systems. and distances up to 950 kin. for different seasons. with an essentially constant level during the day. 13 and 14 show the diurnal variation of pI2 on 16 and 70 kc/s. 12 shows the frequency trend of P12 Figs.2 Lon& Distanoe Characteristics The characteristics inferred from measurements over longer distances will be .Fig. These principally cover distances of about 400-950 ka. in sumaer and winter. (a) Reflection Height The reflection heights determined fra pattern fit the ground interference in with a reflection height of 70 ± 2 km at midday. These have been derived from four sources. It is seen that a pre-sunrise drop and post-sunset rise (The in amplitude takes place. 4. small ripples in the winter daytime curve are considered as probably being due to a two-hop wave. (1) 16 kc/s observations at 540 kin. It is evident that the daily amplitude variation is distinctly different from the daily height variation at short distances.2. The diurnal-variation of reflection height is illustrated by Fig. 15. with no apparent This agrees within a few kilometers with the variation of height with frequency. but will also include some deductions made from observations over distances of several thousands of kilometers. and (4) observations of the v-l-f spectral characteristics of atmospherics. for a 39 .miarised in this section.) The drop in amplitude begins at a solar zenith angle of close to 980. for a frequency of 16 kc/s. measurements near vertical incidence. (2) a series of observations over the Decca navigation chain at frequencies from 70 to about 130 kc/a. 11. Fig.

This is shown in Figo 16. From measurements of v-i-f transmissions on available frequencies analyzed by IA . (b) Polarization Measurements of the polarization of the sky wave showed this to be linear at about 450 to the vertical. This is completely different from the diurnal variation at vertical incidence shown in Fig. the height variation is very much like the amplitude variation near vertical incidence shown in Fig. In fact.2 kc/s over a 4000 km path.. 7. respectively. smaller values of reflection coefficient are found at the higher frequencies. and 0°55 at night. and to the amplitude behavior at short distances. Again. This represents a change from the short distance measurements.27 at summer midday. (c) Amplitude The reflection coefficient at oblique incidence is be greater than at vertical incidence. frequency of 16 kc/s. the sunrise drop in height being substantially complete at midpath around sunrise. which gave the polarization as approximately circular. Heritage. Weekes and Stuart This shows an increasing reflection Also. 10o Similar types of variation were observed at higher frequencies. for which it was assumed that the nighttime height was 90 km. 15. coefficient with distance. round to For 16 kc/s) Bain. Pierce [23] reported a normal diurnal phase variation at 16 kc/s of 200P ± 300 over a 5200 im path. an (21] obtained the results shown in Fig° 17. increase of about 2:1 takes place between summer and winter. while Casselman. For higher frequencies. and Tibbals [24] measured a diurnal change of about 3500 + 300 at 12. The drop in amplitude around sunrise is shown in Fig. 18. This is similar to the behavior of the reflection height shown in Fig. but smaller values at increasing frequency. et al 119] found a value of 0. compared to vertical incidence values of 015 and 0050.

combined with observations of the spectrum of individual atmospherics. Fig. This change in phase appears to be a very sensitive way to detect flares Near vertical incidence. observed near vertical incidence. 4. In general. 20(a). accompanied a solar flare of importance 2+. the decrease in reflection height is substantially the same for frequencies in the range 16-135 kc/s. as shown in Fig. The above characteristics. jPierce i4 [23] showed an example of a phase advance at 16 kc/s over a 5200 1m path This SID. at 70 kc/a than at higher frequencies. a change both in phase and amplitude of the sky wave is The change in phase corresponds to a decrease in reflection associated with an SID.3 This shows a minimum around 15 ke/s. 20(b). The amplitude near vertical incidence suffers a decrease during an SID. 19. the change in amplitude being greater at higher frequencies. Chapman and Macario [26] deduced the attenuation vs. frequency curve shown in Fig. it appears that the amplitude Imcrease may be a maximum for frequencies around 30 kc/s.2. 21 for 16 ke/s. undergo a drastic change at oblique incidence associated with the longer ranges (>500 ka).I I Eckersley [25). and a maximum around 2 kc/so SIRggeg The effects of SID associated with solar flares have been observed both at the short and long distances used to obtain the results discussed above. during an SID. From observations of SEA. . as shown by Fig. height. relative change in amplitude is The roughly proportional to the decrease in reflection height. of importance 3. The amount of this increase is greater. 22 shows an example of the relative phase and amplitude changes observed at a distance of about 900 km during an SID. [27]. while the amplitude increases markedly. for example. This Is illustrated by Fig. The phase change associated with the reduction in height of reflection decreases with increasing frequency.

corresponding to a height change of about 1. accompanying a 2 flare. a 60 kc/s signal over the same path experienced a phase advance of only 700. 4.4 Eclipse Effects Observations of the phase of the sky wave on 16 kc/s at steep incidence were made during a partial solar eclipse by Bracewell [311 Although the 42 . and becomes less for higher frequencies. or No amplitude equivalent to a reduction in height of reflection of about 9 km. Pierce suggested that the primary physical phenomenon produced by the SID might be a steepening of the ionization gradient. The progressively greater at higher frequencies. To sumarize the SID effects observed on v-. the SID produces a reduction in reflection height and a change in amplitude of the sky wave. and roughly At 100 kc/s the decrease may on the other hand. raised. which may be by a factor of 5 or more. This increase. appears to be a maximum at frequencies around 30 kc/s. while the sky wave amplitude increases markedly. Near vertical incidence the reduction in reflection height appears to be substantially independent of frequency.I A phase advance of 1000 was observed.l-f wave propagation. however. be by a factor of about 100. This is half the normal diurnal change. On the other hand. while the amplitude change is amount of this decrease is a decrease. so that the frequency of minimum attenuaAlso. in reflection height is less for higher frequencies. -_equency et al [29. At oblique incidence. the decrease proportional to the decrease in reflection height. Gardner [28] and Obayashi. while the amplitude increased considerably.SID. corresponding to a decrease in height of the reflecting region. change was observed. with an accompanying reduction in the phase lag at reflection. the low-frequency cutoff of the ionospheric waveguide was tion was raised.2. during a 3.30] showed that an SID shifted the spectrum of atmospherics upwards.6 km.

I

I

greatest eclipsed area was only 0.3 of the solar disk, a definite phase anomaly was found, as shown in Fig. 23. shape of the obscured area curve. From this result, Bracewell deduced that the relaxation time of the reflecting region probably did not exceed 6 minutes. change about 35 degrees Furthermore, the magnitude of the phase The form can be seen to agree roughly with the

represented an increase in height of reflection of

about 1 km, while for a Chapman layer a change of only about 0.2 km would be expected. 4.3 D-Lay'er Production and Structure A proper interpretation of SID effects on ionospheric propagation ultimately requires a knowledge of the composition of the ionizing agents, and of the reactions which lead to the prevailing ionization densities. some of the pertinent available information will be suarized. 4.3.1 The Two-Layer Model In order to explain the diurnal phase and amplitude variations In this Section,

discussed in Sec. 4.2.1 and 4.2.2, Bracewell and Bain [32] proposed an ionospheric model containing a two-layer D-region. The height of the upper layer, which they

denoted by Da, was supposed to be under solar control in accordance with the formula

h = 72 + 5.5 log sec x km.

This is shown by the upper curve in Fig. 24,

(5)

Below this layer, a layer denoted by

D4 was postulated to exist, with height variations as shown in the lower part of Fig. 24. The upper layer was supposed to be the reflecting layer for 16 kc/s waves

at steep incidence, while the lower layer was considered to be responsible for absorption of the waves. At sufficiently glancing incidence, however, reflection

would take place at the lower layer. Bracewell and Bain based their two-layer model entirely on the observations of

43

I

16 kc/s propagation at short and medium distances. They gave no suggestions as to the mechanisms by which these two layers could be formed. 4°3.2 Bracewp1ls Exhaustion Region In order to explain the observed type of solar flare and eclipse effects on the D-region, Bracewell

L31]

postulated the existence of a so-called

"exhaustion region", in which the ionizable constituent exists in a small concentration. With respect to a two-layer D-region, this mechanism was supposed to take

place in the upper region, denoted by Da in Seco 4o3olo Bracewell showed that an exhaustion region would explain the amount of change in reflection height, observed during a partial solar eclipse, whereas a mucn smaller change would result from a Chapman region. He. also showed that an exhaustion

region would produce h-f absorption whose variation with cosX agreed in general with experimental observations. Bracewell also showed that the characteristics of an exhaustion region would explain satisfactorily the observed reductions in reflection height during solar flares. For example, a reduction of 15 km in height would require an increase in However, no attempt

intensity of the incident ionizing radiation by a factor of 15.

was made to deduce the accompanying effect on the amplitude of v-I-f waes° 403o3 Ionization Mechanisms The existence of several separate mechanisms for the formation of ionization in the D-region has been brought out in the last 'ecad.e o7 soo Brown

and Petrie L32]., pursuing a suggestion attributed 'o Ratcliffe, have evaluated the role of photodetachment of electrons from Oj icnso ment of an electron to a neutral oxygen moleculea This ion, formed by the attachstarts building up in corcentration The nighttime

arcund sunset, resulting in the disappearance cf the normual D-layero

level of ionization below the E-layar is maintained by cosmic rays, which vary in

l4

I

intensity with latitude. Visible light, extending down into the infrared, can

supply the energy required to break up the attachment, and thus liberate free electrons* Since visible light can reach the altitudes >50 km appreciably before

ground sunrise, electrons released by the photodetachment process build up D-layer ionization appreciably before sunrise. Brown and Petrie (33], and Moler [34]

showe, that this explained satisfactorily the pre-sunrise drop in amplitude discussed in Sec. 4.2.1. Aiken L35] verified the fact that a two-layer D-region

would be produced at sunrise, the lower layer being due to cosmic rays, and the upper layer to photoionization of nitric oxide by Lymanradiation0 Thus, in the

two-layer model of Bracewell and Bain discussed in Sec. 4.3.1, these mechanisms would account for the layers D* and Da, respectively. Nicolet and Aikin [36], in a discussion of the formation of the D-region, pointed out the following mechanisms of ionization which are possible at levels

below 85 kn:

(1) X-rays of X < 10 A;

(2)

(3) (4) (5)

**Lyman-a radiation (0 = 1215.7A);
**

Ultraviolet radiation, X > 1800 A; Cosmic rays; Photodetachment by visible radiation.

The normal E-layer, which is ascribed to the combined affect of soft X-rays in the range 30-100 A and ultraviolet radiation (Lyman-p) is penetrated by cosmic rays, ultraviolet radiation of X > 1800 A, Lyman-a and hard X-rays (X < 10 A), Of these, cosmic rays and hard X-rays are capable of ionizing all atmospheric

constituents.

In addition, Lyman-a, due to a narrow window in 02 absorption at the A minor constituent, NO (l1 part in 1010

Lyman-a line, can penetrate to low levels.

was proposed by Nicolet [37] as the ionizable constituent responding to Lyman-a to

45

account for the daytime D-layer. In view of the presently-accepted view that the upper part of the D-region, Da, is due to photoionization of NO by Lyman-a, it is tempting to suppose that NO

is the ionizable constituent responsible for the exhaustion region postulated by Bracewell. The concentration of NO has been estimated by Nicolet [381 as about

3 10-10 of the total concentration below about 85 km, or about 105 cm- at 75 km

[36].

In order to give ionization densities to fit

the v-i-f observations, however,

the NO concentration would have to be lower than this by about two orders of magnitude, or about 103 cm-3 at 75 km. Although Bracewell believed the exhaustion region would also explain solar flare effects, this must be rejected on the basis of later evidence. For example.,

Friedman and collaborators [39] observed no large increases in Lyman-c during flares, whereas Bracewell requires a factor of about 15. In a recent report, Chubb,

et al [40] stated that no increase in Lyman-a occurred during a 1+ flare, but X-rays in the range 1-10 A were observed. As mentioned earlier, the solar flare enhance-

ment of ionization has been shown to be explainable by the appearance of hard X-rays in the wavelength rr.nge 1-10 A, which ionize all atmospheric constituents, and can penetrate to low levels because of the low absorption coefficients in this spectral region. The resulting ionization would be even less sharply distributed in height

.-than a Chapman region. 4.4 Comarison With SID Results The two features of the experimental results shown in Fig. 5 which require explanation are the following: (1) (2) The lack of correlation between the magnitudes of SWF and SSE; The mechanism which produces the SE.

It will now be shown that the first is explainable on the basis of D-layer structure

46

and solar flare radiation. For example.2 Mechani=m Associated With §§E The observations reported in Seco 3 show that SSE on vlf is one of the phenomena accompanying SID produced by solar flares. at present. It will now be shown that this qualitative explanation cannot be substantiated on the basis of presently accepted ionization processes and present theoretical knowledge concerning v-l-f propagation For the ranges involved in the observations reported here.1ol Abence of Correlation Between HaVitudes of SWF and SSE The absence of any correlation between SWF and SSE in Fig. the normal-mode theory of propagation is more advantageous than the ray theory. this would result in the absence of any clear-cut statistical correlation between the v-l-f and h-f effects of flares. but that an adequate explanation of the second is not available on the basis of present knowledge. It was also stated that such enhancements can be understood in a qualitative way as due to reduced normal-mode attenuation as a result of increased conductivity of the ionosphere. The relative intensifications of these two regions There is no reason will depend on the spectral distribution of the X-radiation. 47 . while the v-l-f SSE would respond only to changes in the lower layer. acting as t he upper wall of a waveguide. 4o3o1. so that the relative increases can be expected to change from flL'e to flare. Consequently. D*. that all solar flares have the same spectral distribution. 5 Is understandable within the framework of the two-layer model discussed in Sec. since only one mode is effective. A number of analytical treatments of this theory have appeared [41-55]. if the flare produces a burst of hard X-rays without any enhancement in Lyman-a radiation. then both the regions of the Da and D layers will be intensified. 4. The increase in h-f absorption leading to SWF is the sum of the increases in the two regions. 4. to believe.

The ionospheric parameters then -nter the analysis in an effect}e conductivity wr given by where wN' v. which dductions ofI a general nature can be drawn. Also. the attenuation de of the boundary decreases. these tend to show that.L1t4s from which one must drabI general inferences are rather scanty.y Because of the inabil : to produce an t e analytical solution of sufficient generality. A yet.OP density and collision frequency with height. and fPr special directions of Ubhii earth'a magnetic field (usually taken to be vertiiLL). so that the rea. Fig. 40 . for a constant collision frequency. One of the idealizations which reduces greatly the comple dty of the calculations is that of a sharply bounded homogeneous ionospher. Nevertheless. collision. and longitudinal gyrt angular frequencies. respectively.'] 6 assumption that the quasi-longit-Ldinal approximation of Booker [ 1] may be used. 25 (from [53]) shows the attenuation of the first rode i n db/lO00 km s a function of frequency for various ionosphere heights for a value of Wr of 2-105. Cail ulaticcs using uder the further eaes as the height such a model have been made.but none treats the problem in a maufficiently general way to allow definitive conclusions to be drawn pertinent to the present observationao Aalytical solutiona have been obtained only for speci&l distributions of ionizati . by Spies and Wait [5. iL are the plasmas. amotg others.. only a rather small nwbe of examples has been worked out. the attenuation decresxes as the ionosphere boundsa7y becomes sharper. but a very large number of special casil out in order to produce a suffici ently extensive catalog froa. . efforts have tb a directed towards obtaining numerical solutions [56-60o This approach is not rastricted to special needs to be worked height distributions. otl er things remaining unchanged.

hancement will cause a given ionization density to appear at a lower level.2 db. the value of wr. this decrease in attenuation is based on a custant effective conductivity. is assumed to remain unchanged. Thus the qualitative Instead. attenuation decrease is the result of a decrease in the grazing angle of the first mode to the ionosphere. would result in a reduction of slightly more than 0. at most. the total reduction in attenuation would be about This the height rechiction oourred uniformly over the whole path. Wr.2 db/1000 km for a frequency of 30 kc/s. Hence one might argue that it is reasonable to suppose that a decrease in attenuation as a result of a decrease of 15 km in reflection height of the same order as that calculated for the sharply bounded ionosphere would occur. if For a 50 km path. an appreciable decrease in effective conductivity would result. semi-quantitative. say. calculated by Nicolet and Aiken [36]. then the effective conductivity would decrease by a factor of about 2. The above conclusion. a solar en- between a quiet sum and a strong flare. is. expectation of an enhanced ionospheric conductivity would not be realized. but with substantiAlly the same gradient. However. it must be emphasized. 1. the effective ionosphere conductivity. since it is based on the behavior of an idealized sharply bounded ionosphere having 49 . in virtue of the approximately exponential increase in critical frequency with such a height decrease. 26. Assiming a value of wL of about 50106 as a representative value for the transatlantic path in the measurements with which we are concerned. so that the collision frequency.I It can be seen from these curves that a reduction of height from 75 to 60 km. However. v. about 103 a3 show no appreciable change in shape at a density of Consequently.3. a 15 km height decrease would bring about an increase in v of a factor of 10. would be decreased 0 On the basis of Kane's (62] measumment of collision frequency. The electron density distributions in the D-region shown iA Fig.

5. the electron density distributions calculated by Nicolet and Aiken. If we grant. while the upper layer is due to photoionization of nitric oxide by Lyman-a radiationo Hard X-rays (in the range 1-10 A) emitted by a solar flare 15 . that an increased sharpness of the lower boundary of the D-layer is required to explain the SSE produced by the flare. show no appreciable change in shape at the electron densities required. CONCLUSIONS Simultaneous observations of short-wave fade-outs (SWF) of a 13. and hence. Again.5-Mc/s signal and sudden signal enhancements (SSE) of a 31ol5-kc/s signal over substantially the same transatlantic path of approximately 5400 km show no evident correlation between the magnitudes of the two effects of the SID.I I "average" properties given by the Nicolet and Aiken results. then it is necessary to adduce the mechanism which produces this effect. in addition to increasing the ionization densities at fall levels in the 1-region. it appears that the flare must increase the sharpness of the lower boundary. and that an adequate quantitative theory is needed before a firm conclusion can be reached. an increased gradient at these lower heights appears to be required. In order to obtain an increased conductivity at the lowered heights due to the onset of a flare. is understandable on the basis of a two-layer D-region8 This absence of correlation The lower layer is produced by cosmic rays. This would result in a decreased penetration of the waves reflected therefrom. it must be emphasised that this line of argument is only qualitative. which are shown in Fig. As mentioned above. could outweigh the effect of the increased collision frequency encountered at the lowered reflection height. 26. In other words. for the time being. for a sufficiently sharp boundary.

an adequate explanation of the mechanism of the v-i-f enhancement is not available on the basis of present knowledge.penetrate to the low levels of the D-region and ionize all constituents (principally 02 and NS). it appears that an increase in the sharpness of the lower boundary of the D-region is required during the onset of a solar flare. no correlation should result between the two effects. The mechanism by which I | 51 . The relative intensifications of the two D-regions will depend on the On the assumption that the spectral spectral distribution of the X-radiatioa. which should result in a decrease in the effective conductivity of the layer if the ionization gradient remains the same. distribution varies from flare to flare. Since the increase in h-f absorption is the sm of the increases in the two regions. Phase measuremets show that a definite decrease in height of the lower boundary of the D-region is caused by the flare. the relative increases also cm be expected to vary frm flare to flare. while the v-i-f enhancement is occasioned only by the changes at the lower level. Conse- quectly. this takes place needs to be determined. On the other hand. This reduced height causes reflection to take place at a level of higher collision frequency.

"Anomalies ionospheriques a debut brusque". Budden and J. pp. V. 1936. Proc. Atmos. (6] M. D. 203. I . A. Proc. V. Katzin and A. "Observations of very-low-frequency propagation during sudden ionospheric disturbances". 1954. 5. pp. 1954. Nature. V.. U-31. Atmos. Appleton. Phys. V. V. V. & Terr. "Regularities and irregularities in the ionosphere . Phys. Roya Soco. Vo 1.R. Nature. pp. V. Ratoliffe. G. 141-172. "An effect of catastrophic ionospheric disturbances on low-frequency radio waves". Nicolet. 1275-1278. & Terr. Dellinger. 20. [3) R. [7] E. 1938. pp.. Jour.E. 1937. und den St&rumgen den magnetischen Feldes der Erde. A. "Effects of the atwospheric scale height gradient on the variation of ionization and short wave absorption". p. M6gel. "Effect of catastrophic disturbances on low-frequency radio waves". "'Uber die Beziehungen swischen Rapfangs-Strungen bei Kurzwellen Telefunken Zeit 0 . 11. pp. H. I. 1060-1061. pp. 141-146. (2) J. 140. [9] M. C. Piggott. "Ionospheric absorption measurements during a sunspot cycle". "Sudden disturbances of the ionosphere". 25. 141. [4] K.I [1] H. [8) E. Haire. Jour. pp. Appleton and W. R.I". 162. Washington. 1937. 1947. 451-479. presented at the joint meeting of the International Scientific Radio Union and the Institute of Radio Engineers. Bureau & J. M. Comptes Rendus (Paris). 1937. V. Bureau. Braaten. 646. 1930. (5] R. Oct. V. 1253-1290.

Proc. 888-894. pp. pp. 1936.. "Ionosphere observations at 50 ko". Pt. "The ionospheric propagation of radio waves of frequency 16 kc/s over distances of about 200 kin". and C.. G. Budden. . "The ionospheric propagation of radio waves of frequency 16 kc/s over short distances". I.E°E. ICSU Mixed Comission on Ionosphere. 55. 188-214. A. 575-579. "Ionospheric virtual height measurements at 100 kc/sa.R. [12] R. Ratcliffe. Jour. Proco I. 1951.E. of Second Meeting. Straker. 122-133. and M.. V. Pt. 614-633. [15] K. Westcott. 217-229.Eo. Wilkes. 41. 1949. "Further investigations of very long waves reflected from the ionosphere". T. A. pp. "The ground interference pattern of very-low-frequency waves n .I (10] E. Bracewell. Pt.. Best. "Studies of the ionosphere using very long wave pulse techniques". W. Ratcliffe. J. I. Proc. Taylor. C. Bain. NBS. pp. Geoph. Waynick. V. V. Vn 171. [13] A. 37. Io. Wilkes. N. 97. Wattes. R. Proc. [16] K. J.. W. "Absorption of radio waves reflected at vertical incidence as a function of the sun's zenith angle".3 . 1950. V. [14] J. [18] R. pp. Royo Soco. [ii] J. Weekes. Straker. "Experimental investigation of very long waves reflected from the Ionosphere". 179-181. Proc. pp. M. Jour. Brown and J.. V. 156. W. 1939. Proco ioE. Bracewell. [19] W. IV. 102. Proc. pp. N. E. 1955. and M. 99. j 5. 161-165. V. 1952. A.Eo. pp. Helliwell. 1950o [17] T. 1948. Proc.. Vo 99 Pto IV. C. 100-107. Soo. A. H. N. V. A. V. pp. Res. ppo 250-259. 1952. Res. V. Roy. "The ionospheric propagation of radio waves of frequency 16 kc/s over distances of about 550 ka". H.EoE.

[29) V. N. Pt. W. C. 1952. & Geodec. pp* 38-46.101.Ze. IVe. L21] K.R.E. 1+05-4549 1932.. 47-55. "Propagation of audio-frequency radio waves to great distances". 28-45. T. Proe. 1949- [28) F.LL. L. Stuart. [22) K. Pt. 1954. [26) F.. Nature. pp. pp 41-48. 1956". 41? pp. Fujii. "~The ionospheric propagation of radio waves of frequency 30-65 kc/s over short distances".. 29-37. D. and T. Tibbals. "The ionospheric propagation of'radio waves with frequencies near 100 kc/s over distances up to 1000 kmn". Proc.. V. "Measured frequency spectra of very-low-frequency atmospherics". Casselnan. 1952o [23) Jo A. D. Geoph. Stuart. S. 930-933# 1956.(20] R. pp. Jour. "Mh phase shifts associated with the disturbance of February 23. and T. [27] R. MV.. Bracewell. 47.N. Res. 8298399 1959. Weekes and R. F. Jour. [25] T. Chapman and R. PP. Pierce. and M.. Phil* YNa. Weekes and R. Heritage. D. Geomag. pp. V.. NBS-D. Bracewell and T. N. Proco I6EE.L. Macaz'io. [30] T. Proc. W. "Propagation measurements for the Rsdwc-Omega navigation system". P. L. I.S. "Studies in radio transmission". V. 71. 109. 475-483t 1956. Radio Propagation. V. Eckerass. pp. "IAn experimental proof of the mode theory of VLF ionospheric propagaction". Jour. Obayashi. 177. Straker. V. 10. W. V.A. 1950.IL V. 1960c j 54 . pp. 1259-1269. 1959. Jour. Kidokorv. J. 61. 99.9 Pt. M. pp. IV. ILE. [24] C. Io. "The study of solar flares by means of very long radio waves". Straker. 99. pp. Res. "The ionospheric propagation of radio waves with frequencies near 100 1cc/s over short distances". Obayaabi. V. 154-162. Harwood. Ve 641). V. Gardner. J. V. "The use of atmospherics to study the propagation of very long radio waves".

Inst. 226-235. F. Lab. pp. Roy. Aiken. 32. [42] M. pp. Ionosphere Res. W. Votaw. 1947. V. Beg. Proco Roy. V. "Theory of formation of an ionospheric layer below E layer based on eclipse and solar flare effects at 16 kc/sec". 1960. Vo 189. June 1. and Me J. W. & Terr Phys. Ratcliffe. 1952c [32] R. No. [37] M. Symposium on Upper Atmosphere and Medicine. 1960. Baint. C. Soo. Penn. IGY Bull.i L31] Re N. Phys. Unsicker. Petrie. 216-225. 41. C. 143-163. pp. No. 175. [38) M. 1945.. Res. 1952. Aiken. A. "The formation of the D-region of the ionosphere". 65. 133.. Geoph. Brown and W. Kreplin. Geoph. Res.m. Roy. pp. Nov. 2. Nicolet. V. Jour. pp. 1958. Academic Press. A. "The oblique reflexion of very long wireless waves from the ionosphere". [41] M. Jour. Jour. Proc. Nicolet. E. Atmoso Physe. "A preliminary study of sunrise effects in the D-region". 1459-1468. [33] S. V. Ho Friedman. Soc. "An explanation of radio propagation at 16 ko/sec in term of two layers below E layer". B. C.. Jour. State Univ. oler. Soi. of low frequency waves". 199. Atmos. Bracewell and W. Jour. [39) J. "Aeronomic chemical reactions". 1 55 . A.Uo. Bracewell. Wilkes. M. 42. New York. pp. Can. "The effect of sunrise on the reflection height jTerr. V0 2. 19.. 1940. [35) A. V. pp. Met. V. 83-244. N. 1954. p. [34] W... No. [36] Mo Nicolet and A. AoG. Rep. ppo 130-147. 717-721. V. 1960. San Antonio. A.. 1469-1483. A. pp.. [40) T. 90-98. "Contribution & l'etude de la structure de 1'ionosph~re". Trans. "VLF propagation effects of a D-region layer produced by cosmic rays". Nichols. 65. Wilkes. "Solar Radiation Satellite. 1960. "Physics of the Upper Atmosphere*. R. V. 1960. "The theory of reflexion of very long wireless waves from the ionosphere". ed. A. Chubb.

surface of the earth". T. V. Jour. Roy. Phys. G. Wait. 0. P. "An extension to the mode theory of VLF ionospheric propagation". A. 1955. "Mode calculations for VLF propagation in the 56 . pp. A theoretical investigation". "The propagation of a radio-atmospheric". [14] K. Wait and K. "Terrestrial propagation of very-low-frequency radio waves. L. "The reflexion of vertically-incident long radio waves from the ionosphere when the earth's magnetic field is oblique".. V. Soo. 63. Jovw. 45. [48] Ia. Proc. A. P.. Trans. V. 65. V. R. [53] K. Phil. Soo. pp. 1947. Stanley. Wait. R. I. Heading and R. 1951. 231. 65-72. 125-135. Budden. Nag. 4. [50] J. 244. 1951. Wait.. "The reflection of very low frequency radio waves at the surface of a sharply bounded ionosphere with superimposed magnetic field". Res. Res. Geoph. [49] J. 833-850. Spies. P. 414-435. V. R. V. 1952.R. "Lw frequency electromagnetic wave propagation over the Moscow (1955). pp. pp. R.143] J. "Influence of earth curvature and the terrestrial magnetic field on VLF propagation". & Terr. Wait. Whipple. 42. Jouro Geoph. Phil.. VD 64D. Roy.. pp. 2325-2331. R. 1958. Nag. 760-767. "The mode theory of VLF ionospheric propagation for finite ground conductivity". 469-503. pp. "The oblique reflexion of long wireless waves from the ionosphere at places where the earth's magnetic field is regarded as vertical". 1960. pp. Atmos. 153-204. [52] J. Spies and J.. Radio Propagation. V. Res. [45] K. 42. NBS-D. 1950. 1960. Al'pert. 1-19. Jour. [46] J. Proc. pp. Budden.Eo. "The absorption of long and very-long waves in the ionosphere". Phil. V. [51] J. [47] J. pp. Heading..

1961.earth-ionosphere waveguide".. "On the propagation of VLF and ELF radio waves when the ionosphere is not sharply bounded". V. Wait. 1959. of Com. [58] D. "The numerical solution of differential equations governing refleion of long radio waves from the ionosphere. 387-401. NBS-D. IV". A. Barron. Radio Propagation. 57 . Proc. V. pp. 65D. Soc. Res. 1961..R. 37-46. 53-61. [57] K. Dept. Barron and K. Soo. Mag. A. Note No. V. 1959. G. pp. A. Jour. [55] J. 1962. Proc. 1955. G. "A new approach to the mode theory of VLF propagation". Phil. 516-537. V. 1959. pp. 1961. "The 'waveguide mode' theory of radio wave propagation when the ionosphere is not sharply bounded". [60] D. Soco.. pp. J. 114. Barron.. "The application of the magneto-ionic theory to the ionosphere". A. G. Roy. II". 249. Serv. 133-139.. "The numerical solution of differential equations governing reflexion of long radio waves from the ionosphere. Budden. Roy. U. Budden. V 227. 1935. "The numerical solution of differential equations governing reflexion of long radio waves from the ionosphere. [591 Do W. Roy. pp. W. pp. V. III". Roy. 393-408. [62] A. Radio Propagation. A. pp. Off.S. Phil. Joiro Geoph. G. Res. 1068-1081. Jour. V. [56] K. NBS-D. Wait. 150. Trans. V. Soco. [61] H. 64. Booker. of Tech.. pp. 45-72]. Proc. 260. Res. J. W. R. 4. V. Budden. Proc. 248. 1955. "Arctic measurements of electron collision frequencies in the D-Region of the Ionosphere".. "The numerical solution of differential equations governing reflexion of long radio waves from the ionosphere". [54] NBS Tech. pp. 267-286. Roy° Soco. 66D. Kane.

0 0. - l 0 00) CQ 0 ti 17r A00 3Ab r 0d0C) Iq Jd58 . ICd 0 0 .

*W'V 9.1 '4 HN 4 0 bD 010000 'loll d V0 59 .

.-ti. - - -~ ~~ J 7 COINCIDENCES . BETWEEN GLH FADES and GLC it ENHANCEMENTS . - ~~~~ - - - - -. -4--r 05 06 0708 0910 11 12 1514 15 1617 IS81920 2122 2324 01 020304 05 UNIVERSAL TIME Fig. . 60 .. ~-AUG. and off coincidences between GLH fades and GLO enhancements. . o1NOV. .. 3--Histogram of GLH fades. *r 1938-JUNEI194 I - *NUMBER -- OF GLH FADES 1936-JUNE 1940.

10Z L NACMNS CD z 060 8 la 9t 07080910If 12 131415 16 171IS1920 21 22 2324 0102 0304 05 UNIVERSAL TIME Fig. 61 . during common operating periods. 4-Histograms of GLO enhancements and GLH fades of intensity "moderate" or greater.

) 0 4-.H0 0 0 * -I 1VNOIS H-19 NI 3SV383G S13681330l 62 ..) U) V 0 .__ U U) C.

* C 1 C. 00 1. . 0 0 00 63 .

Seasonal variation of the apparent height of reflection' at 'night (upper curve) and at local noon~ (lower curve). 9-.4 .. x Obeazoaa min 1948 e 0 Obewvatom made to 1949.3[ J40-1 OL-r 0 0600 ita 24. 0.90 400_ _ _ __ _ _ A__ _ _ 0-- x *X -70 FES APR AM UIS OCT OC Fig.2-5Jly ovrincofcsi 98 * .1-ira PI aito o n16k/.0 Fi.

.00- 0.IU23 frequenWinfer different season's.4 a 0- ~0 1 0 0 ic1z 1 1Frequency. 12T11c variation of conversion coefficient 0~ L2With Fic.Ngt1.jRepreents 00 Sammer noon.4. an ur~r limit when measurementp am confused by aime 65. epuobeeevailam mae&R i~ 1% 00Ppeaaoeraomad .e. . 16 kc/s for midnight (upper curve) and midday (lower oie) ox Winter night. 01 Jan Feb Mar Apr May Jun Jul Aug Sep Oct No Fiji. 11-The seasonal varnuon ot 01 observed on a I4ii.

in summetr. 0000lp 010 0200n 0300' ob400ton (a) At 7114 kC/l 0uin 00.. in sunLwbr. 1949. h ie fgu . 1'At 70 -9 kc/s. -V --A At 127 -5kc/sin summper.x-x t At 71 ki. . At M58. 0-6 C . 66 .4 0 0) 6"~ 75 on e/ atugus. n :11car Sib L Fig.sin surmer.rise at the mid-point of the path. xx~xx The &now inicate All obirTd resultsat 7114 kcis.1kcs. S X 04 0-2 0 Ln 2O 40 60 0 10 0 000 0100 020 GAI 030 000 00 Distiicefromsendr~ki vrosdi~stance from !* . in winter.199 -At 83 kc/s. The arrow indicates the time of ground snrise at the mid-point of the pith. inaprn hih frelcin ear surise.- t ender. aqadover seven days in July. (e) At 1133Itch. (b) AtIs We.

67 .TAIL 0SILAT0YV with freuen~cy.x Fm(QqVcNCV %LOW .

0 volsjOK. 0 7 * *0 0 C) *I0 0 -9- +C * r un a 0 0 r -L - - -L 0 x -001 ~ 0 0 0400 . 68a0* .II C...

4. hi 69 . 5- I I I I I I I I l h I~cee.I II d .l Nex/num rnib S 07 N Fit. 7T1. 4The diurnal height variations of tho layers Da and DP. s .Eclipse anomaly obtained by subtracting mean of control days from eclipse day. 23-.

~t< ..4 7. i2 ' Py f* 105< Xi Wit :4.. T. / 70- . .... 7. * 1"0 FREQEN7.

I I I I t I S I I 0 REGION AND SOLAR ACTIVITY 0 12 3 4 6 75 - X 1. 2 6--Vmriatton or electron concentration vit. Specil events 6 Strong flares 6 0 1011 L0 4 loll ELECTRON CONCENTRATION (cm' 3 ) F14g. 71 .h height for various solar conditions. Very quI "ma 2 Quiet sun 65-/ 5 3 Lightly disfurlb 4 Disturbed sun 5.