The correlation of VLF propagation variations with atmospheric planetary-scale waves

8/6/2019 The correlation of VLF propagation variations with atmospheric planetary-scale waves

1/30

PRhCEDING PAGE BLANK NOT FILMED

THE CORRELATION OF VLF PROPAGATION VARIATIONSWITH ATMOSPHERIC PLANETARY-SCALE WAVES

D. J. CavalieriR. J. Deland

Polytechnic Institute of New YorkT. A. PotemraR. F. GavinJohns Hopkins University, Applied Physics Laboratory

ABSTRACTVariations in the received daytime phase of long distance, cesium-controlled,VLF transmissions are compared to the height variations of the 10-mb isobar icsurface during the f ir st three months of 1965 and 1969. The VLF phase valuesare also compared to height variations of constant electron densit ies in the E-'region from Brown and Williams (1971) and to variations of f-min fromDeland and Friedman (1972) which have been shown to be well correlated withplanetary-scale variations in the stratosphere by Deland and Cavalieri(1973). The VLF phase variations show good correla tion with these previousionospheric measurements and with the 10-mb surfaces. The VLF variationsappear to lag the stratospheric variations by about 4 days during the 1965 period,but lead the la tte r by about 4 days during the 1969 period.The planetary scale waves in the stratosphere are shown to be travelling on theaverage eastward in 1965 and westward in 1969. The above correlations areinterpreted a s due to the propagation of travelling planetary sca le waves withwestward tilted wave fronts. Upward energy transport due to the vertical struc-ture of those waves is also discussed.These correlations provide further evidence for the coupling between the lowerionosphere a t about 70 km altitude (the daytime VLF reflection height) and thestra tosphere, and they demonstrate the importance of planetary wave phenomenato VLF propagation.

INTRODUC TlONNumerous observations support the view that ionization variations in the D andE-regions a re coupled to metoerological variations in the stratosphere. Evidencefor this coupling is the connection of ionospheric variations, which have beendetermined almost exclusively from ground-based MF o r HF (>1MHz) radio

397IREX=EDlNG PAGE BLANK NOT FILMEDI


2/30


3/30

Eliassen and Palm (1960) have related the upward propagation of energy to thevertical structure of quasi-stationary planetary scale waves. Deland (1973) hasshown that the theoretical results of Eliaseen and Palm (1960) ar e also applica-ble to transient planetary scale waves. The resu lt s presented below a re consis-tent with upward propagation of energy into the lower ionoephere.In th is paper, varia tions in the received daytime phase of a long-distance VLFtransmission are compared with the height variations of constant densit ies fromBrown and Williams (1971). the variations of f-min from Deland and Friedman(1972), and the height variations of the 10-mb isobaric surface from Deland andCavalieri (1973), which were al l observed during the fir st three months of 1965.Daytime VLF phase values a re also compared to variations in the 10-mb iosbaricsurface for the f ir st th ree months of 1969.

VLF PHASE DATAThe paths of VLF transmiss ions monitored a t the Applied Physics Laboratory(APL) and nearby at the U. S. Naval Observatory (USNO) in Washington, D. C.a re illustrated in Figure 1with projections of magnetic L shells on the ionosphereat 100km altitude. The frequency, path length, highest geographic latitude,highest L shell (and corresponding invariant latitude) for each VLF path a re istedin Table 1. The phase advance produced by a uniform 1km lowering of the day-time VLF reference height is also listed in this table and was computed in thefollowing manner.VLF transmiss ions a r e often analyzed a s waves propagating in the waveguideformed by the earth 's surface and the lower ionosphere. The total phase delayT between transmitter and receiver separated by a distance do is 7 = do/vosecs., where vo is the VLF phase velocity for undisturbed conditions. Varia-tions in the ionosphere w e r this path which resul t in a different VLF phase velo-city v will be observed a s a change in the phase delay at the receiver, AT, and isexpressed by the formula, AT = do(l/v - l/v,) secs. A uniform lowering ofthe effective VLF waveguide hoight, due for example to ionization enhancements,will increase the VLF phase velocity and cause the phase to advance (i.e. anegative phase delay) at the receiver measured with respect to the undisturbedvalue. The time scale of these disturbances is lees than a few hours, so thatthe ir effects are not important to the present analysis of variations of a few days.The VLF reflection height for undisturbed daytime conditions is usuallytaken a s about 70 km (Po temra e t al., 1970; Johler, 1970, Westerlund andReder, 1973). Using phase velocity values for the lowest order VLF modesfrom Wait and Spies (1964) and Spies (private communication, 1964) which


4/30

WEST LONGITUDE (degrees)Figure 1. VLF Tra nsm ission Paths Monitored at APIA and th e U.S. NavalObservatory with Projections of Magnetic L Shells at 10 0 krn Altitude (fromWiley and Barish, 1970).

Table 1VLF Propagation Paths

Path

GBR-APL(Rugby,Eagland)NLK-APL(Jim Creek,Wash. )

L

GeographicLatitude

54.4O

48.2'

Freq.k ~ z

16.0

18.6

H i g h e ~ tL(A)

4 .0(6 0 O )

3 .5(57.5 O )

Distance,km

5615

3730

AT/& a t 70kmr sec/km

2.9

1 .8


5/30

employ the exponential conductivity model of the lower ionosphere , a uniform1km-lowering of the ionospheric reference height without a change of gradientwill produce a 2.9 ps ec (2.9 x sec ) advance in the phase of the GBR trans-miss ion a s received at APL. The phase advance (or re tardation) correspond-in g to a 1km-lowering (or rais ing ) of the effective height ne ar 70km f or theNLK-APL path was computed in the same manner and is also l is ted in Table1. These phase calculations may be applied to the VLF t ransmiss ionsreceived at the U S. Naval Obs ervato ry becau se th is stat ion is c lose( -30 km) o APL.During the first thr ee m onths of 1965 (the fi rs t period analyzed here) the fr a-quency of mo s t VLF tran sm itter s was controlled by cr yst a l oscil la tors whichdrifted in frequency and therefo re in phase to such an extent tha t meaningful studie sof long-period ( > l o days) varia tions a r e difficult if not impossible, The 16kHztransm ission from station GBR in Rugby, England, was unique during thisperiod since the frequency of this transmission was compared on a dailybasis to a ces>:r>; atomic standard located nearby at the National PhysicalLaboratory, T e d d i n p n , Middelsex, England. The avera ge frequency de-viation of the GBR transmissions over a 24-hour period was measured andrecorded (P ie rce e t al . , 1960; Red er, p rivate communication, 1973). Therece ived t ransmiss ion a t APL was a lso compared to a ces ium re fe renc e whichis part of the receiv ing facility. The NLK tran sm itt er osci llat or was put underdirect ces ium controi in May 1967. Before then the day-to-day var iati on s in theNLK data (k10 psec) were often very much la rg er than the GBR varia tions (i2-3P sec). Th is ma kes the us e of the NLK data extrem ely difficult fo r a long-periodanalysi s and may explain in par t the poor correla tio ns obtained using the NLKtransm iss ion data for 1965.With transm itter and receiv er os cil la tors controlled by cesium standards , theprecision of the frequency measurement is bet ter than a few pa rts in 101 , otha t the rel ativ e pha se delay a t APIA o r USNO can be determiried with a precis ionles s than a p sec in a 24-hour period. Thus, varia tions in the ionospheric wave-guide height that produce phase changes m or e than a few CI s e c s in a 24-hourperiod can be detected. Since a klkm uniform change over the GBR-APL pathwould produce a k2.9 p sec change in relativ e phase, we may expect to be ableto det ect height fluctuations of this magnitude.The GBR tran sm itter o scil la tor was placed under direc t control of a rubidiumstandard in 1967 which considerably reduced the lon gt er n frequency dri ft (al-though not as effic tively as by the ces ium standard). A parabolic phase varja-tion was subtracted to co rre ct f or th is dr ift in the GBR-USNO phase data duringthe 1969 period presented here.


6/30

All the VLF data presented here were subjected to a five-day running mean.The VLF phase variations may be considered as representative, approximately,of the variations of the reflection neight averaged over the transmission path.For comparison with electron density variations calculated from the ionosphericsoundings at Aberystwyth, and with meteorological data estimated for particularlongitudes (see next section) we can consider the VLF phase changes to corres-pond to horizontally averaged observations over the midpoint of the path, that isabuut 4 0 ' ~or the GBR path and 1 0 0 ' ~or the NLK path.

METEOROLOGICAL DATAThe geopotential heights a t various latitudes and levels had been subjected tolongitudinal Fourier analysis previously in connection with another study (Deland1973). Since the VLF paths ar e relatively long, and also because previous workhas indicated that the largest scale variations extend upward to a greater extentthan the smaller scales, we have calculated values of the geopotential height ata particular longitude by summing the contributions of the f ir st three zonal wave-numbers 1, 2 and 3. These longitudinally smoothed values of geopotential heightwere then subjected to a longitudinal and time-lagged auto-correlation analysis("autov becawe it is the same valliable at different places and times that isbeing correlated), the height fluctuations at four different longitudes being cor-related with the fluctuations at zero longitude. Lag correlation coefficients werecalculated for y ~ ( tT ) and gdt ) where y~ and yo are the geopotential height valuesat longitude X and zero, reepectively, and 7 is the delay in days at longitude Xrelative to longitude zero. The result s are presented ln Figure 2.The zero and 9 0 " ~ongitude graphs for 1965 (Fig. 2a) ar e almost opposite inphase, so the fluctuations appear to correspond to a wavelength of approximately180 degrees , that is, the three harmonics average out to essentially a "wave two''pattern. It is also apparent from all four graphs of 1965 that the best positivecorrelation is found for increasing lag as longitude X increase8 correspondingto the composite wave moving eaetward with an average speed of the order of 8degrees of longitude per day.In 1969, a comparison of the 45 O ~nd 9 0 " ~raphs (Fig. 2b) indicates that theaverage half wavelength of the comp,l; - tvave 1s of the order of 135 degrees,somewhat longer than in 1965. The composite wave for 1969 is apparently mw-ing westward (increasing lag westward) with a5 average speed of about 15 degreesof longitude per day.


7/30

46oE

r

0.8 - 909N goow

0.8 -1 I 1 I 1 L a 1 I I-15-10 -6 0 5 10 15 -15 -10 -5 0 5 10 16GBR-APL PHASE LAG (DAYS: GBR-USNO PHASECORRELATED WITH ZN CORRELATED WITH ZN

Figure 2, Auto-Correlations of the Variations of 10 mb GeopdentfalHeight Correepmdingto the Firet Three Harmonlce Relative to ZeroLongitude for the Firet Three Mo d e in (a) 1965 and (b) 19f8,

COMPARISON OF IONQBPHERICAND METEOROLOGICAL DATAThe running 54ay average of the daytime relative pham delay far the GBR-APL path is plotted in Figure 3 for the first tbree months in 1965. Therelative phaee ie measured in mita of r eec (10-"ec). The effective reflectionheight, computed by the method described earlier, relatlve to a 70km height,ie a160 indicated fn this figure. A180 ahown in Figure 3 are the variations of the


8/30

0.10MHz

0.05

Figure 3, The Height Zf: of the Constant Electron Deneity Surface for N =4.5 x lo4 elec/cm3 in the E-Region over Aberystwyth from Brown and Williams(1971), the Interpolated Valuee of f-min at Zero Longitude Correeponding toZonal Wave Number 1 (f I ) from Deland and Friedman (1972). and the Running5-day Average of the Daytime Relative Phase Delay of the GBR-APL Path for1965.


9/30

interpolated values of f-min a t zero l ongiMe corresponding to zonal wave num-ber l(f ,) determined by Deland and Friedman (1972) and the height ZE f theconstant electron density surface for N = 4.5 x l o 4 elec/cm3 in the E-region de-rived by Brown and Williams (1971) from ionograms measured a t noon atAberystwyth (located near the GBR transmitter a t Rugby). The threedifferent iono-sphere measurements shown in Figure 3 appear to be well corre lat ed when theVLF data a r e delayed by about 3 days with respect to the f ,and ZE ata.The sa me GBR-APL phase variations ar e plotted in Figure 4 with the geopoten-tial height corresponding to the sum of the fi rs t three Fo uri er harmonics com-puted for each day a t S OON geographic latitude and 90"W longitude. The Itdailyequivalent planetary amplitude1' Ap (Rostoker, 1972) is used her e a s the dailyover-all index of magnetic activity and is al so plotted in Figure 4. The times ofpolar cap absorption events (PCA1s)and geomagnetic storm sudden commence-ments (SC) are also indicated in this figure and their effect on VLF' signals willbe discussed later.The running 5-day ave rage daytime relative phase delay for the NLK-USNO trans-mission during the f ir st th re e months of 196C a r e plotted in Figure 5. The heightof the lOmb surface at zero longitude calculated from the fi rs t three zonal har-monics a t 50N geographic latitude is also plotted in thi s figure, but shifted to 4days later with respect to the VLF data. The Ap indices and times of PCA1sandSCVs r e plotted in Figure 5 on the sa me time s cal e as the VLF data.In Figure 6 the resu lts of a lagged cross-corre lation between the VLF phase dataand the lOmb geopotential height data a re presented for several longitudes fo rboth periods analyzed.

DISCUSSIONThe comparison of the GBR VLF phase fluctuations with the 90w component ofthe 10mb data for 1965 (Fig. 4) shows the two time series to be well corre latedwhen the latt er is lagged by about fourteen days. The correlati on is 0.67 whichi s sigr!iE.?ant a t the 0.025 level assuming 11degrees of freedom for a sample of56 d a b pi n t s using the VIStudent'sll statistic (but of course the choice of lag isalso relevant in estimation of significance). There is also a good corre lationbetween the VLF and the 10mb data ak 45OW longitude when the VLF is laggedby 9 days. The correlat ion is 0.65 a t the 0.01 level assuming 12 degrees of free-dom from a sample of 61. The best positive correlat ion a t zer o longitude is 0.36at the 0.2 level with a lag of 4 days.In Figure 6a the lag correlation coefficients mentioned above ar e marked by ar-rows at each longitude. The lag correlat ion curves for 45' E and 45"W ar e almost


10/30

SC l*scA Ooi y Average30

20 281 10 ISJon. Fob. March.

Figure 4. The 10mb Geopotential Height Corresponding to the Pun? of the FirstThree Fourier Harmonics at SOON Geographic Latitude and 90w Longitude, tineSame GBR-APL Phase Variation from Figure 3, and the Daily Equivalent Plane-tary AmpliMe A (Solar GeophysicalData, U. S. Department of Commerce).180' out of phase corresponding to a wave two pattern (that is the V LF data ar ecorrelated with en average two pattern) in agreement with the auto-correlationof the lOmb height data for 1965 (Fig. 2). It is apparent from an inspection ofthe correlation curves for al l four longitudes that the VLF correlated fluctuationsa t 10mb a re travelling eastward (increasing lag eastward) with an average speedof aanroximately 8 degrees of langitude per day which is again in agreement with


11/30

Figure 5. Running 5-Day Average of the Daytime Relative Phase Delay for theGBR and NLK Transmissions Received a t the U.S. Naval Observatory, theHeight of the 10mb Surface Calculated from the First Three Zonal Harmonicsat a 5 0 ' ~ eographic Latitude and oOI,ongitude, and Daily Equivalent PlanetaryAmplitude Ap for the F irst Three Months of 1969.the speed of the wave at 10 mb. Therefore, from a comparison of Figures 2a and6a, it may be tentatively concluded that there a r e fluctuations at 70km movingeastward with approximately the same speed a s those a t 30km (10 mb). Also,from F igure 6a, the correla tion between the VLF data (for 0-75"W) and the 10 mbheight data is greatest a t zero lag around zero longitude a 2 5 " ~ , ndicatingweatward tilt with height. Since the waves appear to be moving eastward, the3-day delay of the GBR-APL phase with respect to the f-min and z~ variationa(which ar e dependent upon ionization changes at higher altitudes than the day-time VLF reference height) indicate an eastward tilt above 70 km.In 1969 the phase fluctuations a r e leading the 10 mb geopotential height data(Fig. 5). The correlation of the VLF phase data when leading the ~Olongitude


12/30

Figure 6. Corlelations Between (a) he GBR-APL Phase Variations and the10mb Geapotential Height Data in 1965 and (b) the GBR-USNO Phase Variationsand 10mb Geapotential Height Data in 1969 for Several Longitudes.

45. E.4 ', I0

-.4---

- 8 - qc -8 r II 0z -w .4 -E o -LL

hr

W -0 -.4-0 -Zf -.a- I0

45' W

--.8-.8r 4 90.W-.4 -

-0 -

- -4 -b

-.8- 1 I I I I I I 1-15 -10 -5 0 5 10 15 5 0 - 6 0 5 10 1s

LAG (DAYS )


13/30

geapotential height data by four days is 0.52 which is significant at the 0.025level assuming 15 degrees of freedom for a sample of 76. The CBR fluctuationsa re almost in phase (time wise) with the geopotentkal fluctuations a t 45"E (Fig.6b). Inspection of Figure 6b shows that the variations appear to be dominatedby a wave one pattern (the correlation nearly reverses f rom 45OE to 90"W) andthat this pattern is moving westward (increasing lag westward) with an averagespeed of about 14 degrees of longitude per day, in agreement with the resul ts a t10 mb for the same period. Since the VLF phase fluctuations ar e almost in phasewith the geopotential fluctuations a t 45"E, the wave apparently ti lt s westwardwith height. Summarizing the above, longitudinal phase relationships both inspace and time presented in Figures 2 and 6 provide evidence of the existence ofa westward tilted eastward travelling planetary sca le wave during the 1965 periodand a westward tilted westward travelling wave in 1969.The longitudinal phase relationships discussed above for both periods analyzedwere a lso present before smoothing of the VLF data, so these phase relation-ships ar e apparently rea l and ar e not due to averaging techniques.Comparison of the phase fluctuations along the two paths analyzed (NLK and CBR)did not indicate any definite phase (longitudinal) relationship in 1965, which seemslikely to be due to the severe frequency drifts of the NLK transmit ter oscillatorduring that period.In 1969 correlation of the NLK and GBR fluctuations computed over the wholeperiod was weak. However, there appears (Fig. 5) to be a good correlation be-tween the two paths for January and February. The reason for the poor correla-tion in March is not obvious, particularly since the entire 1969 period i s markedby relatively high geomagnetic activity. A discussion of geomagnetic effects onVLF propagation is presented in the next section. The correlation between thephase flr!ctuations of these two indicate that the fluctuations a t least during Januaryand February were of large scale both in longitude and latitude.The fact that the phase fluctuations along the GBR path correla te better with the10 mb data than do the fluctuations along the NLK path may be due in part to thelower geographic latitudes of the NLK transmission path. Studies made withshipborne absorption experiments (Schaning, 1973) have suggested that there maybe a geographic latitude "cut ofP1 around 3 5 ' ~ 4 0 " ~atitude for such eventsas variations of D-region absorption that are apparently due to planetary waveeffects. Therefore, a southern boundary may exist ( - 40N) south of which plan-etary scale wave transmission may be inhibited. Dynamical models (Dickinsnn,1969; Matsuno, 1970) of upward transmission of planetary scale wave energy alsoindicate that such upward transmission should mainly occur in high latitudes.More significant correlations would then be expected at higher latitudes.


14/30

Deland and McNulty (1973) have derived an approximate relationship for the t imeand zonal average energy conversian &om the zonal flow to a traveling planetary-scale wave:

o a cosg 2 apwhere

A = the phase of the traveling wavep = pressureZ+ = the amplitudeg = the latitudeo = the hydrostatic stability factork = the zonal wave numberU = the eastward wind velocityup = au/bp

From this expression it follows that if the eastward zonal flow decreases withheight (i. e. 31, /ap > 0) the waves must tilt eastward with height for t,fa energyconversion to be positive. On the other hand if the zonal flow increases 1.ithheight then ETwill be positive only with a westward tilt wi th height.From observational studies Deland (1973) has found that in the %wer stratospherea t mid latitudes the traveling planetary scale waves apparently adjust their struc-ture relative to the zonal flow rso that energy is converted from the zonal flow tothe waves a t the levels studied.In view of the fact that in winter the eastward zonal flow increases with heightfrom 30km up to the stratapause and than decreases, it is not surprising to findthat the VLF fluctuations a r e lagging behind both the geopotential data a t 10mband f-min and E-region electron density isopleths fo r the 1965 period (see Fig.3). Although the height of the stratosphere is taken to be at about 55 km at midlatitudes, various studies have shown that winter mid latitude west wind maxi-ma vary greatly. Maxima heights as great as 70km have been reported (Batten,1961).The result s obtained from the above correlations seem to be consistent with thestudy made by Deland (1973). The tilt of the wave fronte of these transient plane-tary waves are such that energy on the average i s converted from the zonal flowto the wave up to the E-region and possibly higher. Such an energy supply forthe waves could compensate the losses due to radiational cooling, for example,a s analyzed theoretically by Dickinson (1969).


15/30

PCA AND GEOMAGNETIC STORM EFFEC'IS ON VLF PROPAGATIONThe phase of midlatitude VLF transmissions during daytime conditione is not of-ten affected by geophysical disturbances in comparison to high l a t i w e o r night-time VLF transmissions. The daytime phase is sometimes disturbed by 1-10Ax-rays during solar flares, but these effects usually las t for less than 1hourand cannot affect the slow planetary-scale variations analyzed in this paper.However, during PCA events, the sun can provide a sufficient number of energe-tic part icles to penetrate down to a 70km altitude and disturb VLF transmissionsfor long periods of time (e.g. 1 o 10 days). The excess ionization during PCAVsis confined to the polar caps of the earth (>63" geomagnetic latitude) exceptduring severe magnetic storms when these effects extend to lower latitudes(Zmuda and Potemra, 1972).The only PCA event that occurred during the first 3 months of 1965 began onFebruary 5, 1965 and was relatively minor (producing a peak 30MHz polar capriometer absorption of 1.8 db in comparison to 12db for more severe events).Riometer measurements at several latitudes during this event by Bailey andPomerantz (1965) indicate that ionization effects were negligible at o r below L = 4(the highest L shell reached by the GBR-APL p ~th) .The comprehensive review of VLF and LF propagation disturbances at mid-latitudes associated with geomagnetic storms by Belrose and Thomas (1968) in-dicates these disturbances are llmost marked during twilight and night hours,and a re usually absent a t noon. l1 The geomagnetic storm which accom-panied 'JLF disturbances presented by Belrose and Thomas were charac-terized by a range in the daily equivalent planetary amplitude Ap = 100 to 150.The largest A p value in the period 1 January to 20 March 1965, a s shown in Fig-ure 4, was equal to 31 following the storm sudden commencement (SC) duringthe PCA on February 5, 1965. Except for this minor PCA, the entire periodduring the beginning of 1965 can be characterized a s magnetically quiet. It ap-pears unlikely therefore, that the long-term variations in the GBR-APL phaseshown in Figures 3 and 4 can be attributed to PCA o r geomagnetic storm effects.During the period 1 January to 27 March 1969, three minor PCA events occurred,which began on January 24, February 25, and February 27, with the peak 30MHzriometer absorption equal to 1.7db. These are indicated in Figure 5 and theredo not appear to be any clear effects on the GBR or NLK-USNO phase variations.For example, the relative phase on both paths began to decrease on 20 January1969 preceding the January 24 PCA and the phase delay increased after this event,instead of decreasing as would be expected for polar VLF transmissions duringPCA events.


16/30

A large number of geomagnetic storm sudden commencements occurred duringthe bmnning of 1969 and these are shown in Figure 5 with the daily planetaryamplitudes Ap. The magnetic activity for the month of January 1969 i s relativelylow (maximum Ap = 29) when the GBR-USNO phase delay reached its most nega-tive value during the en tire period shown in Figure 5. Because of the numberof SC*s in February and March 1969, i t i s difficult to prove conclusively thatm e of the VLF variations shown in Figure 5 ar e associated with magnetic stormeffects. However, the level of geomagnetic activity during this period (maximumA p = 62 for February and maximum Ap = 79 for March) is only moderate in com-parison to the larger geomagnetic storms that have been observed to producedaytime mid-latitude VLF disturbances (Belrose and Thomas, 1968).

CONCLUSIONComparison of VLF phase measurements with stratospheric geopotential heightdata indicate the presence of traveling planetary scale waves in the ionosphere.The use of long distance VLF transmissions a s an ionospheric probe i s usuallylimited by the fact that localized disturbances (small in spatial extent comparedto the path length) a re difficult to detect. But, a s shown here, the VLF phasedata can be very effective for the study of planetary-scale disturbances. Fur-ther, the VLF is affected by a smalle r altitude range of ionization in the D-region,in comparison to MF or HF absorption measurements, and is therefore, a moredirect measure of smal l changes (about 51km) in the effective height of the iono-sphere near a 70km altitude.For these reasons and as demonstrated here , stable-frequency VLF transmis-sions can se rve as a useful tool for the study of stratospheric-ionosphere coup-ling. They should be especially useful in studying the vertical propagation ofenergy into the ionosphere in terms of the vertical structure of both the quasi-stationary and transient planetary scale waves.The results presented in this paper appear to be consistent with requirementsfor ~pvmrd 20pagation of energy.It i s hoped that further use of VLF transmissions in the study of transient plane-tary scale waves in the ionosphere will make possible the forecasting of meteoro-logical effects on the ionosphere in the near future.

ACKNOWLEDGEMENTSThis work has been supported by the National Science Foundation, under GrantNO. 25820.


17/30

The work of T. A. Potemra and R. F. Gavin was suppor ted by th e Naval Ord-nance Sys tems Command, Department of the Navy, under C ontract N00017-72-C-4401.In the preparation of thi s paper, helpful discussio ns were held with Dr. F. H.R d e r and B. W. Shaw.The authors are grateful to Dr. G. M. R. Winckler and Miss M. Raines for th eVLF data fr om the U.S. Naval Observatory, Washington, D.C. The Fo ur ie rharmonics of the geopotential dat a were computed by J. Shiau, using the CDC6600 of the Courant Computing Center, New York University.

REFERENCESBailey, D K and Pomerantz, A. M., 1965, J. Geophys. Res. -0, 5823.Batten, E. S., 1961, J. of Meteor. 18, 283.-Belrose, J. S. and Thomas, L., 1968, J. Atmos. Terr. Phys. 30, 1397.Bowhill, S. A., 1969, Annls. IQSY 5, 83.-Boville, B. W., 1966, Space Rese ar ch 1, 20.Bracewell, R. N., Wldden, K. G., Ratcl iffe, J. A., Straker, T. W and

Weekes, K., 1951, Proc. Instit. Elec. Eng. -8, 221.Brady, A. H. and Crombie, D. D., 1963, J. Geophys. Res. 68, 5437.Brown, G. M. and Wi ll iams, D. C., 1971, J. Atmos. Terr. Phys. 33, 1321.Charney, J. G. and Drazin, P. G., 1961, J. Geophys. Hes. 66, 83.Deland, R. J., 1970, Q. J. R. Met. Soc. -6, 756.Deland R. J. , 1973, Tellus-5, No. 4, (In press).Deland, R. J. and Cavalieri, D. J., 1973, J. Atmos. Terr. Phys. 35, 125.Deland, R. J. and Friedman, R. Me, 972, J. Atmos. Terr. Phys. 34, 295.Deland, R. J. and Johnson, K. W., 1968, Mon. Wea. Rev. 96, 12.


18/30

Dickinson, R. E., 1968a, Mon. Wea. Rev. E, 405.Uickinson, R. E., 1968b, J. of Atmos. Sci. -5, 984.Diekinson, R. E. , 1969, J. of Geophys. Res. -4, 929.Ellassen, A. and ?Im, E., 1960, Geofys. Publikajnoner-2, No. 3, 1.Gregory, J. R and Manson, A. H., 1969, J. Atmos. Terr. Phys. -1, 703.Hirota, I., 1971, J. Met. Soc. Japan 49, 439.Johler, R. J., 1970, Phase and Frequency Instabilities in Electromagnetic Wave

Propagation p. 154, Technivision Services, Eng.Lauter, E. A. and Taubenheim, J., 1971, Space Res. XI, 1005.Matsuno, T., 1970, J. Atmos. Sci. 27, 871.Noonkester, V. R., 1972, J. Geophys. Res. -7, 6592.Pierce, J. A., Winkler, G. M. R., and Corke, R. L., 1960, Nature-87, 914.Potemra, T. A., Zmuda, A. J., Shaw, B. W. and Haave, C. Re, 1970, Radio

Sci. 5, 1137.-Potemra, T. A. and Rosenberg, T. J., 1973, J. Geophys. Res. 78, 1572.-Reder, F. H. and Westerlund, S., 1970, Phase and Frequency Instabilities in

Electromagnetic Wave Propagation p. 103, Technivision Service, Eng.Rostdrer, G. , 1972, Rev. Geophys. Space Phys. -0, 935.Schdning, B., 1973, J. Atmos. Terr. Phys. 35, 1003.Schwentek, H., 1969, Space Res. IV, 405.Thomas, L., 1971, J. Atrnos. Terr. Phys. 33, 157.-Wait, J. R. and Spies, K. P., 1964, NBS Tech. Note 300.Westerlund, S., Reder, F. H and Abom, C., 1969, Planetary Space Sci. -7,

1329.


19/30

Westerlund, S. and Reder, F. H., 1973, J. Atmoe. Terr. Phye. (in press).Zmuda, A. J, and Potemra, T. A,, 1972, Rev. Geophya. &ace Pbya. -0, 981.Reference is Also Made to the Following Unpublished MaterialPeland, R. J. and McPSulty, R., 1973, Paper presented at the Fifty-faurth annual

meeting of the American Geophysical Union, April 16-20, 1973, Washington,D. C.

Wiley, R. E. and Barieh, F. D,, 1970, Plots of geomagnetic field geometry,Tech. Rept. AFWL-TR-69-144, Air Force Weapon6 Lab., Kirtland AFB,N. Mexico.


20/30

QUESTION AND ANSWER PERIODDR. REDER:Any queetions, please ?MR,MERRION (Defence Mapping Agency):I wae wondering, what other implications would this have, other than forweather forecasting? Also is it possiblt to go backwards, do you think, to takethe weather forecasts and predict something about your VLF transmisslone ?

Abeolutely. As a matter of fact, in response to your latter point there, that isexactly what we have been using.A t one point we weren't really wnvinced that the meteorological. disturbancesdid have an effect on the ionosphere, There are many, many reasons why itshould not, because there a re various temperature minima as one plots temper-ature versus altitude.And one would suspect that anything that happens on the ground would be inda tedfrom the Earth's ionosphere that is so high up. There is no reason to expect whyit should propagate upwards,But we believe, by looking at the meteorological data fi ret , and then correlatingit with the VLF data, that there is a connection.So, this is the direction that we have been goiag now.I only mention that as a very, very low possibil!ty of using the VLF to perhapspredict what is going to happen on the ground, because there Is a great deal ofinterest; ae you know, Walter ReRoberts is now advocatink * is theory that theinterplanetary magnetic field can be ueed to predict weather, and things like this.So, thie ie quite a controvereial issue. But I think at the present point we canclaim that by using the meteorological data, we can correla te it directly withionospherlc oecillations.Now, as far ae implication is concerned, I euppose if you are using the VLF asa time and frequency reference that you have to be concerned in the wintermonthe as to long period stabilities and disturbances caused by the ionosphere,becauee some of these can come out to 20 microseconds o r so.


21/30

But I am just not cle ar how thi s would work into a tim e and frequency network.DR. REDER:Coming back to you r question, maybe that will be the only useful application ofweather forecasting.

DR. REDER:Any othe r questions 3 Yes, please.DR. KLEPCZYNSKI:I think there might possibly be another application. I haven't done any or de r ofmagnitude est im ate s on the back of an envelope yet , but if we have ani deafrom VLF how the atmosphere is acting, as tronom ers might be very interestedin this because they might be able to get data on refraction.DR. POTEMRA:Oh, absolutely. If we look at som e of the data that was pres ent ed on VLBI,for example those oscilla tions in the ear ly morning hours se em to exhibitper iods, I think, of 15 to 20 minutes. But in any case , it was remin iscent, notof planetary waves with 15 to 20 day period s, but of acous tic gravity waves thathave been looked at in gr ea t detail. They have peri ods of 10 to 30 minutes, andthey ar e often found after s unri se, becauue when the atm osphe re gets a big blastof heat from the Sun i t starts shaking. Also it is seen a s the Sun set s , becausetha t i s when things cool down.So, yes, I think that would be an important input to people for very long base-line interferometry.DR. REDER:I have a question, D r. Pot em ra. Maybe I mi ssed it when 1was outside.Isn't i t so that th is kind of an effect will be mostly seen on paths which arefairl y high in lati tude?


22/30

DR, POTEMRA:Yea, I didn't mention that, bu t thie ie a latitude restricted phenomena, and wejust can't get anything from high latitude patha, because they ar e very disturbedby +heaumrae, ae you w i l l know from your work, but there appears to be alower latitude cutoff also, and thia cutoff occurs at about a 40 degree northgeographic latitude. For example, the pathe that we have down to Panama o rto Trinidad, that is around the Equator o r the Southern Hemisphere, thereappears to be no variations of this sort. Profeseor Deland hae an explanationfor thia, and it hae to do with the propagation of the atmoepheric disturbances.But they are definitely rest ricted at latitudes above 45 degrees north geographic.If one gets higher than that, for example, up to 60 o r 70 degrees geographic;then we would like to talk about geomagnetic latitude; and then we would have tobe more concerned about ehort period disturbances due to particle precipitationand things like thin.DR. REDER:Any more questions on this paper?MR. CHI:In your data which was presented, apparently you have correlated the nighttimephase record with magnetic disturbcee.DR. POTEMRA:Yes,MR. CHI:What other parameters have you identified, temperature o r --DR. POTEMRA:Now, you are referring to the nighttime disturbances correlated with th e mag-netic activity, ia that right?MR. CHI:Right*


23/30

DR. POTEMRA:I went very quickly over that, because it i s related to an analysis that we didabout a year ago. I think Dr. Reder has quite a bit of experience in this area aswell, but the situation is this, during the nighttime we very often see very smalldisturbances that a r e correlated with magnetic activity.The question i s why. Now, on one occasion, during a so-called magnetic sub-storm, measurements of precipitating electrons were made down a t the SouthPole, of all places, and also whistlers propagation, VLF emissions. These a relong, very long frequency waves that propagate back and forth on the magneticfield lines,They were also correlated with the onset of the same type of VLF disturbances,and on at least this one occasion we put together an argument that the VLFtransmission phase disturbances were due to precipitating electrons that werebeing dumped out of the Van Allen radiation belt, and that these w e r e due tothese whistlers propagating back and forth, and that another manifestation of thesubstorm - now, it wasn't a blg storm, a smal l storm - was the ground basemagnetogram deflection.Now, we have been trying to advocate the theory that when one sees these night-time VLF disturbances, they ar e due to precipitating Van Allen e!ectrons thatare associated with magnetic disturbances.Now, unfortunately, they occurred so often and it i s very difficult to get allthese things coordinated. But we think the evidence is very st rong that this isthe case.Now, that has nothing to do at all with meteorological disturbances. I justwanted to paint out that we have to sort out magnetospheric disturbances frommeteorological disturbances, and one has to be very careful.MR. CHI:Did you correla te with respect to temperature, for instance ?DR. POTEMRA:' the metoerological disturbances, yes, we have done that a s well becausepressure and temperature would certainly work together, and there i s an effectwhich has been long known, when one looks at absorption of HF radio waves,called the winter anomaly. During the winter months when the atmospherecools down, they have observed for many, many years, 20 years, that the


24/30

abearption, now, not VLF, the absorption d~sapp eared 8 well. Not completely,but reduces, except on certain days, when the temperature increases on theground fo r a few days - hese ar e called stratospheric warnin gs - nd theabsorption also increases. This was first detected, 1 believe, by Profes sorInuckes in 1950 or 80 in Berlin, and i t was called the "Berlin warming,"So, that i s another manifestation, but V L F hasn't yet been used for this. Butcertait ly temperature corre lations have been made, yes.DR. REDER:Any other questions on this paper?(No response, )DR. REDER:So, let's start now with our general discussion of all the papers which have beengiven up to now, and who wants to ask a question o r comn~ ent r anything?Yes.MR. MONTGOMERY (WSM, Nashville):Fo r seve ral ye ar s we have been running phase recordings on WWVB t Nash-ville, and we have noticed on a number of occasions that there will be a shiftabout noontime on certain days.I was just wondering i f anyone el se has noticed this ? This is a shift that lookslike the st ar t of a diurnal shift, but it i s only for a short period.DR. REDER:1s anybody here fmm the Bureau of Standards, o r from the power companies,somebody who uses WWVB(No response. )DR. REDER:May 1 ask you, Mr. Montgomery, when was th is ? When did you observe it ?


25/30

MR. MONTGOMERY:On a num ber of occas ione in the pas t , but I don't have th e d a b with me tpresent, but I can look thisq. W e ha v e t h e r e c o r d s f o r th e p a s t three o r fo uryears.DR. REDER:And i t happened a t noont ime, local noont im e?MR. MONTGOMERY:Right, loc al noontime.DR. REDER:Well, was it an SID, mayhe? How long did It last?MR. MONTGOMERY:A m a t t e r of a n h o u r o r so.DR. REDER:Well, it could be a n SID.A r e t h e r e any o t he r ques ti ons ? Yes.MR. MERRION (Defense Mapping Agency):I have a ques t ion fo r Mr. Am lie of FAA.I asked Dave Cal l this sa m e ques t ion , but I guess I didn't ge t the intent of myques tion acro ss .The A i r F o r ce , o r whoever I s r espons ib le fo r changing the nam e of tha t GPS somany t i mes , i s abou t to go ahead with the GPS- NAVSTAR syste m.My que stion is no t r e f e r r ed to the aynchrodabs , hut the astradabs. Wouldn't itbe r easonable to p lan on us i ng an a s t r odab s ys t em to be comp at ible with theGPS y s t e m , r a t h e r than a s a s ynchr onous o r b i t s ys t em ?


26/30

MR. AMLIE:Let's me, how do I handle that answer?Fir st of all, the astrodab really isn't real. You know, it is fashionable to havea satellite program, so we have one.

I am serious. I own an airplane, and my wife and I share a checkbook, aad 1simply couldn't afford the kind of avionics that i s required, to participate in asatellite system.The military have a need for global coverage, they have a need for securecriptogmpbic navigation and communications. Their needs are entirely differentfrom the civil community, and they a re willing to pay, a s all we taxpayers know.It i s entirely different. So, I think it i s not reasonable.DR. REDER:Yes.MR. MERRION:In reference to what you j u t said about general aviation, someone who i s inoperations thougbt that proximity warning devices ar e the ideal collision avoid-ance system, and wouldn't something in this area, say, in infrared sensor s,wouldn't this be the . .. .MR. AMLIE:Again, it is a matter of economics. The system that was used at Fort Ruckerby the Army was a short range system, because they had a severe problem withhelicopter training. A s you know, they have a couple of square miles and anenormous number of helicopters, and they had a problem, they had people killedin collisions.The equipment they bought was very short range, it was $5,000 a unit. If youare to use it for fixed wing aircraft , it has to be, you know, much fancier, andthe price m s p.Our goal in the DABS operation is to have the ent ire avionics units under athousand dollars.


27/30

DR. REDER:Any more questions on anything?DR. WINKLER:I would like to make a comment though and that goes back to the slide that youjust gave.I think there i s one more point in the considerations, and that is 99 percent ofall general aviation is not interested to find their location in the middle of thePacific. They want to have something to go here in the Continental UnitedStates.Now, for those few who a r e in the middle of the Pacif ic, o r who a r e bush pilotsin Northern Canada o r in Alaska, the re a r e additional grs tems which a r e eco-nomical and which a r e on the market W ay . 1 wanted to mention that, and alsothat they do not only use Omega, which at the moment i s kind of frust ratedbecause of a lack of operational transmitters. I imagine, of course, that thiswill eventually be done. There a r e stations on the ai r which are entir ely com-patible with Omega. These a r e the high power VLF statim, and I hope to heara little bit more about this th is afternoon.They a r e being used fo r navigation by a la rge number of ai rcraft , already goinginto many hundreds of users.What general aviation can do, other people can do a s well, and I think the appli-cation of precise frequency control of VLF transmissions, be that now Omegatransmissions, or be that the communications transmiss ions , i s somethingwhich is still an important item, and it is fo r that reason that I think thatresearch on prediction of propagation phenomena must continue.We have had this morning the correlat ion with atmo spl~eric henomena, and 1think that is j u t one of the things which we have not yet completely under con-trol, and I think until we a r e in a position to set up somewhere and to have apredicted, accurate time of propagation from a s tation for a epecific frequency -we ar e sti ll a li ttle bit away from that.Would you agree with that, Dr. Reder ?DR. REDER:Yes.


28/30

MR. WILSON:I am Robert Wilson, from Aerospace Corporation, fmm the division that i sinvolved with NAVSTA R.AS so rt of an instant lay expert, I have been studying a lot of data that has beenproduced by many people in the audience. I would like to know why I have seena gr eat deal of data plotted a s a variance of df/f, but essentiall y no data plotteda s a variance of 6 t/t, which is of considerable more interest to us, for example,than the frequency variation.DR. REDER:Well, if you give me your addres s, we could send you, for instance, a s far a sVLF i s concerned, data on phase as a function of time.I understand there are some data, fo r instance, on the change of the total elec-tron content which can be related to the change in the time delay of satell itesignals to ground stations. I am sur e they ar e available from many sources,perhaps Dr. Soischer o r Mr. Gorman who is here. He can take your addressand send you data on that.DR. WINKLER:You may not have zeroed in on exactly the sense of the question here,Frequency and P h a ~ e ,he two a r e related. The sigma of a time variation, ofcourse, is related to sigma of frequency variations. It is very simple to convertone plot into the other, once you agre e what you want to accept a s a good statis-tical measure of time deviations.One misunderstanding which I find most often in discuseions about probable timedeviations is the simple fact that the most likely position of your clock in anyfuture moment will be with no time deviation. There is an equal probability fort h e clock to be late o r slow (in relation to its extrapolated rate).One has to keep that firmly in mind, that the most probable clock closure, o rclock error when you resynchronize, is zero.What we a re talking about is the width of the distr ibut ion function of these clockerrors when we make many synchronizations, This width i s quite clearly relatedto the sigrna/tau plots for frequency variations,


29/30

I would use the following measure &it(?) = r .SAf(r) and apply an additional fac-tor 6 n ord er to be conservative. -fMR. WILSON:Let me make the point that while these ar e convertible, it is not always easy todo, particularly fo r people who aren' t experienced in the field of stati stics.Both for the Air Force and for engineers who a r e not experts in the field of fre-quency, and the field of time determination, it would be extremely convenient tohave curves and data that show the way in which the errors in clocks ove r longperiods of time will develop, and these simply donf seem to exist. A t least wehaven't been able to ru n them down.Now, I understand from what you have said that thi s is available, and I will beglad to talk to you. But I did want to make that point, it is perhaps mo re a mat-t e r of laziness o r inconvenience than the actual overall capability of being ableto convert.DR. WINKLER:One source which is widely distributed and available i s one of the older Hewlett-Packard catalogues. I don't know why HP , in the most recent catalogue hasomitted the right-hand scal e of the sigma tau plots. I think that they have beenpaying tribute to some perfectionist, because of the lack of standardization insigma tau, o r sigma subscript tau.But I think i t was a useful device, and maybe HP would like to respond to thatquestion, why did you omit in your catalogue the right-hand si de ?MR. BOURDET (Hewlett-Packard) :I think it was just a very simple economic move ra ther than anything else. Wethought we could simplify the graph. We had to make it small in the catalogueand it was getting very confusing with s o many lines.DR. WINKLER:Maybe, since this i s a generally in teresting question, I should rep~yo i t morefully.There is a considerable amount of information in the paper on characterizationof frequency stabili ty by Barns and co-authors , members of that committee,particularly see equation 39 on page 113of the IEEE IM20 paper (May 1971).


30/30

In addition, there have been publications where di rect measures in time havebeen cited, e.g. ou r paper in metrologia Oct. 1970 o r Cutler & Venot, NEREMRecord page 68, 1968. If you have a sigma tau frequency plot you addone to theelope, e.g. fo r a -1/2 slope in frequency you get a + 1/2 slope in time, etc.DR. REDER:Any mor e questions ?(No esponse. )DR. REDER:Well, let me as k you one question.Is t he re anybody here who has personal experience with the problem of precipi-tation s tat ics on antennas used in aircraft navigation ? Anyone?MR. AMLIE:Well, I can give a sort of tm answer. It is a problem, precipitation static hasbeen a problem fo r a long time in aircraft. There a r e some excellent littleplastic widgets with very sharp needles which seem to solve the problem, cer-tainly on HF, VHF i s not a problem. It is a little plast ic widget that worksfantastically well down to VLF.DR. REDER:Well, maybe you should also get in touch with your people a t Atlantic City, be-cause they seem to have a problem.MR. AMLIE:

Maybe they don't have some of these gadgets. I have one on my desk I can givethem.DR. REDER:Any more questions, comments ?(No response. )

Documents

The correlation of VLF propagation variations with atmospheric planetary-scale waves