Occurrence of Sporadic E in the Nairobiregion

I.",Joana) ef Radio " Space PllJsies Vol. 3, SeptemHr 1974, pp, 227-231

Occurrence of Sporadic E in the NairobiRegion

H. KASBNALLY &: R. F. KELLEHER Physics Department, University of Nairobi, Nairobi, Kenya - ,

Rece;lIed 23 February 1974; accepted 8 April 1974

Sporadic E ocearrenee near Nairobi bas been stadfed by means of a 27'S MHz swept-azlmutla obIIlpae backscatter sounder from September 1969 to Mayl971. The results are r.ompared with ionosonde data for lEI •• fbEl ~ 5 MHz obtained over tile period 1965-1970. ~ comparison Is also made witb predictions pubUsbed by the lustituteoC Terrestrial Magnetism, Ionosphere and Radiowave Propagation, Moscow ..

,.. Introduction records were exam.ined for the presence' of Es e¢hoea. FIXED-FREQUENCY,swept-azimuthoblique 'back• in each of the four compass quadrants andthe ra.pge, scatter sounders have been used by various workers to mean azimuth and azimuthal extent of all 'sqch study the characteristics of sporadic E (see, for exam• echoes noted. The data were then further, pro<;osse

receiver was 0'5 p.V for a (signal -t noise) - to-noise of/Es (or IbEs) ;> 5 MHz for 3 seasons are ".bownc . ratio of 2, The common antenna was as-element in Fig. 1. horizontally polarized Yagi with a half-power beam• width of about ± 20°. The Yagi was ,mounted at the 3. Occurrence top of a 70 ft tower and was rotated in azimuth at Fig. 1 shows the diurnal variation of' percetlta.e the rate of one revolution every four, and a half occurrence for backsCatteC(all azimuthS) andlBsftl~ minutes. The returned backscatter echoes were record• the three seasons of the year: The resultS inboth~ e<\photographically, initially from an A-type display refer to the period Septembed 969 to May 1971. It with verti~l position on the ~RT varying with the is difficult to predict how the absolute: :values fOIth! azimuth of the antenna, and subsequently from a two methods of observation should compare" ~i~ conventional PPI unit, The sounding schedule from this will depend on. such factors as the ante~Da .~.~ September 1969 to May 1970 was quarter-hourly diagrams, the exact distribution of criticalf~uetI,cies; intervals pn four days per month. From June 1970 and patch sizes and so on. Also, it is ,clear that'tite this was changed to half-hourly intervals on twelve Hs patches observed by the ba~kscatter sounder,~., days a month. ' those at which the radio wave is rcflectcdrather'" The difficulties in interpreting backscatter records scattered. This implies tbat t~e blanketing freqJCl~ and in identifying sporadic B returns have already IbEs rather than the top frequency' re:ft~ed,(!f:s)'f~ been discussed by several workersltll• The individual greater than 5 MHz. During the day at Nairohi.tho INDIAN J. RADIO SPACE PBYS., VOL. 3, SEPTEMBER 1974 (To some extent this may be due to a peculiarity of the results forthe years when the backscatter was 25 studied. It can be seen from a comparison of Figs. 1 and 2 that the ionosonde Es for 1969-1971 tended to os increase significantly from 1200 to 1800 hrs, whereas the average for all the years showed a much flatter curve). The backscatter graph is broader than that for fbEs but this can be explained by the fact that it is the average data for clouds observed at all azimuths and therefore, in particular, over an east-west distance 25EQ corresponding to a range in local time of about 2 hr. Q)~0::J c Q)I Or-50(.)(.) ~ lt is obvious from Fig. I that the occurrence of ~ Es is greatest in the June solstice months, i.e. in what is formally local winter at Nairobi. The details of the \ month-to-month variation of percentage occurrence \ \ averaged over the day are shown in Fig. 3. The June! \ " July maximum is clear. It has been shown by Kelleher and Kasenally3 also by Oyinloye4.,that near the geo• graphic equator sporadic E may have a maximum. occurrence in local winter rather than in local sum• mer as is the case at middle latitudes. In order to illustrate this point further, Fig. 4 shows the seasonal

-.- fEs f ,,:--, ' \ '" en - -- fbEs ,i \ .•.••'-. I\ W ;t: - BKSTR /,, \ .. ,'\ •••.J U s:: " 20 Z :J ~W ~>• Q:: ~ JS ::> ±: oUu .cdI / /, ....._

06 12 hrs LT 6 12 18 Fig. 2- Diurnal variation of backscatter, IEs and /bEs ---I hrs LT , Fig. 1- Percentage occurrence of backscatter echoes (conti• I nuous line) and oflEs ~ 5 MHz (dashed line) for the three seasons as a function of local time (September 1969- May 1971) oQ) c: Es value is of the high or cusp type for whichfEs and Q)~ fbEs are closely related. The sporadic E which occurs ~ at night (flat type) is only weakly blanketing with o:J fbEs rarely exceeding 5 MHz. It does not, therefore,' o support oblique reflection at 27'8 MHz. This means o ~ that we'would expect the diurnal occurrence curves o for backscatter to be similar to those for fbEs and to agree in shape with the f Es curves only for the day• time period. Fig. 2 contains a comparison of the diurnal occur• rences of the three quantities (plotted with vertical scales adjusted for approximate equality). The iono• MJ JASONDJ FMA sonde curves are for all the available years taken Months together. It can be seen tbat the backscatter shows Fig.3-Percentage occurrence of backscatter (continuous better agreement with fbEs althougb tbe latter is line) and/Es ~ 5 MHz (dashed Jine) averaged over the day considerably higher during the period 0900-1400 hrs. for all months from May 1970to April 1971 228

.,.'·-'-L 1--1 - _··;.;\.lliRflH I "1'_1 "'j Iii. KASENALLY "KELLEHER: SPORADICE OCCURRSNCB NEAR NAJR.OBI

30 • " ,I. I TAM.23N I 20 I

fEs I~ ... 10

c~n-20A It ~M !I\ o QI :J ' : (J ,', '1 ,11 •• NAt. 1S o 20 f I'. I II oo I; II I 1 I N t: ~IU . ::c 10 0' \r~ ::!\ [J V ~ ~ t _ ~ : :: I• .i .~'I 'I[:t Ln 0 IA\ 20 ~. en \ ..' W 10 '65 '66 '67 '68 '69 '70 •••••o Year ' Pig. 5-Percentage occurrenceoff& (continuOQ&lU1e) .•J1I.4 o jbEs (dashedline)~ 5 MHz averagedovereacII monthf'rodl 20 January 1965toPebruary1971 It is well known8 that the incidence of as •.ca." .vary considerably from year to year, Fig. 5shows~' percentage occurrence of fEs and flEs :> 5 MlIZ.fQt all' available months at Nairobi and illustrates'it_jj,; o fluctuation well. It is clear that the curves forthe~, parameters agree very c]osely, with theocCll~: .. 10 for fb being about one quarter that of/Es .. ThGtc'd: no obvious so]ar-cyc]e dependence but the 'data '. o J ~ MAM J J A SON 0 is pI:obab]y too short to show any varia.tion.wh~Ji Months might exist.. " .. ',' •••.....,.:~ P'I'. 4-Mthon - b y-mon th variation .. 0ffOEs r'.••.5 MHz at a A study o.f hac.lcscatter' .... and lonosonde' da...t~c .' .dj; .,., numberof Africanstations not reveal any .defimte relationshIp between oce~-'., variation infOEs:> 5 MHz at five stations in the rence and magnetic activity." African zone. Tamanrasset (22·goN) has a maximum '4. Range and Critical Frequency Distribatlmt ,.' 'J" in the June solstice, as is normal for a northern hemi- / A histogram of the relati~e frequency of OGc!.lr-t sphere station. Both Nairobi and Lwiro (2'3°8) which rence of the minimum range of backscatter«i,.~ ", are formally in the southern hemisphere but close. to echoes is given in Fig. 6. This is fdt all .~ta,~ the geographic equator have definite June solstice- together. When' division' into seasons ~smad~.,t_ ,,:.·,,:,;i· +; maxima. The seasonal anomaly extends south to median range is smallest in the June so]stic~ •m~~~ Elizabethville (II OS)where the local winter maximum indicating that in this.season of maximum~ during the June month is.still conspicuous; the minima there is a greater preponderance of cloudswit~ .l1ii~_ occur at the equinoxes and there are secondary maxima electron densities. during the December solstice. The seasonal variation The range distribution of obsetvedclouds,will~' for Tsumeb, which is in the region of the change-over pend both on the distribution. of critical rteqUe~', latitude (19°S) has almost equal maxima during the of the patches and on the polar diasrI!-JIl<>!t~~'.;' June and December months and minima at the oblique sounder antenna1• ThelEs vadationis~~.·· equinoxes. easily established from ionosOnde data.·; It .is\li!n0~. It is also noticeable that the annual total of Es that the values of sporadic E frequencies are, subject . . . ' .. " . " - - ,:- -; . - j,-, '>:, occurrence decreases steadily from north to south, a to tlie ~Phillips frequency dependence ~e'~: .•...•...• trend whicb is clear in the world maps given by IOg10P = a + bj,!Ili' SmithS. where INDIAN J. RADIO SPACE PHYS., VOL. 3, SEPTEMBER 1974

E• 80

UI :e ',0 1-0 c 40 ::l >• •... c~L a E •... •... +• J ..0 U •... u C Of-eo I o,.!! WU 0'5 0-5-1 Z w 40 a:: a:: ::l u o hrs LT Fig.7-Diumal variation in four directions about Nairobi 600 1000 1400 of observed occurrence of backscatter echoes (heavy con• Height) k m tinuous line) and predicted occurrence 'based onfEs (con• tinuous light line) andfbEs (dashed line) during equinoxes Fig. 6- Relative frequeney of occurrence of backscater range

80 P is the probability of occurrence > I ; I, the frequency in MHz; and a, b are constants. An analysis of 20 months of data at Nairobi showed .. Ju0 40 ~~ 80 ~ ~0c•... that this law was obeyed very well for IEs '> 5 MHz cu ~40 and that the constant b was equal to - 0'35 (MHzt1• 40 A similar period of data wasT analyzed for fbEs values '> 4'5 MHz. Once again, the equation was found to describe the data quite well; the constant b in this case was equal to -. 0'45 (MHzt1. It has been shown by SmithS and Oyinloye8 that the best value of b varies from year to year and even from hour to hour during the day.

S. Cloud Sizes, Velocities and Durations hrs LT Fig. 8- Diurnal variation in four directions about Nairobi The cloud sizes were estimated by measuring the of observed occurrence of backscatter echoes (heavy con• azimuthal extent of the echoes and subtracting the tinuous line) and predicted occurrence. based on fEs (con• known beamwidth of the antenna. This method is tinuous light line) andfbEs (dashed line) during June solstice necessarily somewhat uncertain but it gives the most - 80\ 612181218I\I I 4080 W probable cloud size as about 200 km. c:cu•..0 0 40 ~cu S 0 80' Velocities qf the Es clouds were derived by follow• I ing the movement of the centroid or edges of observed 40 40 patches. Only those which gave consistent results (\ over several successive records were used. The values ~r obtained are in agreement with the speeds and direc• tions expected, i.e. about 100 msec-1 towards the west during the day. The durations of the observed clouds varied from 30 min to 8 hr and the median value was 2'7 hr. lhere was little variation of this quantity with season.

6. Predicti~n of Sporadic E It is interesting to compare the observed back• hrs scatter occurrence 'with values derived from world• Fig. 9-Diurnal variation in four pirections-about Nairobi of wide communications predictions of sporadic E observed occurrence of backscatter echoes (heavy continuous line) and predicted occurrence based on fEs (continuous frequencies. For this purpose we have used probability light line) andfbEs (dashed line) during December solstice 230 KASENALLY & KELLEtlBR : SPORADlC-E OCCURRENCE NEAR NcAtRO'ltI , . N It is interesting to note that, on the clouds were observed to thewest 50 to the north and east (Fig. 10). To SOme fall-off to the north may be due to the fact direction the region of suppressed blanketin,&~sporadic E is being approached' even though there is a caCDetal 7 • tendency for the occurrenge to increase southwardS in this part of Africa. The west-east asymmetry s~ans E to be,real and may even be greater than suggested by 50 the present results if, as has sometimes been Sllgg~$t• ,ed10, the sea,(to the east) is a more effiqetWscattering medium.

Acknowledgement , ' , ,," ,,' c'" This work was carried out under ,a contraCtwitli the British Ministry of Technology (AviatiQP$upply) s sponsored by the Government' CommuQ,ieatio~s, Headquarters, Cheltenham, from whom thel?acJc-• Fil!. lO-Azimuthal variation of occurrence .or' backscatter scatter equipment was taken on loan.,' The authon echoes wish to express their gratitude to Mr R.. MaU~~ and Dr P. J. D. Gething for valuable sug3~tfolts values obtained by applying the secant law to average during the writing of this paper audto, Dr G~thipg • v~lues off Es and fbEs as predicted by the Institut e for arranging the contract and giving useful advi~" of Terrestrial MagnetisItt, Ionosphere and Radiowave throughout. Propagation, Moscow in 1964. The data for ,the four directions about Nairobi are shown in Figs. 7-9. As References expected from the remarks in ,Section 3 above, the 1. BAGOALEY,W. J., J. geophys. Res., 74 (1969), 1169. predictions based' 'on fhEs show very much better 2. EOAN,.R. D. & PETERSON, A. M., Backscatter obiel'f. agreement with the observed backscgtter results, in lions of sporadic E, in.ionosp1}eric sporadic E, cdhe