JOURNAL OF RESEARCH of the National Bureau of Standards-D. Radio Propagation Vol. 64D, No. 3, May-June 1960

An Analysis of Propagation Measurements Made at 418 Megacycles Per Second Well Beyond the Radio Hori­zon (a Digest)

H. B. Janes, J. C . Stroud, and M. T. Decker (D ecember 1, 1959)

During an IS-mon th period in 1952 and 1953, t ransmission loss measurements at 41S megacycles were m ade over a 134-mile p ath between Cedar R apids, I owa, and Qilin cy, Illi nois. Continu ous r ecordings m ade simultaneously at severa l receiving a ntenna heights from 30 to 665 feet yield ed information on diurn al and seasonal variations in both the hourly m edian bas ic t ransmiss ion loss a nd in height gain . These data are compared to p redictions made using t he m ethod d evelop ed by Rice, Longley, a nd Norto n a nd a re found to be in good agreement, particularly at t he lower a ntenn a heights. An a nalysis of the correlation of short-term signal level variations observed at horizontall y and ver tically spaced antennas is described .

1. Introduction

This is a digest of an NBS T echnical Note [1] J

which dcscribes in considerable detail a series of transmission loss measurements made by the National Bureau of Standards at a frequency of 418 M c over a 134-mile path extending from Cedar Rapids, Iowa, to Quincy, Ill.

The principal purpose of the measurements was to study: (1) The hourly, diurnal, and seasonal variations in basic transmission loss experienced in transmission made well beyo nd the radio horizon ; (2) the corresponding long-term variability of height­gain; (3) t he comparison of measured transmission loss and height-gain wi th predicted values; and, (4) the correlation of instantaneous signal levels measured at vertically and horizon tally spaced antennas. (A discussion of item 4 is given in ref [1] but will not be summarized here.)

The experiment covered a period of approximately a year and a half from J anuary 1952 to May 1953. The transmitter was located at Cedar Rapids and was operated by the Collins R adio Co. under contract

) with the National Bureau of Standards. The receiving and recording equipment were installed and operated by NBS. Space on a 750-ft tower was obtained through the cooperation of WTAD- FM in Quincy, and the receiving antennas wer e mounted on this tower at heights ranging from 30 to 655 ft above ground. The transmitting antenna was 39 ft above ground.

The equipment was operated for 13 recording periods, each of approximately 2 to 3 weeks duration.

I During each period, continuous recordings of basic transmission loss were made simultaneously at 3 to 5 different antenna heights. Table 2 of reference [1] shows a schedule of the recording periods and the antenna heights used during each period. Also shown are the inclusive dates during which each antenna heigh t was used and the total n umber of hours of data recorded at each height during a given period.

1 Figures in brackets indi cate tbe literature reference at the end of this paper.

The angular path distance [2] ranged from 20 .3 to 15.7 milliradia ns for the 30- and 665-f t receiving antenna heights, r espect ively. Insofar as the long­Lerm median basic tr ansmission loss measured over paths having angular distances of this order (i .e., greater than abo ut 10 milliradians) agree quite well with values predicted from scatter theory [2], this migh t be consider ed to be a tropospheric scatter propagation path. However, analysis of the hort­term variations in sig nal level reveals tha t for sig­nifican t percentages of th e time (especially during the night), mechanism s other than sca ttering appear to be importan t. It wou ld eem tha t this path is in a tran sitional region between the shorter paths wher e processes such as diffraction and ducting may provide most of the signal power and longer paths wher e scattering is th e principal contributor.

2 . Description of the Data

The data output of this experiment was r educed to hourly distributions of instantaneous signal levels obtained from time totalizing recorders or, in some instances, from paper chart recordings. The hourly median basic transmission loss , L bm , and the fading range (ratio , in decibels, of levels exceeded 10% and 90% of the hour) wer e read from t hese distributions. A complete tabulation of all hourly L brn values and fading ranges is given in reference [1]. Also given there are graphs of individual hourly L bm values plotted versus t ime of day for each an tenna heigh t and each recording period . These graphs illustrate the fact that the range of the variations in L bm

measured at the same hour for 15 or 20 consecutive days is, in general, of the same order of magnitude as the range of the diurnal trend as shown by th e median of hourly L bm values.

Although the hourly fading ranges lis ted in refer­ence [1] show considerable variance , they consistently cluster in the vicinity of the 13.4-db fading range that would be obtained from a pure Rayleigh dis­tributed signal. This tende ncy is apparent regard­less of time of day, although the variance of fading ranges measured at a given hom is somewhat larger

255

TIME BLOCK 2, NOV.-APRIL ,NOON -6 P. M.

160 160 30' ReceivIng Antenna /65' ReceivIng Antenna

170

180 Z oen 190 --I en w !QCD 200 ::!E-en u 210 ZW «0 160 Q:Z f--uen 170 -en eno 180 «-I CD

190

200

~ 't:::: ::::::-

1"-, .-r----.

---"-

365'Receivlng Antenna

I'- i'----- - r--~ t--t::--'~ ~ l'--..

170

180

Zen 190 Q-I enw

200 en CD it> enw 210 zo «z 160 Q:-f-en 170 uen -0 ~-I 180 CD

190

200

~ ~ ' .... .::: ~r---:--

-. -- :-::-r-----~ 565'Receivlng Antenna

........ r---- [:':---'- -p::,~~

.... .~ I'-J:::.....,

210 210 0.1 0.5 I 5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.9 0.1 0.5 I 5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.B

Z ISO Q~ en w 170 !QCD ~t> en w 180 Zo ~ ~ 190 f-en ~ ~ 200

~-I 210 CD

PERCENTAGE OF HOURS PERCENTAGE OF HOURS

Z 0 iii en ~ en Z « Q: f-u en « CD

en -I w CD u w 0

Z en en 0 -I

ISO

~ 665' ReceivIng Antenna 170

180

190

200

210 0.1 0.5 I

'l""-i'--....

1" r--r---

..... :::- I'-~

5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.9

PERCENTAGE OF HOURS

--Predicted ---- Observed

TIME BLOCK 5, MAY- OCT., NOON -6 P. M.

30' ReceivIng Antenna t---- 365 ' ReceivIng Antenna

i""- t'---

~ en 160

iii ~ 170 ~m ::!E-

--::::::: +::::-t-

0.1 0.5 I

1'-. t'--r- ~I- t--

-I---

Cf) ~ 180 Zo « Z 190 Q:-f-Cf) UCf)2OO -0 ~ ...J 210

5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.9 m 0.1 0.5 I

PERCENTAGE OF HOURS

Z 160

:t-.. .... )::"p:: C:- .........

5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.9

PERCENTAGE OF HOURS

OCf)

iii ~ 170 ~m

~ r---- 665 ' ReceivIng Antenna

~ U 180 ZW « a 190 Q:Z f--~ ~ 200

~ 3 210 m 0.1 Q.51

.... __ r---

I'-r---- ~- :: I--r--- ........

5 10 20 30 40 50 60 70 80 90 95 99 99.5 99.9

PERCENTAGE OF HOURS

F IGU RE 1. Obser ved and predicted cumulative distrib1itions of how ·ty median basic tr·ansmission loss (1952 an d 1953 data) .

256

during the night than during the day, particularly in the summer months. The lack of diurnal cycle in fading range and the tendency to approximate the Rayleigh distribution fading range should not, how­ever, be interpreted as evidence that the signals recorded at Quincy were the result of scattering regardless of the time of day. It can be seen in the data samples shown in the technical note that the character of sbort-term variations in signal level changes co nsiderably in going from afternoon to night. The slow, deep fading occurring at night often covered a range equal to, or greater than, the Rayleigh fading range.

One of the principal objectives of the Quincy ex­periment was to study t he effects of antenna height on the signals received far below the radio horizon. During most of t he recording periods simultaneous transmission loss recordings were made at three or more receiving anten na heights. W' e define the heigh t-gain associated with two antenna heights to be the difference between hourly median Lb values measured simul taneou ly at t hese two heights. A positive height-gain indicates a reduction in L bm with increasing height . The hourly height-gain values obtained for each pail' of antennas during each re­cording period were plotted versus time of day. It was found that althou gh the median of hourly L bm

values tended to decrea e slowly with height, there was a large variance in the height-gain observed on anyone pail' of anLennas at a given hoUl' of the day during a given recording period. Examples of these height-gain vers us time-of-day plots are shown in reference [1]. Also in eluded are graphs showing the median of the hourly height-gain values for eacb recording period plotted versus the rat io of receiving a ntenna heights, and the median basic transmission loss, fadingrange, and height-gai n versus time of year.

3 . Comparison of Observed a nd Predicted Basic Transmission Loss

Rice, Longley, and Norton [3] have developed a method of predicting the cumulative distribution of basic transmission loss at frequencies above 10 Mc for wide ranges of path lengths, antenna heights, ter­rain configurations, and atmospheric refractive index gradients (the latter as deduced from observed sur­face values of refractivity). Using this method, the predicted cumula tive distributions of basic transmis­sion loss were determined for the 30, 165, 365, 565, and 665-ft receiving antenna heights for time block two (November through April , from noon to 6 p .m .) and for the 30, 365, and 665-ft antenna heigh ts for t ime block five (May through October from noon to 6 p.m.). These antenna heights wore chosen because of the relatively large amoun t of observed data avail­able for comparison. The surface refractivity data used in this determination were the average of values for these t ime blocks during 1952 and 1953 obtained at the U.S. Weather Bureau stations at Des Moines, Iowa, and Joliet, Ill. These slations lie west and east of the propagation path, respectively. However, t he data obtained at these two points are so well

correlated that we may reasonably assume that they closely approximate conditions on the path.

The predicted distributions are shown in figure 1 along with the corresponding distributions of ob­served values. The latter include all hourly m edians observed during the time block in both 1952 and 1953. Figure 2 provides a comparison of the ob­served height-gain data with the corresponding pre­dicted values. A predicted value is shown at each ratio of antenna heights; the observed value for each ratio is simply the mean of the median height gain values for all recording periods. On the assumption that t he observed decibel values of height-gain are normally distributed, there is a 68 pOl'cen t proba­bility that the true mean lies within the "wings" on the observed points. The fact thaL Lhc predicted values are all above the measured values (and, in­deed, lie outside the wings in most case) indicates a consistent bias in the prediction which is larger at the higher ratios, i.e. , those involving the 30-ft antennas.

I·x Predic ted I Observed I

I

RATIO OF ANTENNA HEIGHTS

T x 1 -0

"' -",

"' "'

T T x 1 1

-0 -0 "' ~--- ·ri--", "' "' "' '"

FIG UR E 2. P redicted and observed height-gain versus mtio of antenna heights for all hours 1952 and 1.953.

The experimen t described }lOre was performed under the supervision of J . VV. Herbstreit. The sur­face refractivity data were supplied by B . R. Bean . Much of the reduction of t he radio data was don e by F. L . Anderson. K . A. Norton and P. L. Rice offered many valuable suggestions concerning Lhe analysis.

4. Refe rences [I} n. B. Janes, J. C. Stroud , and M . T. D ecker, An analys is

of propagation m easurements rnade at 418 M c \\'ell be­yond t he radio horizon, NBS T echni cal Note 6, (PB 151365) $2.25, available from the Office of T echnical Services, U.S. D epartment of Commerce, vVas hington D.C. Foreign remittances must be in U.S. excha nge a nd must include one-fourth of the publication price to cover mailing costs .

[2] 1( . A. Norton, P. L. Rice a nd L. K Vogler, The use of angular distance in estimating Lransmiss ion loss and fading range for propagation through a t urbul ent at­mosphere over irregular terrain , Proc. IRE 43, 10, 1488, (1955) .

[3} P. L. llice, A. G. Longley, and K . A. Nor to n, Prediction of th e cumulative distri bution with time of ground wav e and t ropospheri c wave transmission loss, N BS T ech . Note 15 (PB151374) $1.50, available from the Office of T echnical Serv ices, U.S. D epartment of Commerce, Washington, D .C. Foreign r emittances must be ·in U .S. exchange and must inclu de one-fourth of the publication price to cover mailing costs.

BOULDER, COLO. (P aper 64D3- 55)

257