Writes Kohl: "Most of the rainfall in the U. S. and many spectacular windstorms stem from thunderstorms or cumulo-nimbus cloud activity characterized by the separation of electric charge. When the stress of these charges exceeds the dielectric strength of the atmosphere, a discharge occurs that has far-reaching consequences. The uniform pattern of the electric field is disrupted, and if lightning accompanies the discharge, sferics are radiated. Both the field disturbance and sferics can be detected and measured with relatively simple apparatus, the former over an area of 50 to 200 square miles and the latter for hundreds of miles. The resulting data provide meaningful clues to approaching weather. My initial interest in the electrical aspects of weather has grown into a full-scale avocation which I can recommend to anyone with more than a passive interest in meteorology.

"To the eye, lightning appears deceptively simple: merely a big spark which, it is said, never strikes twice in the same place. This impression vanishes when a discharge is recorded with responsive instruments. Strokes from; cloud to cloud, from cloud to ground or from the top of a cloud to the bottom are nearly always preceded by complex minor discharges, or leaders, which precede the main arc. The leaders advance in stepwise fashion, following routes along the most highly stressed regions between centers of opposite charge. These junior discharges trigger the main arc by moving the centers of charge progressively closer. Fields of extreme intensity result, and the final transfer of charge follows the ionized pathways created by the leaders.

"The leaders usually begin about a thousandth of a second before the main arc, and often advance in as many as 30 steps. Each step produces an electromagnetic disturbance and an abrupt change in the electric field. Not all leaders end in an arc, particularly when they are associated with rapidly growing cumulo-nimbus clouds. They can still be detected electrically.

"Another characteristic of lightning is the multiple stroke of the main arc. Single strokes are uncommon. Multiple strokes are usually completed within a few millionths of a second; otherwise they extend over several thousandths of a second. Few strokes are of intermediate duration.

"Although oscillograms show that no two sferics are alike, all cover a broad spectrum of frequencies. Most of the energy is concentrated in the region below 20 kilocycles. The sferics are detected by means of a rudimentary radio receiver; the number picked up during a given interval can be counted automatically by equipping the receiver with a register. In monitoring sferics on three frequencies (430, 2,000 and 5,100 kilocycles) I have observed that the count at each frequency is related both to the distance the sferic is propagated and to the type of discharge from which it is radiated. The ratio of counts among the three frequencies changes characteristically as the storm approaches or recedes. During late-afternoon storms at a distance of 100 to 200 miles, for example, every 100 counts at 430 kilocycles are accompanied, on the average, by 46 counts at 2,000 kilocycles and 13 counts at 5,100 kilocycles. By the time the storm has moved into the local area, the ratio has changed from 100:46:13 to 100:96:50.

"Cold-front advances and storms characterized by a high percentage of cloud-to-ground discharges show considerably less change in ratio. A comparable ratio for long-distance frontal activity would be 100:74:40. Normally one is chiefly interested in the rate of counts at each frequency. It is possible, however, to set up a coincidence-counter to identify sferics that appear simultaneously on two or more frequencies. This counter shows that nearly 90 per cent of those sferics registered on the 5,100-kilocycle channel also appear on the 430-kilocycle channel, thus confirming their common origin.

"Recordings made in the vicinity of Minneapolis are almost as dramatic as the lightning which produces them. The 480-kilocycle channel has provided consistent data up to a radius of 200 miles with counting rates frequently rising as high as 150,000 sferics per minute! The current flowing in a main arc, which may register as a single sferic, can exceed 20,000 amperes. The duration of the arc is of course measured in millionths of a second. But at the rate of 150,000 sferics per minute the average current flowing between the earth and the cloud is on the order of 500 amperes. Other investigators report a mean energy of some 200,000 joules per full-sized lightning discharge. The corresponding energy liberated over the monitored area could thus amount to 80 billion joules per minute, or 500,000 horsepower!

"On clear days the distribution of the electric field between the ground and the ionosphere is relatively uniform. But violent changes are observed when a storm passes overhead, and the build-up in charge which results in a stroke of lightning is easily observed. Many such build-ups occur in less than 80 seconds. Interactions in the field are as fascinating as they are complex. A single discharge can warp a relatively smooth field so that another stroke follows in a matter of seconds. One also observes many polarity reversals in which the normally negative earth swings positive! These changes, like the leaders which initiate lightning, follow a step-like pattern. In most instances the highly stressed field relaxes slowly without evidence of an arc. From this it would seem that the potential electrical energy in a widespread storm may be much greater than is indicated by the record of sferics.

"For purposes of weather forecasting, measurements of sferics and field intensity need be combined only with readings from a local barometer and wind direction indicator. After a reasonable amount of practice in interpreting this information, reliable eight-hour forecasts; are easy to make. During the summer of 1947 our method gave only one incorrect eight-hour forecast in 50 attempts. The score of the local Weather Bureau for the same period was less than 50 per cent. The Bureau's forecasts were handicapped, however, by an unusual pattern of weather.

"The effectiveness of sferics and field intensity in forecasting lies in the fact that cumulo-nimbus cloud activity provides a sensitive indication of potentially turbulent air conditions during the summer months. An array of these clouds usually appears in late afternoon. They may either develop into thunderstorms or dissipate with the setting sun. On the basis of visual information alone it is impossible to predict which of these two courses they will follow. In contrast variations in the strength of the electric field, coupled with a moderate sferics level, nearly always signal the advance of thunderstorm activity into the local scene, even though the visible clouds have not begun the familiar consolidation that marks the first stage of a maturing storm.

"The sferics record at any location constitutes the summation of weather activity throughout the area covered by the recorder. The profile shows the consolidations, growth and maturity of various thunderstorm cells. A typical record made in the late spring of 1956 is reproduced in the illustration above.

Weather conditions at that time were re ported as a stalled front to the north with high humidity and temperature carried into the region by several days of southerly winds. The heat of the sun caused thunderstorm activity to begin around noon; the thunderstorms were then dissipated as the atmosphere cooled off at night. Full scale on the graph is 200,000 sferics per minute. The rate varied from 21 counts per minute (9:30 a.m. on May 27) to more than 200,000 per minute (after 6 p.m. on May 23). This 10,000-fold increase was intimately related to the weather disturbance.

"In this as well as in many other periods of activity one or two fairly intense consolidations take place prior to the change of weather in the area. These are noted at about noon on both days. Apparently in the complex interactions within a region close to major weather landmarks-high-pressure ridges, low-pressure cells, fronts and so on-the advent of one or two thunderstorms will either trigger general regional activity or else the cumulo-nimbus formations will quickly dissipate. The sustained activity early in the evening of May 28, shown by the record, persisted as the cold front moved across the area. The southerly flow of air apparently blocked the movement of the front a second time; the temperature drop and northerly winds did not materialize until May 31, after local thunderstorms were observed during the night of May 29. Sferic activity from the region west and north persisted during the entire period.

"As for the forecast that was based on this record, it should be noted that thunderstorms were not visible in the distance until about 5 p.m. on May 28. In spite of strong southerly winds on May 27 the decrease in sferic activity early in the evening led to the correct prediction that no change in weather would occur in the immediate vicinity during the night and following morning. A stronger southeast wind appearing at mid-afternoon the next day, coupled with the rapid increase in sferic activity, led to the forecast that turbulent weather would enter the vicinity within several hours. The nearest storm was seen and estimated to be 12 miles from the sferics station. A slow decrease in the sferic level, which occurred during the following morning, added to occasional field disturbances beneath passing clouds and resulted in a further prediction that the weather had not cleared and other storms were imminent. These developed on schedule.

"For rapidly moving storm systems the advance indication on the sferic recorder amounts to only a few hours, and reliable eight-hour forecasts are not difficult. As the observer acquires the knack of interpreting the records, he can also make accurate forecasts under more passive conditions. Observations of the electric field then become particularly useful. Occasionally tremendously impressive cloud formations, including the formidable roll clouds associated with maturing thunderstorms, appear; by conventional standards these would signal foul weather. But when the recorder shows them to be charge-free the observer knows at once that they will be nonviolent. Locally severe storms, on the other hand, are accompanied by violent changes in electric field even though the general sferic level may be fairly low-a clue that the enveloping storm covers a small area.

"With the exception of the electrical pen recorder, the equipment for measuring field intensity and making sferic counts is relatively inexpensive. The sferic detector is essentially a special radio receiver equipped with a means for counting short pulses of current [see illustration on page 156]. The antenna consists of a bare copper wire at least 20 feet long suspended 10 feet OF more above the ground between glass insulators. It serves the dual purpose of picking up sferics and sensing field intensity

"A sferic signal excites the tuned 456-kilocycle transformer, and the resulting oscillations are amplified. The train of amplified oscillations is then rectified and used to trigger a pulse-forming circuit. The output is averaged in a vacuum-tube voltmeter circuit. The antenna should be equipped with a lightning arrester; as a further protection the receiver is equipped with a choke coil designed to block damaging bursts of energy at high frequencies.

"The radio-frequency amplifier consists of two standard 456-kilocycle intermediate-frequency transformers and a high-gain pentode tube. Incoming signals are rectified by a crystal diode that delivers negative pulses at the output. The negative pulses tend to drive the grid of the 6AU6 tube to cutoff, the point at which no current can flow between the cathode and plate. The output of this tube is fed to the 2D21 Thyratron by way of a resistance-capacitor network designed for a pulse duration of 50 microseconds. Most multiple-stroke lightning discharges are thereby registered as a single count. Corresponding pulses generated by the Thyratron tube persist for 100 microseconds. The sensitivity of the circuit to sferics is controlled by adjusting the bias of the Thyratron. Normally this control is set to register not more than one pulse per hour during the winter or during fair summer weather.

"The output of the Thyratron is coupled to a capacitor-resistance 'averaging' circuit. This smoothes out random fluctuations in the sferics counting-rate, and thus gives meaning to recorded values of 50 counts per minute or less.

"As mentioned earlier, the rate at which sferics are detected spans a broad range: from less than 20 sferics per minute to more than 200,000. Hence the scale of the recorder must be compressed as the sferics rate increases to prevent the pen from being carried off the paper. This is accomplished by taking advantage of the grid-cutoff characteristic of the 12AU7 voltmeter tube. As the counting rate increases, progressively higher negative voltage is applied to the grid of the 12AU7, thus reducing the response of the meter.

"The leakage current of the antenna, which varies with the intensity of the electric field, is measured by another averaging voltmeter circuit which utilizes a second 12AU7 tube. One grid is connected to an averaging resistor-capacitor circuit which is connected in turn to the antenna. Some capacity and resistance are inserted to counteract the effects of abrupt changes in the electric field which accompany local lightning discharges. The antenna is so effective that transient changes in the field can induce potential differences in the circuit as high as 10,000 volts! To protect the meter a 22-megohm resistance and a .l-microfarad capacitor are connected in series between the antenna and ground, and the grid of the voltmeter tube is tied to the junction between the two. The capacitor requires an appreciable interval to reach full charge through the resistor. This delay protects the meter against abrupt voltage surges. Sensitivity to local weather is lost by the circuit, however, if transients arc suppressed completely. The resistor is therefore bridged by a small neon tube. When the input voltage exceeds thc firing potential of the tube (70 to 100 volts), the neon glow short-circuits the resistor and in effect transfers the antenna directly to the capacitor and the grid. The five-millihenry choke and .0002microfarad capacitor connected between the antenna and ground present a low resistance to slowly varying fields, such as those produced by nearby power lines, and thus prevent these sources of voltage from firing the neon tube."

Amateurs are occasionally faced with the problem of equipping an instrument with a very fine mechanical adjustment. In November, 1956, for example, this department described an interferometer, the accuracy of which in the measurement of short distances depended on the delicacy with which the moving carriage of the apparatus could be positioned. The problem was solved by equipping the instrument with a machinist's micrometer which moved the carriage by depressing a lever.

George O. Smith, an engineer of Highlands, N.J., suggests an alternative device: the differential screw. Two sets of threads, differing in pitch, are cut in separate segments of a rod. One set engages a nut fastened to the bed of the instrument. The other engages a similar nut attached to the movable carriage. When a knob attached to the rod is turned, one set of threads tends to advance the carriage and the other to return it. The difference in pitch produces a net movement of the carriage, as shown in the drawing on page 164. The threads can be made as coarse as desired and the net movement as fine as desired. Assume that the portion of the rod that engages the bed nut is threaded with 10 turns per inch, and that the portion that engages the carriage nut is threaded with eight. Ten turns of the knob will advance the lead screw one inch. Simultaneously the carriage is returning eight tenths of an inch along its portion of the lead screw; this results in a net movement of only a quarter of an inch. One turn of the knob thus advances the carriage .025 inch.

"The device is not without limitations. For one, it does not lend itself well to long throws of the carriage. The differential screw gains its mechanical advantage at the cost of displacement, which means that quite a lot of lead screw is required to produce a given motion of the carriage. In the case of 10 threads per inch working against eight threads per inch, a carriage translation of one inch requires a total screw length of seven and two tenths inches. Matters worsen as the differential is made less. In the case of 10 threads per inch working against nine, one turn of the knob advances the carriage only a ninetieth of an inch, but requires a total screw length of 19 inches for a carriage translation of one inch!

"All screw-driven devices are subject to errors built into the lead screw, among the worst of which is 'drunkenness,' or variation of pitch within a single turn or fraction of a turn. Drunkenness is present in some amount in the lead screw of the lathe on which the screw is cut, and the error is reproduced in the work. In the case of the differential screw, however, we stand a good chance of averaging out the error because the two portions work against each other."

Roger Hayward, who illustrates this department, has had considerable experience with differential screws. He writes: "Smith's statement that errors due to drunkenness in his screw are as likely to cancel out as to add to each other is in conflict with Murphy's law. (If something can go wrong, it will.) But this is not the only bug in the device. A very real one will show itself in the fit between the screw and the nuts, and in the alignment of the screw with the ways on which the carriage rides. To minimize the effect of drunkenness the lead screw should be lapped against the nuts with which they mate, and the nuts should be supported in gimbals.

"Incidentally, I used a variation of the differential screw in a Slichelson interferometer I once designed for the Mount Wilson Observatory. In this scheme the bed nut engaged a sleeve cut with eight threads per inch. The sleeve served as one bearing for the lead screw, in which 10 threads per inch were cut. Rotation of the threaded bearing was limited by a stop to something less than one revolution [see illustration above]. This variation enables one to position the carriage roughly at the rate of one inch per 10 turns of the knob. The differential principle is brought into play by permitting both screws to rotate (advancing the carriage at the rate of one inch per 40 turns of the knob) for the desired fine adjustment."

Bibliography

ATMOSPHERIC ELECTRICITY. J. Alan Chalmers. Clarendon Press, 1949.

PHYSICAL METEOROLOGY. John C. Johnson. John Wiley & Sons, Inc., 1954.

PRINCIPLES OF MECHANISM. F. Dyson. Oxford University Press, 1939.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.