Examining the Explorer I data

There was a substantial delay, sometimes exceeding three weeks, between the record­ing of data at the ground receiving stations and appearance of the data in our Iowa City laboratory. That was due, of course, to the time required for the shipment of data tapes from the widely spread sites, for clearing customs in some cases, for some limited processing and examination of the data at JPL, and for shipment of the tapes and charts from JPL to Iowa City. As of 5 February, the date on which I left Iowa City on my way back to Pasadena after the Explorer I launch, none of the Explorer I data had yet reached Iowa City for analysis.15

Even after the next week or two, only a sparse set of data was available. It consisted mostly of short segments (of the order of a minute each) from stations widely dispersed geographically. The positions of the satellite at the times when those bits and pieces were recorded were uncertain, as the accuracy of the satellite orbit computation in Washington was still evolving.

Even by the end of February, the situation remained quite tentative, as seen by examining an SUI progress report released on 28 February.16 Sanborn strip-charts from 74 passes over Microlock stations had been received from JPL. Of those passes, all but seven were from U. S. stations, that is, from locations where the satellite was near its northernmost excursion and not far above its lowest height, that is, outside the region of trapped radiation.

In addition to the Sanborn charts, magnetic tapes with copies of the appropriate channels from the ground station recordings had been received for 66 passes. In most cases, those were not the same passes as those for which the Sanborn charts had been received.

By 23 February (the cutoff date for the data analysis that was summarized in the 28 February progress report), SUI had examined, read, and plotted the data from 54 of those passes. Thirty-four (largely from stations in California, where the satellite

OPENING SPACE RESEARCH

was near its lowest altitude) produced clean and readable data. They showed cosmic ray counting rates within the expected range of 12 to 80 counts per second.

Although a substantial number of tapes from the Naval Research Laboratory’s (NRL’s) Minitrack stations had arrived at JPL by that time, none of them had been processed. Thus, none of the recordings made during deep excursions into the region of intense radiation were yet available.

The 28 February report did briefly mention a curious anomaly:

An apparently valid case of a temporal variation was observed in a record from the J. P.L. receiving station at 0123 UT on 2/5/58 [5 February]. An inadequate amount of data has been received to confirm the Japanese report of temporal variations at 0709 UT on 2/5/58 and at 0857 UT on 2/7/58. Reports of apparent temporal variations must be viewed with caution since the normal counting rate varies markedly with latitude and altitude.17

The complete tabulation of Explorer I data published much later reveals that the JPL pass mentioned here had shown a normal rate until about mid-pass, when it began edging upward, reaching 109 counts per second at the end of the pass. Thus, it was making a very shallow approach to the trapping region. Final analysis of the Tokyo passes revealed that the rate peaked at 324 counts per second and then decreased to a normal rate during the 5 February pass. The tabulations reveal that no data were received from Tokyo or any other station on 7 February at the time cited above. Thus, the one Tokyo pass did also indicate a mild incursion into the trapping region.

Records from Singapore and Nigeria deserve special comment. A later assessment revealed that the instrument had frequently risen well into the high-intensity radiation region over those stations during the 24 day period covered in the 28 February report. Such incursions occurred on 8 days at Singapore and on 14 days at Nigeria. However, no significant portion of those data had yet reached Iowa City by the time of the February report.

As a side note, the tapes mentioned above were just the beginning of the torrent of Explorer I data that eventually descended upon our laboratory. During just the period from launch through 23 February, a final total of 474 station recordings were of good enough quality that the beginning and ending times of radio frequency signal reception were eventually determined. Of those, 334 yielded at least short bursts of useful cosmic ray data. By the end of the satellite operational lifetime, the accumulated totals grew to 877 and 592, respectively.

In summary, the report on 28 February indicated satisfactory performance of the Explorer I instrument and included a few guarded comments about what were being referred to then as “temporal variations.” The general situation in Iowa City at the end of February can best be characterized as a mixture of jubilation that we were receiving useful data from the first U. S. satellite, frustration that we were not receiving

CHAPTER 12 • DISCOVERY OF THE TRAPPED RADIATION 327

the data more quickly, and perplexity about several mildly abnormal readings. The photograph on the cover is an accurate portrayal, even though staged sometime later, of the puzzling situation that existed at that time.

The first Deal II launch attempt was made on 5 March, but the satellite failed to enter orbit, as related earlier. I stopped at Iowa City over the 8-9 March weekend on my way back from the Cape to Pasadena. Our small cosmic ray group gathered in the physics building that Saturday to discuss the early Explorer I data. Van Allen, Ernie Ray, and I were joined by Carl McIlwain, who had recently returned from his Fort Churchill rocket-launching expedition and had just joined the Explorer I data effort. The three of them, with help from student data readers, had been trying to make sense of the unexpected “anomalous” readings.

The situation remained additionally clouded by the continuing lack of accurate orbital information, since minute-to-minute satellite orbital positions from the IGY Vanguard Processing Center had also not yet started to arrive. Joe Siry and his people in the Vanguard Computing Center in downtown Washington were working heroically but were still having difficulties in computing an accurate orbit, partly because the software and tracking system were new and had to go through a period of initial shakedown, but also because the orbit apogee was substantially higher than expected.

To further complicate the situation, both the tasks of interpreting the data and computing the orbit were complicated by the transition of the Explorer I motion from its initial axial spin to an end-to-end tumbling, as described in the previous chapter. The signals from both the high – and low-power satellite transmitters, although they were operating at their full design power levels, were very weak by any ordinary standards. The deep spin modulation that resulted from the change in the spin configuration resulted in additional dropouts in the data that complicated their interpretation.

Thus, even by the time of my 8 March visit, we still possessed only isolated segments of Explorer I data, mostly with durations of a minute or less. And the locations of the instrument in space where the data had been captured were uncertain.

Despite those factors, a discernable pattern was emerging. There were periods during which the counting rate was as expected from the primary cosmic rays, other periods during which the counting rate was much higher, and still other times during which the counting rate appeared to be zero. One especially intriguing case had been seen of a smooth transition during a station pass from a zero counting rate, to a high rate, and then to a normal cosmic ray rate.

My clear recollection of that 8 March meeting is of the four of us (Van Allen, Ray, McIlwain, and me) gathered around the conference table in Van’s office for an extended assessment of possible causes of the strange readings. The moment was later reposed and captured on film, and that photograph became the model for a 1962 painting by artist Robert Tabor that still hangs in Van Allen Hall at the University of Iowa.

OPENING SPACE RESEARCH

Either there was some systematic misbehavior of the instrument, or we were seeing a real physical phenomenon that was completely unexpected. We pored over every possibility we could think of to explain the anomalous rates. We debated, for example, whether the unusual observations could be the result of temperature effects on the GM counter or associated electronics. But there was no correlation between the anomalous rates and the passage of the satellite through the Earth’s nighttime shadow, where the instrument temperatures dropped to their lowest values. The only consistent correlation appeared to be with the satellite’s position relative to the Earth’s magnetic field. The high and zero apparent rates seemed to occur consistently only when the satellite was both closest to the magnetic equator and at its highest altitudes.

In preparing for the Fort Churchill expedition from which he had just returned, Carl Mcllwain had wanted to find out how his detectors would respond to the high-intensity radiation that had been seen in the auroral zone during our earlier rockoon expeditions. To do that, he talked Van Allen into buying an expensive 250 KV direct current X-ray machine and set it up in the north end of the basement. His preexpedition calibrations revealed that that machine could easily drive the GM counters to zero counting rates.

During our March meeting, Carl noted that the zero counting rates being seen in the Explorer I data seemed similar to those he had seen during the calibration of his Churchill instruments. He suggested that the Explorer satellite instruments might be observing a high flux of particles that was blanking the GM counter. He further suggested that study of transitions from normal to abnormal rates might help resolve the riddle. Abrupt transitions would suggest an instrument malfunction, while smooth transitions would suggest that the satellite was moving into and out of a region containing a high radiation flux.18

Thus, at that 8 March meeting, we were already beginning to seriously entertain the possibility that there might be a higher than anticipated flux of charged particles in certain regions above the atmosphere. That information, however, was held very closely to our collective chests until we could be more certain of the situation. Not even other members of the physics faculty were aware of the development at that time.

The Iowa satellite experiment as originally envisioned did not call for any particular urgency in dealing with the flight data. As previously stated, its original purpose had been to assemble a set of data that was comprehensive in its spatial and temporal coverage. We envisioned that the data set would be assembled throughout the satellite’s lifetime and that the analysis process would be done in due course during the following months.

We had no way of anticipating the intensity of the spotlight that soon focused upon our work. The cold war space race introduced a new reality—pressure for early

CHAPTER 12 • DISCOVERY OF THE TRAPPED RADIATION 329

results precluded the more measured pace that had been envisioned. Added to that, the unexpected readings created in us an even greater sense of excitement and urgency. We were beginning to understand that we might have come across a new finding of great importance.

During the weeks following our 8 March meeting, while I was busy at JPL with preparations for the second launch attempt of our more complete Deal II instru­ment, Van Allen, Ray, and McIlwain, with a team of student data readers, continued their examination of the Explorer I data. Nevertheless, the data remained largely in­conclusive.

No indication of our growing belief in the possible existence of a naturally occur­ring region of high-intensity radiation appears to have been made at the previously mentioned 11-12 March meetings at JPL. Throughout the rest of that month and early April, there were other external contacts in which preliminary results from Explorer I were discussed. On 20 March, for example, Van Allen talked to a CBS reporter about our results—again, there is no indication that the anomaly was mentioned.19 On 17 March, I talked to a physicist at Northrop Aircraft who was working on a paper on radiation hazards in space. During that discussion, I outlined our general progress to date but made no mention of elevated counting rates.20

Van Allen wanted to be sure of our analysis before mentioning it outside our group.

In the general excitement that followed the successful Explorer I launch, a new satellite program was approved and announced that would capitalize on the now-proven launch capability. The program, with Juno II as the vehicle name, would use a larger Jupiter Intermediate-Range Ballistic Missile instead of a Redstone as its first stage, combined with the Juno I upper stages. It would be able to launch a substantially heavier satellite. On 17 March, Van Allen informed me that our Iowa group would be furnishing an instrument for that so-called IGYHeavy Payload.

Its initial schedule called for a very short developmental period. We were expected to have our instrument prototype ready by 1 June. Thus, the development of our new instrument became an urgent priority for me. As circumstances evolved, that instrument was not actually launched until over a year later as Explorer 7 (after an intervening failed attempt). But those delays could not have been foreseen during the period March through May, when I had to scramble to meet the original schedule. The story of the development and launch of that satellite is recounted in Chapter 14.

Meanwhile, time marched inexorably toward the second attempt to launch Deal II. The Deal IIb payload was launched from Cape Canaveral at midday on 26 March (12:38 EST or 17:38 UT), as described in the previous chapter. After launch, it was renamed Explorer III.

OPENING SPACE RESEARCH