Todd pushed a key on the podium and a line drawing map of the American east coast, including the area from Labrador through Cuba, appeared on the left screen. “The missile was a final test version,” he continued, “in the exact configuration of the production flight vehicle, except for the command test set and the warhead. This was to be the longest test flight yet conducted and was designed to demonstrate thoroughly the new 4.2 version of the software that was recently installed in the APRS. So of course the missile was not armed.”
The lieutenant picked up a light pen from the podium and marked on the small monitor in front of him. His markings were immediately translated to the larger screen behind him so that everyone could easily follow his discussion. “On the screen you all can see the predicted versus actual overflight path of the bird yesterday. Here, roughly ten miles east of Cape Canaveral on what appeared to be a nominal flight, the sequencer turned on the cameras. After a couple of hundred calibration images, sort of a self-test of the APRS, the terminal guidance algorithms were activated as scheduled. As far as we can tell from the realtime telemetry, nothing strange had occurred until this time.”
The right screen now showed a detailed map of south Florida and the Keys that included the target in the Bahamas. The maps on the two flanking screens remained in view during the rest of his presentation but Lieutenant Todd constantly changed the word charts in the middle to keep up with the discussion. “The a priori location of the target, which was where the cameras should first have looked for the aircraft carrier, was here at Eleuthera, in the Bahamas. The search algorithm should have fanned out in a circle from there and, if it had operated properly, found the target in about fifteen seconds. This (Todd pointed toward a dotted line on the more detailed map) should have been the impact trajectory.
“However,” Todd continued dramatically, “based on the telemetry data that we have analyzed to date, it appears that the missile veered sharply westward, toward the coast of Florida, soon after the terminal guidance system was activated. We have only been able to reconstruct its path up to this point, where it was about three miles west of Miami Beach at an altitude of ten thousand feet. After that the telemetry becomes intermittent and erratic. But we do know that all the terminal guidance engines were on at the time we lost complete data. Projecting the total control authority for the missile, the area highlighted here, covering the Everglades, the Keys, and even as far south as Cuba, represents where the bird might have landed.”
Lieutenant Todd paused for a second and Commander Winters, who had been writing down major points in a small notebook during the presentation, immediately jumped in and started taking charge of the meeting. “A couple of questions, Lieutenant, before we proceed,” Winters began in a businesslike manner with an obvious overtone of authority. “First, why was the missile not destroyed soon after it veered off course?”
“We’re not exactly certain, Commander. The command test set and the small ordnance had been installed, of course, specifically for that purpose. The change in the motion of the vehicle was so sudden and so unexpected that we reacted a little slowly at the beginning. By the time we sent the command, it’s possible that we were out of range. All we know is that we never saw an explosion of any kind. We can only assume — ”
“We’ll come back to this operational error later,” Winters interrupted him again. Todd blanched at the word “error” and fidgeted behind the podium. “Where would the impact point have been according to the guidance constants active at the time of the last complete telemetry packet? And how long is it going to take us to extract additional information from the intermittent data?”
Lieutenant Todd noted to himself that the commander was sharp. Winters had obviously been associated with anomaly investigations before. Todd then explained that if the active guidance constants had not changed again, the continued firing of the terminal engines would have brought the missile to an impact point about twenty miles south of Key West. “However,” Todd added, “the constants were allowed, by the software, to change every five seconds. And they had changed in two of the last five internal data updates. So it’s unlikely they stayed the same as they were when our complete telemetry terminated. Unfortunately, although all the constants — even the future predicted ones that are being calculated by the APRS-are stored in the onboard computer, because of bandwidth limitations we only transmit the active constants with the realtime telemetry. We are now going through the dropout data manually to see if we can find out anything more about the constants.”
One of the other staff officers asked a question about the probability of the missile actually having reached Cuba. Lieutenant Todd answered “very low” and then activated an electronic overlay that placed a dotted and blinking trajectory on the right screen inset map. The blinking dots followed a path that started just off Coral Gables, south of the city of Miami, and then continued across a portion of south Florida into the Gulf of Mexico, across the Keys, and finally into the ocean again. “It is along this line that we intend to concentrate our search. Unless the bird suddenly changed its mind its general heading would have been consistent with a perceived target located anywhere along this path. And since we have no reports of any land impact near a populated area, we assume that the missile landed in the Everglades or the ocean.”
Lieutenant Todd had consulted briefly with Winters the previous evening on the agenda for the meeting. It had been scheduled to last only an hour, but the number of questions caused it to stretch to an hour and a half. Todd was thorough and precise in his presentation but was obviously dismayed by Winters’ continued probing into the possibility of human error. The lieutenant freely admitted that they had blown the procedure to destroy the missile when it went awry, but defended his men by citing the unusual circumstances and the nearly perfect previous record enjoyed by the Panther missile. He also explained that they were going to equip their search vessels with the best possible instrumentation (“including the new ocean telescope developed by the Miami Oceanographic Institute”) and begin searching the outlined areas in earnest the next day.
Winters asked many questions about the possible cause of the missile’s strange behavior. Todd told him that he and his staff were convinced that it was a software problem, that some new or updated algorithm in the 4.2 version of the software had somehow scrambled both the initialization sequence and the optically stored target parameters. Winters accepted their opinion eventually, but not until he ordered them to prepare a “top down” failure modes analysis that would list every possible hardware, software, or operational error (Todd winced when Winters mentioned operations again) that could lead to the kind of problem observed.
Toward the end of the meeting Winters reiterated the secrecy of the activity and pointed out that the Broken Arrow project was to remain completely unknown to the press. “Commander,” Todd broke in while Winters was explaining the press policy. The lieutenant had begun the meeting with confidence but was feeling increasingly unsettled. “Sir, I had a call late yesterday afternoon from a reporter, a Carolyn or Kathy Dawson I think, from the Miami Herald. She told me that she had heard of some special activity down here and that I was supposedly connected with it. She claimed her source was someone in the Pentagon.”
Winters shook his head. “Shit, Lieutenant, why didn’t you say something before this? Can’t you imagine what will happen if the word gets out that one of our missiles wandered over Miami?” He paused. “What did you tell her?”
“I didn’t tell her anything. But I think she is still suspicious. She called the public affairs office after she talked to me.”
Winters gave an order that the existence of the Broken Arrow investigation was to be kept classified and that any and all inquiries about it were to be referred to him. He then called for the next status meeting at 1500 on the following day, Friday, by which time (he told Lieutenant Todd) the commander expected to see the results of the analysis of the intermittent telemetry, a more complete logic breakdown of the failure modes, and a list of recent open items with the 4.2 software.
Lieutenant Richard Todd left the meeting aware that this assignment was going to have a significant impact on his career. It was clear to the lieutenant that his personal competence was already being questioned by this Commander Winters. Todd intended to respond to the challenge in a positive way. First he called a small postmortem meeting of the junior officers in his group. He told them (they were all young ensigns, just out of the university after completing a Navy ROTC program) that their collective ass was on the line. Then he defined a series of action items that would keep all of them up working for most of the night. It was imperative to Todd that he be properly prepared for the next meeting.