V. Allegations Regarding Sensors
Electronic sensors buried in the ground in the San Diego Sector alert Sector Communications when they detect movement. When a sensor is triggered, Sector Communications radios the sensor's number to agents in the field, who can then determine whether the movement was caused by alien traffic.
Union President T. J. Bonner made two allegations regarding sensors. First, he claimed that sensors were "inhibited" to prevent them from signaling that alien traffic was present. Second, he alleged that productive sensors were removed from "northern" locations and placed closer to the border to avoid the detection of alien traffic that had made it past the first line of defense. During its investigation, the OIG also received a series of allegations that agents were prevented from responding to sensor reports indicating alien traffic. Some agents alleged that supervisors had told Sector Communications that no agents were available to respond to sensors when in fact there were. Another agent presented us with a sensor list that appeared to instruct agents not to respond to two particular sensors. Others alleged that agents were instructed to acknowledge the call from Communications when a sensor hit, but to do nothing.
After describing how sensors operate, we address these allegations in turn.
A. Background on sensors
Electronic sensors are buried in the ground along the various trails used by the aliens going north. Each is given a unique identifying number, with its location recorded on a map.172 This information is then provided to the agents at the relevant station. The sensor works by detecting seismic movement.173 When a sensor registers movement it is called a "hit." It then resets itself. Thus, if a group of people walks by, the sensor will register a "hit," reset itself, and continue to register additional hits until the group is beyond the sensor's range. The number of hits will depend on how short a cycle the sensor is set and how long the person or thing causing the disturbance is within range. A single person jumping up and down in the area of a sensor may register numerous hits. On the other hand, several people walking past may register fewer hits than the actual number of persons present because some will pass while the sensor is resetting itself.
Because the sensor only registers seismic movement, it cannot distinguish between human traffic, animals, vehicles, or other ground-shaking disturbances such as earthquakes or thunder. One person likened it to a doorbell; it merely signals that someone or something is there and you have to answer it to determine what caused it to ring. A sensor hit, therefore, does not necessarily mean that undocumented aliens are in the area.
When a sensor in one of the five line stations detects movement, it sends a radio signal to the computers in Sector Communications (also known as "Dispatch").174 A "ticket" appears on the computer screen, indicating the sensor number, the number of hits, the date, and the time. A communications specialist then radios the sensor number and the number of hits to agents in the field. An agent responding is supposed to acknowledge that fact over the radio. After determining what set off the sensor, he is to report what he found to Communications. The apprehension of aliens as a result of the hit also should be reported and recorded. The information as to whether someone responded and the result is entered on the ticket and stored in the computer system.
The computer system that records this information - the Intelligent Computer Assisted Detection System (ICAD) - indicates which sensors report the most hits and which ones produce the most apprehensions. The system can also report the percentage of sensor hits responded to, the percentage of identified traffic that is undocumented, and which sensors are not working. This data is analyzed by station sensor coordinators to measure sensor effectiveness and determine the best places to put sensors. Sensor placement decisions are made at the station level, usually by the sensor coordinator with advice from agents in the field. When alien smugglers figure out where sensors are planted, they move to new trails. The sensor coordinator keeps abreast of changes in alien traffic patterns and redeploys the station's sensors accordingly.
There are four ways to alter the sensor functions. First, sensors may be "inhibited," a setting that allows the sensor to continue recording and transmitting information regarding hits, without creating a ticket, for an indefinite period. This generally is used when the need to inhibit a particular sensor depends on conditions that are neither predictable nor constant. It also is used for sensors in areas where no manpower is available to respond. The system continues to record hits, allowing supervisors to determine whether there is sufficient traffic to warrant assigning manpower to the area.
Second, a sensor can be inhibited by setting a timer so that, during some regularly scheduled period, the computer will continue to receive indications of hits but will not create a ticket; the hits will therefore not be called out to the field. Timers are generally used when there is a predictable and regular period when legitimate traffic is expected. For example, sensors near construction sites may be put on timers during the daytime hours when construction traffic will set them off. They will, however, be fully functioning during the night hours, after the legitimate traffic has departed.
Third, a sensor may be turned off so that it will not record hits or transmit any information. This setting is used when the sensor is undergoing maintenance. Finally, the sensor may be deleted from the system. This occurs when a sensor is being moved or withdrawn from service.
B. Allegations that sensors are not being used effectively
Some of the allegations the OIG investigated stated that sensors were being inhibited to eliminate any record of alien traffic and to prevent agents from investigating possible alien traffic in the area. They also claimed that productive sensors far from the border were moved to less productive areas closer to the border, in an effort to understate the amount of alien traffic getting past the initial lines of defense. Finally, they alleged that agents were frequently not permitted to respond to sensor hits that were called out. The purported intent of all these actions was to permit alien traffic to elude apprehension and to eliminate any record of the alien traffic able to get away.
The OIG questioned 270 persons regarding sensors, including 12 communications specialists (the individuals responsible for calling out sensor hits and recording the responses) in Sector Communications, 5 technicians from the Electronics Unit (the unit responsible for physically planting the sensors), and 25 individuals in the sensor units in the various stations (the people responsible for deciding where sensors will be deployed and analyzing the effectiveness of particular sensors). We received numerous maps of sensor locations, records of sensor movements, reports regarding sensor effectiveness, and memoranda and electronic mail messages regarding sensors.
We also requested copies of all of the electronic data stored in the ICAD system. Unfortunately, we were unable to retrieve much historic data because the storage capacity of the system is very limited. By requesting that data be regularly downloaded and forwarded to the OIG, we were able to build over a year's worth of sensor data, but this library begins only in mid-1996, shortly before the OIG began its investigation. Any information from before that period comes from printouts of analyses conducted at various stations. These printouts are far from comprehensive.
1. Allegations that sensors are inhibited to hide alien traffic
Bonner first raised his allegations regarding sensors in the San Diego Union-Tribune on July 7, 1996, when he said that sensors at Imperial Beach and Chula Vista had been "turned off."175 In his July 15 State Assembly testimony and his August 9 Congressional subcommittee testimony, he alleged more generally that sensors had been deactivated to conceal the movement of illegal entrants north of the border. During a July 25, 1996, appearance on a San Diego radio talk show, Bonner claimed that "most of the northern sensors" had been "deactivated."176 He did not identify any particular sensors that had been inhibited, specify dates on which had this occurred, identify who was responsible for ordering the sensors inhibited, or identify any witnesses with knowledge of these events. His allegations were based solely on second-or third-hand information received from agents he failed to identify.
Unquestionably, there are legitimate reasons to inhibit a sensor. It may be malfunctioning. There may be construction or livestock in the vicinity. If the sensor is registering something other than alien traffic, there is no need to create a ticket, call out the sensor number, and ask agents to respond. At every station there are times when particular sensors will be inhibited for legitimate reasons. The issue here is whether sensors were placed on inhibit for inappropriate reasons, such as to conceal alien traffic or prevent alien apprehensions. Although we questioned agents at each of the stations regarding sensors being inhibited inappropriately, the allegations we heard of such misconduct involved only Imperial Beach and Chula Vista.
a. Imperial Beach
Fifty-nine witnesses testified that some sensors at the Imperial Beach Station had been inhibited. Thirty of these did not specify which sensors had been involved. Twenty-six witnesses testified that the "60 series" sensors had been inhibited at various times, and five recalled that the "7-1" and "9-1" sensors had been inhibited at some time. These witnesses included line agents, supervisory agents, station managers, Sector management, and Communications personnel. It was clear, therefore, that sensors had been inhibited. The questions were when and why.
We received widely varying claims about the 60 series sensors - a historically productive group of five sensors located in the third tier of zone three.177 Two line agents, including a Union official, said these sensors had been inhibited prior to Gatekeeper. Other line agents variously said they were inhibited at the beginning of Gatekeeper, sometime in 1995, January 1996, or March 1996. One agent said sensors had been inhibited for months at a time; two said "every night." One supervisor said sensors had been inhibited off and on since 1993. Another said he had inhibited these sensors prior to Gatekeeper. A third said he had inhibited them before and after Gatekeeper.
The general problems with relying on witness recollections regarding the timing of particular events were magnified with respect to sensors because most agents generally had no reliable way to know whether particular sensors had been inhibited, unless they noticed the sudden absence of radio traffic reporting hits on usually active sensors, or saw traffic near a sensor but heard no hits called out. Although several witnesses recalled one supervisor announcing at muster that the 60 series sensors were being inhibited, this was an exception. Thus, most of the agents who testified regarding the 60 series sensors had no direct knowledge that these sensors had been inhibited.
We had hoped to resolve this issue by reviewing Sector records of sensors that had been inhibited. We discovered, however, that prior to 1996 any agent could call Communications and have a sensor inhibited.178 Usually there was no documentation of any such requests, and Communications had not retained what requests it had received. As mentioned above, no electronic records regarding these sensors prior to May 1996 survive. The OIG did obtain copies of sporadic sensor analyses that contain some data regarding these sensors, but because these reports were not created for investigative purposes they often did not contain all of the relevant information.
Despite these difficulties, we believe we can say with some degree of certainty that the 60 series sensors were inhibited on the midnight shift for a short time in 1994 prior to Operation Gatekeeper and for several consecutive weeks in March 1995 and late summer 1995. They were also inhibited on several occasions in late October and early November 1994 on the swing shift. It is unlikely they were inhibited for any additional measurable time, particularly after August 1995.
Several agents recalled these sensors being inhibited prior to Gatekeeper. One recalled calling Sector Communications in the spring of 1994 to ask why these sensors were not being called out. He said that he was told they were on inhibit. Because he actually made this inquiry, we believe this agent's recollection is more reliable than witnesses who relied solely on the fact that they had not heard the sensors called out. His testimony also is consistent with the testimony of a supervisor who acknowledged inhibiting these sensors in 1994, prior to Gatekeeper, and that of a supervisor in Sector Communications who recalled that these sensors had been inhibited sometime in 1994. In addition, Sector Electronics records indicate that all of the 60 series sensors had been pulled by March 8, 1994, and not replanted until October 18, 1994, several weeks after Gatekeeper began. Thus, there were no 60 series sensors to be inhibited from March until October.
One supervisor on the swing shift admitted asking Communications to inhibit the 60 series sensors and several 70 series sensors on several isolated occasions in October and November 1994, after Gatekeeper started. Because this supervisor left Imperial Beach near the end of November 1994 and the 60 series sensors were not replanted until October 18, he could have only inhibited these sensors sometime in late October or early November. He said he did so because he did not have the manpower to cover that area. We also found a line agent who recalled that the 60 series sensors were inhibited on the swing shift for approximately one week during the first few months of Gatekeeper. Because we received no further information that these sensors were inhibited on the swing shift, we believe this agent was referring to the same events the supervisor described.
A supervisor on the midnight shift (No. 1110) admitted putting the 60 series on inhibit periodically - for up to two weeks at a time - before and after Gatekeeper. He was unable to specify when he did so, but explained that he used this setting because he did not have enough manpower to respond to the sensors in the third tier. This suggests that he took these measures in early to mid-1995, when traffic began to increase but Imperial Beach lost manpower because resources were spread further east along the border.
A former supervisor at the Imperial Beach Station (No. 1699) testified that, while working on the midnight shift, he was asked by another supervisor (No. 1110) to draft a request to Sector Communications to inhibit the 60 series sensors from 11 p.m. to 6 a.m.179 He believed that No. 1110 forwarded this memorandum to Sector Communications and that these sensors were placed on inhibit during the midnight shift. He thought they were inhibited for several weeks, but not more than a month. While not certain exactly when he wrote this memorandum, he recalled that it was not while Nos. 1689 and 1694 were also supervisors on the midnight shift. This ruled out any time from October 1995 forward. Because this witness did not arrive at Imperial Beach until January 1995, and was not on the midnight shift between April and the end of September 1995, he must have written the memorandum sometime between January and March 1995.
A second witness (No. 1702) recalled that No. 1699 had showed him a draft of this memorandum and that he had suggested that No. 1699 put the requesting supervisor's name on it instead of his own. He also testified that, shortly thereafter, No. 1110 announced at musters that these sensors were being placed on inhibit because it was too demoralizing for agents to hear these sensors register hits but to be unable to respond. No. 1702, however, recalled that these incidents occurred in March 1996.
One of the two midnight supervisors mentioned by No. 1699 (No. 1694) testified that, to his knowledge, the 60 series sensors were never inhibited while he was a midnight supervisor, a position he assumed in October 1995. This is consistent with No. 1699's testimony. No. 1694 recalled one night when some agents came to him and complained that the 60 series had been inhibited. He called Sector Communications but was informed that these sensors were not inhibited. The other supervisor mentioned by No. 1699, No. 1689, recalled that No. 1110 wrote a memorandum to Sector Communications asking that the 60 series be inhibited because of manpower shortages. However, No. 1689 said this occurred after he had officially transferred to Imperial Beach in October 1995, rather than during one of his earlier temporary details to the station.
No. 774, a supervisor detailed to Imperial Beach from late August 1995 through December 1995, testified that when he arrived at Imperial Beach the 60 series sensors were inhibited on the midnight shift for two or three weeks. He said he and No. 1689 persuaded No. 1110 to take these sensors off inhibit and allow his team to work these sensors. Reports that No. 774 submitted confirm that his detail lasted from late August through the end of December 1995, during which period his unit regularly worked the 60 series sensors.
No. 1110 confirmed this account, testifying that No. 774 came to him when the 60 series sensors were inhibited and offered to have his unit work them. No. 1110 agreed, and, with the approval of the PAIC, the sensors were no longer inhibited. But No. 1110 could not recall when these events occurred. Based on the documents and the persons involved who arrived at Imperial Beach at very specific times, they must have occurred in the second week of September 1995. Given this testimony and a sensor printout for the midnight shift of July 22, 1995, indicating that the sensors were not inhibited that day, we can conclude that the sensors were inhibited near the end of the summer, most likely beginning in late August - but not from mid-September through December 1995.
The sensor coordinator at the Imperial Beach Station through most of 1995 had heard that these sensors had been inhibited on the midnight shift but had never investigated whether this was true. Beginning in December 1995 and continuing through September 1996, however, a new coordinator (No. 1968) began conducting systematic analyses of the sensors. No. 1968, a line agent with no apparent motive to make Gatekeeper appear any more or less successful than it actually was, said he had no information that the 60 series sensors had been placed on inhibit any time during his tenure as sensor coordinator.
A variety of contemporaneous reports confirm No. 1968's understanding that these sensors were not inhibited during his tenure. Weekly Intelligence reports for December 31, 1995, through January 6, 1996, and for January 7 through 13, 1996, prepared by two different line agents, indicated that the 60 series and 90 series sensors were the most active sensors at Imperial Beach. The only complaint was that dispatchers were slow in calling out the sensor hits. No. 1968's testimony was also corroborated by seven detailed reports of sensor effectiveness that he prepared covering various weeks from January through May 1996. In the report for January 10 through 23, 1996, he attributed 165 apprehensions to the 60 series sensors; the 64-1 sensor accounted for 87 of those apprehensions, making it the third most productive sensor at the station. ICAD charts for the 64-1 sensor covering these same two weeks reflect that 46 percent of the apprehensions from this sensor during this two-week period were during the midnight shift. This data makes it unlikely that this sensor was inhibited on the midnight shift during mid-January.
No. 1968's report for the last week of January indicates a 12 percent decrease in apprehensions attributed to the 60 series sensors. Notably, this drop is not particularly dramatic,180 and far smaller than the 63 percent decrease in apprehensions attributed to the 50 series sensors181 - which no one has claimed were inhibited. Moreover, during that same week, the 70 series sensors, which are in the same zone as the 60 series but in the eastern half of the first two tiers, experienced a 32 percent increase in apprehensions. This suggests that the decrease in apprehensions for the 60 series was caused by shifts in traffic and increased operational effectiveness in the first two tiers, not inhibition.182
In his reports covering February 16 through February 29, 1996, No. 1968 remarked on the unusually low levels of apprehensions on the 60 series sensors but concluded, after an investigation, that agents were not accurately attributing apprehensions to these sensors.183 Indeed, other witnesses testified that agents were notoriously poor at calling in the results of their responses to sensor hits. A report for all Sector sensors for February 11 through March 12, 1996, makes clear that the cause of lower apprehensions from February 16 through February 29 was not sensor inhibition. This report lists the total events and total tickets for each sensor for this entire period. When a sensor is on inhibit it continues to register "events" but does not create a ticket.184 If the number of events and tickets are the same, this means that the sensors were not inhibited during that time. The fact that on this report, for each of the 60 series sensors, the number of events and tickets are the same is strong evidence that these sensors were not inhibited from February 11 through March 12, 1996.
We also looked at other sensors in the same zone to determine whether the decrease in apprehensions attributed to the 60 series sensors was unique to that series. Significantly, we found that when two of the five 60 series sensors recorded a large decrease in attributed apprehensions, apprehensions in the overall zone increased, particularly for sensors closer to the border. This suggests that the decrease for the 60 series sensors was due to more effective operations in the first two tiers. Although there were noticeable decreases in apprehensions in the entire Zone 3 during the week of February 23, 1996, they were more than offset by increases in Zone 4 apprehensions, indicating that traffic had moved to the adjacent zone, not that the 60 series sensors were being inhibited.
The significance of No. 1968's testimony that he had no knowledge that the 60 series sensors were inhibited is magnified by the fact that he immediately approached the PAIC when he discovered that the 7-1 and 9-1 sensors had been placed on inhibit for two days. If he had noticed that these two sensors had been inhibited for a mere two days, one would expect him to have noticed if the more prominent 60 series sensors had been inhibited for an equal or greater amount of time. No. 1968 also reported that the PAIC had responded by deciding - and informing Sector Communications - that, henceforth, only the PAIC and the sensor coordinator would be authorized to place Imperial Beach sensors on inhibit. For his part, the PAIC testified that, upon learning that the 7-1 and the 9-1 had been inhibited, he had told the supervisors not to inhibit sensors; he said he had no information that the supervisors had violated his order. A line agent had a similar memory of this order. And an electronic mail message dated December 28, 1995, from the PAIC to the supervisors announced that he had told Communications not to put these two sensors on inhibit because he wanted to monitor traffic.185 Had the 60 series sensors been similarly inhibited at the time, the PAIC presumably would have also ordered the supervisors not to inhibit those sensors for the same reason. And the absence of any such reference is yet a further indication that they were not inhibited. Moreover, this message put the midnight supervisors on notice that the PAIC did not want them to inhibit sensors, making it less likely that they would have done so thereafter.
The communications specialists operating the computer terminals for the ICAD system would likely have known if a sensor had been inhibited because they set the sensors in response to requests from the field. Because they merely responded to orders and were not responsible for keeping apprehensions low, they had no apparent motive to shade the truth regarding this issue. We interviewed 12 of these individuals, including persons from each of the three shifts and several supervisors. Because they rotated on the four terminals, each would at various times have had responsibility for monitoring the sensors at the Imperial Beach. Although eight of these witnesses recalled sensors at Chula Vista being inhibited for up to three months, only four recalled sensors at Imperial Beach being inhibited for any length of time. All four estimated that some sensors at Imperial Beach were inhibited for approximately one month in 1995. None could say when in 1995 this occurred, although one recalled it was "around the beginning" of 1995. This testimony is consistent with a scenario that these sensors were inhibited in March 1995. Significantly, none of these witnesses had any recollection of these sensors being inhibited in 1996.
Having obtained the computer data recording tickets from May 1996 forward, we selected five months spaced over the course of a year to examine sensor data in detail. We found that the 60-1 sensor had been inhibited for approximately two hours in the afternoon of September 15, 1996, during which there were nine events. The 68-1 sensor was inhibited on January 31, 1997, for approximately two hours in the early morning, during which there were 11 events. We found no other evidence that these sensors had been inhibited during the five-month sample.
The evidence as a whole is remarkably consistent, save for several witnesses' recollections of dates. No. 1702 recalled that No. 1699 showed him a memorandum requesting that the 60 series sensors be inhibited on the midnight shift in March 1996 instead of March 1995. But most likely he confused the year. His recollection directly conflicts with that of the drafter of the memorandum. It also conflicts with that of the sensor coordinator, the Communications personnel, and the other midnight supervisors regarding when these sensors were inhibited. We also have conclusive proof that these sensors were not inhibited from February 12 through March 13, 1996, and strong indications from the other sensor reports that they were not inhibited during the second half of March 1996.
No. 1689's recollection that No. 1110 wrote a memorandum to inhibit the 60 series sensors after October 1995 conflicts with the testimony of No. 1694, who joined the Imperial Beach midnight shift in October 1995 and denied these sensors had been inhibited from that time forward. It is also contrary to the testimony of No. 774 that he convinced No. 1110 to take these sensors off inhibit in late August or early September 1995, and that No. 1689 accompanied him when he approached No. 1110. As the rest of No. 774's testimony is supported by contemporaneous documents, we believe he is probably correct as to this detail as well. Although No. 1689 did not become a permanent midnight supervisor until October 1995, he was detailed as a supervisor to that shift for lengthy periods before then. We believe he must have confused the timing of the incident.
No. 1689 also recalled an incident in January 1996 when agents complained to him that the 60 series sensors were inhibited and asked that he request Dispatch to take them off inhibit. In light of the documents reporting the extremely high apprehensions off these sensors in January, particularly on the midnight shift, No. 1689's recollection as to timing is suspect. This would have been just a few weeks after the PAIC had issued his order limiting authority to inhibit sensors to himself and the sensor coordinator - both of whom denied inhibiting these sensors at any time, let alone January 1996. One possible explanation is that this incident occurred earlier. Another may be that this was another case in which agents thought that sensors were inhibited, but learned from Sector Communications that they were mistaken.
We find the evidence described above significantly more probative than the general testimony of the one line agent who claimed these sensors had been inhibited for "months at a time." Save for three months at the beginning of 1995, this agent did not work midnights and had no responsibility for sensors. Thus, he would not have been in a good position to know whether they had been inhibited or not. He was on the midnight shift in March 1995 when the 60 series were apparently inhibited for several weeks, and he most likely exaggerated the length of time involved. The evidence also outweighs the claims of two agents from other stations that they "had heard" from sources they could not identify that the 60 series sensors had been inhibited "every night."
A line agent who was in a specialty unit detailed to Imperial Beach for much of 1995 and 1996 said he recalled working the midnight shift in the vicinity of the 60 series sensors in January 1996 and noticing they were not being activated. On inquiring, he was told that No. 1110 had inhibited these sensors. In light of all of the documentary evidence inconsistent with this claim, we believe it is more likely this occurred prior to that date. Manpower reports indicated that this agent worked the midnight shift at Imperial Beach on numerous occasions throughout 1995.
Another line agent on the midnight shift said that the 60 series sensors were "going off all the time" sometime in "late 1995," and that eventually No. 1110 announced at muster that these sensors were going to be inhibited because there were not enough agents to respond to them. He said he believed that these sensors were turned off "around the clock" for several months. He was the only witness who suggested that these sensors were inhibited for any significant time other than the midnight shift. We know from the evidence described above that these sensors were not inhibited for several months, particularly in "late 1995." Because this agent always worked the midnight shift and did not have any particular reference point that would have limited his recollection to the last quarter of 1995, we believe he was probably recalling the period in August 1995 when the 60 series sensors were inhibited for several weeks.
No. 1922 claimed to have heard from another agent that No. 1110 had asked Dispatch to deactivate "all of the sensors" at Imperial Beach to give the impression that no one was entering the country. There is no evidence supporting this broad claim. Had all of the sensors been inhibited, certainly everyone at Communications and Imperial Beach would have been aware of such an extraordinary event.
A Union official (No. 1750) testified that No. 1688 had reported to him that the 60 series sensors were inhibited and wondered why. No. 1750 said he had explained that all the agents were working up near the fence, leaving no one to work these sensors. No. 1750 did not specify when this occurred or what shift these sensors were apparently inhibited on. Because No. 1688 did not arrive at Imperial Beach until October 1995, were No. 1750 correct, this would show that the 60 series sensors were inhibited after October 1995. No. 1688 testified that after arriving at Imperial Beach, he discovered some sensors were inhibited - but he could not remember which - and asked Communications to put them on. Given the strong evidence that the 60 series sensors were regularly worked from September through December 1995, and the substantial testimony that these sensors were not inhibited after September 1995, we believe it likely that No. 1688 discovered that the 7-1 and the 9-1 sensors had been inhibited.
We also received third-hand information from a Union official that the 60 series sensors had been inhibited around the Christmas holidays. He could not specify the year in which this allegedly occurred or provide any other information - not even the source of this report. We received no other claims that these sensors had been inhibited around any Christmas holidays. We did, however, confirm that the 7-1 and the 9-1 had been inhibited around Christmas 1995, and we suspect that the story was confused in the retelling.
We believe that several factors may have led agents to overestimate the extent to which the 60 series sensors were actually inhibited. As noted above, agents generally have no reliable way to know whether sensors are inhibited or not. Although these sensors historically had been very active, the redeployment of agents closer to the line during Gatekeeper significantly reduced traffic. Agents used to hearing these sensors called out frequently may have assumed, when they no longer heard them, they must be inhibited. Indeed, Sector Electronics sensor maintenance records indicate that the 64-1 sensor was checked on ten different occasions between January 1, 1995, and June 30, 1996. Many of these checks were prompted by agents asking whether there was some problem with this sensor. Usually it was found to be working fine.186
Looking at this information as a whole, we conclude that the 60 series sensors were inhibited on the midnight shift for some period of time prior to Gatekeeper and for several weeks in March and August 1995. They were also inhibited on several occasions on the swing shift in late October and early November 1994. Although there may have been isolated additional days when they were placed on inhibit, the credible evidence rebuts the claim that these sensors were inhibited for any other significant period.
In addition to the 60 series sensors, we received five claims that the 7-1 and the 9-1 sensors - the two northernmost sensors at Imperial Beach, extremely close to Imperial Beach residential neighborhoods - were inhibited. As already noted, these sensors were indeed inhibited in December 1995 for two days, until the sensor coordinator reported this fact to the PAIC, who in turn ordered that they were not to be inhibited. It is possible that these sensors may have been inhibited other times as well. However, because only 4 agents and 1 communications specialist reported that these sensors had been inhibited, in contrast to the 26 witnesses who mentioned the 60 series sensors, we believe they were not inhibited for any length of time similar to the 60 series. We did not locate any supervisors who admitted inhibiting these two sensors, so we cannot identify who was responsible. We do know there was significantly less need to inhibit these sensors than the 60 series sensors because they were not as active. Between February 11 and March 12, 1996, for example, these two sensors (not inhibited during this period) recorded a total of 339 events, while the 60 series sensors recorded 2,336 events.
Regardless of the timing, the more significant issue is whether, as Bonner alleged, any of these sensors were inhibited to hide the fact that there was alien traffic in the tier, in order to make Operation Gatekeeper look more successful. Because inhibiting sensors in early 1994, before Gatekeeper was implemented, would not have done this, we focused on the justification for inhibiting the 60 series sensors in March and August 1995 and the 7-1 and 9-1 sensors in December 1995.
Bonner claimed the sensors at Imperial Beach were inhibited to conceal the extent of alien traffic. Because the sensors still register hits even when they are inhibited his allegation makes little sense. Indeed, many agents, including supervisors, were unaware that the ICAD system still records sensor hits even when a sensor is inhibited. This lack of knowledge probably reflects the lack of training the agents received regarding the system when it was installed. Most agents had no contact with the system other than when Communications called out hits to the field. They merely assumed that if Communications did not call out a hit, the system did not record a hit. Even the supervisors received little information about the system. Although they could obtain reports from the system, they apparently seldom reviewed these reports and received little or no training on what reports could be generated or how to generate them.
Because hits are still registered when a sensor is inhibited and the ICAD reports can show all of these hits, inhibiting sensors does not "hide" evidence of hits. Thus, the supervisors could not make their respective station or shift appear more effective by inhibiting sensors. We believe that this misconception about how the system operates combined with a distrust of management's motives caused agents to inappropriately infer that these sensors were inhibited to hide sensor hits.
Even had the system been able to hide hits, we could find little motive for anyone to have done so. Except for a general claim at the very beginning of Gatekeeper that sensor hits at Imperial Beach had declined due to Gatekeeper, the number of hits was not reported publicly.187 Nor was such information included in the monthly Gatekeeper Updates, the SW Border Weekly, or any of the other reports of Sector operations that made their way out of the Sector. There was no reason to improve the appearance of statistics that few people reviewed.
Moreover, it was generally recognized that the number of hits was not a very reliable indicator of alien traffic because the sensors could be registering movement caused by something other than alien traffic. Thus, such data was not considered outside the Sector when evaluating Gatekeeper's level of success.188 Those primarily concerned with sensor information were at the station level - the supervisors trying to determine whether there were weaknesses in the line of defense and if there were better locations for sensor placement. Any effort to conceal sensor information would merely make it hard to make appropriate deployment decisions.
Of course, the lack of motive does not eliminate the possibility that an uninformed supervisor inhibited sensors to hide sensor hits. If a supervisor believed that the hits were not recorded and that station management was monitoring the effectiveness of his shift, he might order these sensors inhibited on the mistaken belief that he would appear to be a better supervisor. Indeed, according to No. 774, No. 1110 had explained that one of the reasons he had ordered the 60 series sensors inhibited was because there was a rumor of an upcoming efficiency review of sensors and he thought inhibiting sensors for which there was no manpower to respond would affect the efficiency rate. No. 774 said that No. 1110 did not realize that ICAD still recorded the hits on inhibited sensors.
For his part, No. 1110 said he knew that ICAD recorded sensor hits even when a sensor is inhibited. He also testified that he had inhibited these sensors because he did not have enough manpower to respond to the sensors and it was demoralizing to the agents to hear Communications call out sensor hits but not be able to respond. These were reasons that No. 774 - and numerous other witnesses - recalled No. 1110 giving as well.
Despite No. 1110's claim that he was aware that hits on inhibited sensors are still recorded, we believe, based on the testimony of No. 774, that at the time No. 1110 initially inhibited the sensors he was not aware of this fact. No. 774 has no apparent motive to claim that No. 1110 made this statement and it is consistent with the general misunderstanding among supervisors regarding the ICAD system. Although this discrepancy is somewhat troubling, we do not believe it to be very significant. First, No. 1110 may have just forgotten one of his reasons. At the time of his interview almost two years had passed since he had learned how the system worked. And it appears that the rumored efficiency review never took place. Because he had other more significant reasons for inhibiting the sensors, this consideration may not have been sufficiently important to him. Only 1 witness claimed to have heard him mention this reason, while 19 witnesses heard him cite the lack of manpower or the demoralizing effect on agents of unanswered sensors as reasons for inhibiting sensors. Finally, when offered the manpower to cover these particular sensors, No. 1110 readily agreed to have these sensors worked and took them off inhibit. This action belies any claim that he was trying to hide the fact that these sensors were detecting alien traffic.
We also note that without manpower to respond, No. 1110 had good cause to avoid wasting radio time calling out the sensors. We found substantial other evidence that radio time is limited, and eliminating unnecessary air traffic is an important concern. Moreover, agents did not like being assigned to fixed areas of responsibility and hearing possible alien traffic elsewhere that they could not respond to was frustrating to them. Several agents, including some Union officials, agreed that sensors were inhibited to avoid upsetting agents who could not respond and to avoid giving them a reason to leave their assigned posts.
In sum, any suggestion by Bonner and others that the 60 series sensors were inhibited to hide evidence of gotaways is unsubstantiated. The 60 series sensors were inhibited for a relatively short time in 1995, but during that time any hits on these sensors were still recorded. Our investigation shows no attempt was ever made to mislead the public (or anyone else) in this regard. And there were legitimate operational reasons for inhibiting these particular sensors.
Several witnesses also implied that sensors were inhibited so agents would not apprehend aliens. None had any evidence of this other than their suspicions that management wanted to artificially limit apprehensions. To the extent that the claim is that agents were inappropriately ordered to stay on their "Xs," we addressed it in Section II, C-2., of this report. But we have already noted that, when manpower became available from the other stations, these sensors were worked - a fact that belies this allegation.
172 Sensors in a particular area or that follow a particular route may be numbered in sequence so that, for example, at Imperial Beach there are the "60 series" (sensors numbered 61, 62, 63, etc.).
173 There are other types of sensors, but the vast majority of the sensors used are of the seismic variety.
174 Because Temecula and San Clemente are outside of radio range of Sector Communications, their sensors register only at their respective stations.
175 A copy of this article is contained in the Appendix to this report at A-61.
176 Although Bonner used the words "turned off" and "deactivated," there was no evidence from any witness or document indicating that sensors had been turned off or deactivated. As a result of our investigation, we learned that what he was actually alleging was that sensors were being inhibited.
177 At the time, the Imperial Beach Station was divided into six vertical zones across its border with Mexico. It also was divided into three horizontal tiers, with the first along the border and the third the one furthest from the border.
178 After September 25, 1996, only designated supervisory personnel could approve a request that a sensor be inhibited. In addition, Communications created a log to record any action regarding a sensor such as inhibiting, disabling, or removing, as well as the date and time the request was made, who was requesting the action, who had approved the action, and who was taking the action. These new procedures are essential if there is to be any accountability for the sensors. Such accountability is necessary not only for investigations like this one, but also to ensure no one abuses the system in an effort to, for example, assist the illegal entry of vehicles carrying drugs or illegal aliens.
179 This witness testified that he had kept a copy of the memorandum but discarded it with all of his other Border Patrol materials when he left the Border Patrol in the first half of 1996. No. 1110 did not deny that such a memorandum had been written but said he did not recall asking anyone to write a memorandum on his behalf to inhibit sensors.
180 Because early in the month these sensors were very productive and the majority of the apprehensions were on the midnight shift, if they were suddenly inhibited on the midnight shift a decrease closer to 40 percent would have been expected.
181 The 50 series sensors are in the same zone as the 60 series but on the western half of the first tier. Even a small shift east in the traffic could cause a dramatic decrease in apprehensions from the 50 series sensors.
182 The intelligence reports from the first two weeks complained that the dispatchers were slow in calling out sensor hits. This would have given traffic time to move further away from the border before being apprehended and would have allowed more people to reach the third tier. If the dispatchers improved, as was reported by the second of these reports, more alien traffic could be apprehended off the sensors in the first two tiers instead of the last tier.
183 A decrease in apprehension due to misallocation would result in similar numbers of tickets being reported as previously, but fewer apprehensions. A decrease due to sensors being inhibited would report fewer tickets and the number of events (reports of sensor activity) and tickets would be unequal. Thus, No. 1968 could have easily distinguished between a decrease due to a misallocation of apprehensions versus inhibiting sensors.
184 Each event is a report to the computer that the sensor is registering hits. An event may be caused by a single, isolated hit or by a series of uninterrupted hits. When there is a break of a certain minimum time after a hit or hits, that completes the event. A ticket is created for each event, not hit.
185 The sensor coordinator testified that this had occurred in January 1996. Based on the date of the electronic mail message, we believe his recollection was off by a few days.
186 On other occasions this sensor was found to be too weak or not sensitive enough. These problems also might have led agents to assume it was inhibited.
187 Sector Communications personnel confirmed the accuracy of this initial claim. They all noted a significant decrease in the levels of sensor hits once Gatekeeper was in place. Because they actually monitored the computer system recording the sensor data, they were in the best position to know whether such a statement was accurate. Furthermore, because they are not Border Patrol agents and are not managers responsible for the success or lack of success of the operation, they had little or no motive to overstate the extent of the change in sensor records.
188 We did learn that, for at least some time, ICAD data was forwarded to a statistician in INS headquarters for possible analysis as a measure of Gatekeeper's success. That individual testified, however, that he never assessed or reported the data. Furthermore, there was no evidence that the various station supervisors had any knowledge that this raw data was forwarded outside the Sector, let alone any reason to believe that their station would be evaluated on the basis of it.