United States v. Williams

Full title: United States of America, Plaintiff, v. Michael Anthony Williams - 005…

Court: United States District Court, District of Arizona

Date published: Feb 22, 2023

CR-18-01695-TUC-JAS (EJM) (D. Ariz. Feb. 22, 2023)

Opinion

CR-18-01695-TUC-JAS (EJM)

02-22-2023

United States of America, Plaintiff, v. Michael Anthony Williams - 005, Samuel Lee Berrelle Rakestraw, III - 004, Defendants.

Eric J. Markovich United States Magistrate Judge

REPORT AND RECOMMENDATION

Eric J. Markovich United States Magistrate Judge

Pending before the Court is a Motion to Preclude Government's Proposed Cell Phone Location Expert Witness Sonnendecker filed by counsel for defendants Michael Williams and Samuel Rakestraw. (Doc. 1610.) The defense argues that the proposed expert is not qualified to testify about the proffered opinions. The defense also argues that the opinions and methodology used to form the opinions are not reliable because they lack foundation.¹ The government obviously argues that their expert is qualified, and his opinions and methodology used to form the opinions are reliable. For the reasons discussed below, the Court concludes that the methodology used to form the expert's opinions are not reliable because there is insufficient foundation for the opinions. Accordingly, it is recommended that the District Court grant the Motion to Preclude opinion testimony.

1 The defense also argued that the government's Notice of Expert Witness was deficient because it did not detail the specific opinions offered or the basis for the opinions. The District Court ordered the government to provide more specificity. As discussed in text infra, the government did so in a supplemental pleading. (Doc. 2084.)

FACTUAL BACKGROUND

1. The Charges.

On April 6, 2022, a federal grand jury sitting in Tucson, Arizona returned a Third Superseding Indictment against Michael Williams, Samuel Rakestraw, and seventeen other individuals. (Doc. 1425.) The charged offenses pertain to an alleged criminal enterprise operated by a gang known as the Western Hills Bloods (“WHB”). Williams and Rakestraw are charged with the following four felony offenses. Count One charges Williams and Rakestraw (and other co-defendants) with participating in a RICO conspiracy, in violation of 18 U.S.C. §§ 1962(d) and 1963(a), the objects of which are: (a) acts involving murder (18 U.S.C. §§ 1959(b)(1) and 1961(1)); (b) offenses involving drug trafficking (21 U.S.C. §§ 846 and 841); and (c) acts involving the obstruction of justice (18 U.S.C. § 1512). Count Two charges Williams and Rakestraw (and other co-defendants) with Violent Crime in Aid of Racketeering - Conspiracy to Commit Murder, in violation of 18 U.S.C. § 1959(a)(5). Count Three charges Williams and Rakestraw (and other co-defendants) with Violent Crime in Aid of Racketeering - Murder, in violation of 18 U.S.C. §§ 1959(a)(1) and 2. Count Four charges Williams and Rakestraw (and other co-defendants) with Use of a Firearm During and in Relation to a Crime of Violence Resulting in Death, in violation of 18 U.S.C. §§ 924(j), 924(c)(1)(A)(i), (ii), (iii), and 2.

Michael Williams is also charged with the following additional offenses which are not alleged against Rakestraw: (1) Counts 12, 13, and 16: Possession of a Firearm by a Prohibited Person, in violation of 18 U.S.C. §§ 922(g)(1) and 924(a)(2); (2) Counts 14 and 17: Possession of a Firearm in Furtherance of a Drug Trafficking Crime, in violation of 18 U.S.C. §§ 924(c)(1)(A)(i) and 2; (3) Count 24: Conspiracy to Possess with the Intent to Distribute Cocaine, in violation of 21 U.S.C. §§ 846 and 841; (4) Count 25: Possession with the Intent to Distribute Cocaine, in violation of 21 U.S.C. § 841; (5) Count 33: Conspiracy to Possess with the Intent to Distribute Marijuana, in violation of 21 U.S.C. §§ 846 and 841; (6) Counts 34 and 37: Possession with the Intent to Distribute Marijuana, in violation of 21 U.S.C. § 841; and (7) Count 38: Possession with the Intent to Distribute Cocaine, in violation of 21 U.S.C. § 841.

2. The Motion to Preclude.

On May 8, 2022, counsel for the defendants filed a Motion to Preclude the proffered expert testimony of ATF Special Agent Mark Sonnendecker. (Doc. 1610.) The proffered testimony relates primarily to the physical location of the defendants' phones before, during, and after the charged crimes were committed. The defense argues that Agent Sonnendecker does not have the requisite qualifications to provide expert testimony, and that even if he does, his opinions lack sufficient foundation and are therefore unreliable. The defense requested a Daubert hearing to address both arguments.

In their Response, the government set forth Agent Sonnendecker's qualifications and summarized the proffered expert testimony and the basis for that testimony. Pursuant to the District Court's Order, the government subsequently filed a Notice that more fully set forth Agent Sonnendecker's qualifications, the specific opinions he will testify to at trial, and the basis for the opinions. (Doc. 2084.)

In their Notice, the government set forth fifteen subject areas of proposed expert testimony. These subject areas can be divided into two categories: (1) that cell phone numbers attributed to various people - i.e., the defendants, co-defendants, a cooperator, and a victim - accessed certain cell sites whose coverage areas encompassed relevant locations (e.g., crime scenes), which indicates that the phone was near the same location; and (2) that there were cellular events - i.e., incoming and outgoing calls and/or text messages - on various dates and times for cell phone numbers attributed to some of these same people. There is an additional general subject area of proposed expert testimony which is not mentioned within the fifteen subject areas: the movement of cellular devices based on accessed cell sites. However, the Notice does not identify the specific cellular devices that will be the subject of that proposed expert testimony. For that reason, at the Daubert hearing, the defense objected to any testimony about the movement of cellular devices because the government's Notice did not specify the cellular devices that would be the subject of that expert testimony. The Court overruled that objection for purposes of the Daubert hearing, but reserved decision on whether the government's Notice complied with the District Court's Order requiring the government to set forth all of the proffered expert opinions with specificity.

Direct Examination by Government Counsel:

Mark Sonnendecker has been a Special Agent with the Bureau of Alcohol, Tobacco, and Firearms (“ATF”) for nineteen years. 11/15/22 Tr. at 8. He is currently assigned to ATF Special Operations Division where he manages a national cellular analytics program. Id. In that position, he assists “investigators nationwide with not only the proper way to obtain cellular records, but how to analyze them and how to apply them to different types of investigations.” Id.

Agent Sonnendecker obtained a bachelor's degree in organizational communication and a master's degree in criminal justice. Id. at 9. After completing his undergraduate education, he worked as a drive tester for T-Mobile (as a contractor) for four months. Id. at 9-10. Agent Sonnendecker explained that at that time “T-Mobile was building their wireless voice and data system in the Cleveland market, and [he] would work with the engineers and use equipment to test the cellular signals that were being emanated from the various antennas being placed throughout the city, in order to test them.” Id.

Agent Sonnendecker's work as a drive tester helped him understand how cellular networks work. Id. at 10. Specifically, he worked with T-Mobile engineers to build that network and “make tweaks to make things better.” Id. He explained that “when you're in a more densely populated place, there will be more cell sites that are designed to cover less of an overall distance. As you start to move out away from densely populated places into more rural places, there will be less cell sites that are designed to cover larger distances.” Id. Because there are less people in rural areas, “it is not as necessary to have as many antennas covering” that area. Id. The geography also impacts where cell sites are located. Id. at 11. If an area is flat, it is easy for radio signals to travel in “a line-of-sight type scenario.” Id. But if there are rolling hills, valleys, or mountains, those obstructions may create complexities when setting up the antenna systems. Id. As a result, antennas need to be put in different places to take the obstructions into account. Id.

Agent Sonnendecker also worked for a company called TRW, which “was a company that was contracted to build a wireless data system for the State of Ohio, to be used by first responders throughout the entire state.” Id. His job was “to create the documentation to provide to the end users” about how the cellular system was put together and worked. Id. at 11-12. That job required him to have a good understanding of the cellular network. Id. at 12.

Testimony turned to how antennas work for a cellular network. Id. Agent Sonnendecker explained that “[c]ell phones themselves are essentially radios” which communicate with antennas placed on cell sites throughout a city. Id. Antennas can be placed on an actual tower, as well as on structures such as a church, water tower, or light pole. Id. The direction that the antennas point matters for purposes of building a cellular network. Id. “Generally speaking, cell sites are designed to cover 360 degrees, a complete circle.” Id. Many cell sites are broken up into “sectors or portions of that 360-degree circle.” Id. at 13. For example, an antenna may be designed to cover “one-third of that 360-degree circle, or 120 degrees.” Id. That said, there are cell sites that are “omnidirectional and there is no sector; it covers 360 degrees. It's all contingent upon how the network engineers lay out the network.” Id.

Antenna placement was important when Agent Sonnendecker was a drive tester and when he helped develop networks. Id. He explained that “[i]t was critical because as we were testing the network, we were looking to see where there [was] poor signal quality in a given area.” Id. If the drive tester saw poor signal quality, the engineers would “either adjust the antenna itself or possibly increase the strength of the radio signal that was being sent . . . in order to better the network in that given area.” Id. However, the engineers, and not Agent Sonnendecker, were responsible for deciding what needed to be done “to make signals stronger and increase coverage.” Id. at 14.

Agent Sonnendecker has received certifications relevant to cell phone technology while he has been with ATF. Id. He has certifications specific to the world of digital forensics, forensic computer and cell phone examinations, and from Cellebrite, which is a digital forensics company that specializes in mobile forensics. Id. at 14-15.

Agent Sonnendecker has been qualified to testify as an expert on cellular records and geospatial analysis on 19 occasions. Id. at 19. “‘Geospatial analysis' is just a way of saying analysis of location-based data.” Id. at 20. Agent Sonnendecker has “provided training to law enforcement personnel throughout the country on at least 75 occasions relating to either digital forensics or to the analysis of historical cellular records or live time tracking of cell phones[.]” Id. at 16. He has also “received well over 250 hours of formalized education as it relates to cellular records,” which includes analysis of historical cell phone data. Id.

Agent Sonnendecker has been involved in “drive scans” during his time at ATF. Id. at 15. Drive scan equipment is used to conduct a cellular forensic analysis of the cellular networks to identify cell sites and antennas that provide coverage to certain areas. Id. ATF did not have drive scan equipment in 2015; it only became available to ATF within the past year. Id.

Agent Sonnendecker has done well over a thousand forensic examinations on digital devices. Id. at 17. He has analyzed sets of cellular records “well over 2,000 times.” Id. “Live time tracking” is used when law enforcement needs to locate a cell phone in live time to figure out where a cellular handset could be located. Id. at 17-18. Agent Sonnendecker's participation in “live time tracking” has helped build his knowledge of how cell sites work and where a handset might be located given the cell sites that it connects to. Id. at 18.

Agent Sonnendecker keeps up with the latest literature pertaining to “what's going on in the cellular analysis world[.]” Id. He also speaks with peers who are involved in cellular analysis (both historically and prospectively), which includes law enforcement officers and personnel at cellular companies. Id.

Testimony turned to how cellular networks work; specifically, how carriers set up cell sites. Id. at 20. Agent Sonnendecker explained that “it's going to vary based on the cell phone company itself, the technologies that they're trying to deploy in a certain area.” Id. But generally, the companies are “trying to set antennas in an area so that they can provide the appropriate amount of coverage for the demand that exists in that area.” Id. Cell carriers want cell sites to have overlapping coverage, which means being able to access multiple cell sites, to “ensure that you have proper coverage wherever you might be.” Id. Agent Sonnendecker explained that if someone is moving from “the coverage area of one cell site into the coverage area of another one, if there's not overlapping coverage, you could conceivably drop coverage.” Id. at 20-21. So, overlapping coverage is “designed for continuity of the network.” Id. at 21.

Agent Sonnendecker explained that “cell phones are going to stay in their own network” unless there are “roaming agreements between carriers in areas where . . . there [is] not cellular coverage for that particular carrier.” Id. But roaming agreements in the United States are very limited. Id. In an area like Tucson, Agent Sonnendecker would not expect a phone to access a cell site that is not part of its carrier's network because there are enough cell sites. Id. at 21-22.

Counsel asked Agent Sonnendecker to explain how many cell sites would be in a highly populated area as compared to a lower populated area. Id. at 22. He testified that “that's going to depend on the demand” based on “how many devices are trying to access the network at any given point in time in an area.” Id. A densely populated area presumably means more cell phones which will require more cell sites. Id. Agent Sonnendecker would not expect a handset “in downtown Tucson to hit a tower out in . . . far east Tucson or in Vail.” Id. He bases that conclusion on his understanding that a cell site “is designed to cover the area around it.” Id. at 23. There are exceptions in rural areas, but in densely populated areas the cell sites are designed to cover a specific area around the cell site. Id.

Agent Sonnendecker testified that it is not possible “to look at just the data provided by a cellular company and say a hundred percent where a cell phone was located when it made contact with a particular site[.]” Id. As a result, he uses “other information, along with those cell phone records, to determine whether a particular location would be in the area likely serviced by a particular cell site[.]” Id. For example, information about the network development, such as population density and geography. Id.

In the case at hand, Agent Sonnendecker was “asked to look at a large volume of cellular records that were obtained” during the investigation. Id. at 24. In 2015, he created some “visual representations” of the information that he gathered from the records, and he recently updated the visuals. Id. He explained that the updated visuals are “probably a better way to visualize [the] sectors that may help to not lead people to conclusions about potential coverage areas for these antennas.” Id. Some of the visuals are standard twodimensional views where “you're looking straight down at something[.]” Id. at 25. There are also visuals in “calendar view” where the same data gathered from the phone company records is shown in three-dimensional view. Id.

The program used to create the visuals is called GeoTime. Id. at 26. Agent Sonnendecker estimates that he has had 50 hours of training on the GeoTime program. Id. He has used GeoTime thousands of times to do cellular and phone record analysis. Id. at 26-27. He has found the GeoTime program “to be accurate in location analysis and in creating” visuals. Id. at 27. He testified that the GeoTime program is “accepted within the community of people” who do cell phone analysis and records analysis. Id. at 26. He explained that GeoTime “is just a tool to help with the analysis of information,” whether it's the data analytics or visualization of data. Id. The visuals created using GeoTime can be “independently verified outside of the program[.]” Id.

Testimony turned to a PowerPoint created by Agent Sonnendecker. Id. at 27; Gov. Exs. 1 and 2. Defense counsel objected to Slide 2 on the PowerPoint because it “does not correspond to any of the opinions” proffered by the government in its Notice. Id.; see Doc. 2084. Counsel noted that the government just sent him and co-counsel the PowerPoint yesterday evening. 11/15/22 Tr. at 28. Counsel did not know that the PowerPoint would be used at this hearing and neither counsel have had the chance to review it with their expert. Id.

Government counsel represented that the PowerPoint was merely updated and all the slides which relate to the F.D. homicide were sent months ago to the defense. Id. at 29. The only difference in the updated PowerPoint is “that there was a key added” on some slides that make them easier to “understand what dot coordinates” pertain to which person. Id. Also, government counsel anticipates that Agent Sonnendecker “will testify that knowing the movement of phones or being able to see the movement of cell sites helps in [the] analysis of where people ended up later because there was movement towards a certain direction.” Id.

Defense counsel argued that the government's recent disclosure of the updated PowerPoint does not comply with the District Court's Order requiring the government to set forth all of Agent Sonnendecker's opinions. Id. Moreover, the defense does not know which opinion this data related to, and there were no opinions set forth “regarding movement.” Id. at 29-30.

In response to the Court's question of which opinion the PowerPoint relates to, government counsel stated that Agent Sonnendecker can “testify about what cell sites these phones are hitting” because that is “factual information as opposed to an opinion.” Id. at 30. Also, one of the proffered opinions is that phones were hitting cell sites close to the F.D. homicide and were hitting other cell sites further away from the homicide before and after the homicide. Id.

The Court found that it would be helpful to hear what Agent Sonnendecker has to say about what is depicted in the PowerPoint in order for the Court to determine if Agent Sonnendecker is qualified and, more importantly, whether his opinions are reliable. Id. at 31. As a result, the Court overruled the objection and allowed testimony pertaining to the PowerPoint for purposes of the Daubert hearing. Id. at 32. The Court also noted that defense counsel could call their expert to testify on a later date once the expert had a chance to review the PowerPoint. Id. at 34.

Agent Sonnendecker testified that he was provided with the phone numbers relevant to this investigation. 11/15/22 Tr. at 28. Based on the information provided by cell phone providers, he matched the phone numbers with the corresponding providers. Id. at 27-28. Agent Sonnendecker was also provided with the locations of events relevant to the charged offenses. Id. at 37. Slide 2 of the PowerPoint shows data between 4:00 p.m. and 5:00 p.m. on May 15, 2015. Id. at 35. Each colored dot represents a cell site accessed by a particular phone based on the data provided by the phone companies. Id. Agent Sonnendecker explained that each different colored dot “represents a different data set”; for example, orange dots relate to the same data set relating to a specific phone number. Id. at 36. Agent Sonnendecker made clear that he will not testify that a cell phone was located at a particular location because that is not what the data reflects. Id. at 39-40. Rather, all he can say is that a phone number accessed a particular cell site and will provide the location of the cell site. Id. Agent Sonnendecker would not expect a cell phone issued by one phone company to connect to a cell site for another company; for example, a T-Mobile phone would not connect to a Sprint cell site. Id. at 38.

Slide 3 of the PowerPoint pertains to “the time of 5:20 until 5:55 p.m. on May 15, 2015. Id. at 40. During that time frame, there are “42 events across a number of different phones that are captured in this particular visual.” Id. An “event” is a phone call (made or received) or a text message (sent or received). Id. at 39.

Slide 4 pertains to the time frame of 5:59 p.m. to 6:31 p.m. on the same date. Id. at 41. The slide shows cell sites “being accessed south of I-10, closer into the area of the red icon, which ultimately represents the LaQuinta Hotel.” Id. at 41. Specifically, the cell phone numbers assigned to Megan Borges and Ana Rodriguez accessed cell sites that “are in close proximity to where that LaQuinta is physically located.” Id. at 46. The difference in the location of accessed cell sites reflected in slides 3 and 4 show that “those handsets needed to move into the area of coverage for these particular cell sites.” Id. at 43. As a result, he would expect the handsets accessing the cell sites depicted in slide 3 to be moving closer to the cell sites depicted in slide 4. Id.

At this point, the Court interjected and asked, “what kind of area mileage-wise do these cell sites encompass?” Id. at 47. Agent Sonnendecker testified that “[g]enerally speaking, each cell site is designed to cover whatever area the network engineers design it to cover. So I can't say that a particular cell site covers a quarter of a mile or three miles or something to that extent because each one is designed to do something slightly different.” Id. The Court asked whether only the carrier could provide the precise coverage area of a cell site. Id. Agent Sonnendecker testified that “[i]f we wanted to map out the exacts coverage area of a given cell site or a given sector. . .we would need to do a drive scan or try to obtain RF propagation reports from the cell phone companies.” Id. at 47-48. The Court asked Agent Sonnendecker if he could estimate the coverage area for a cell site in a dense urban area. Id. at 48. He testified that in an area such as that depicted in slide 4, “a mile to a mile-and-a-half is an average potential distance that some of these cell sites might cover.” Id.

Government counsel resumed her questioning. Agent Sonnendecker testified that “density and geography is important to understanding how cell sites in a particular area work[.]” Id. As a result, he takes density and geography into account “as a totality of circumstances type consideration in order to make some of [his] opinions[.]” Id. at 49.

Counsel posed the following question: if Borges and Rodriguez had phones belonging to different carriers and they were standing next to each other, “would you expect them to still be hitting off two different cell sites” as reflected in slide 4? Id. Agent Sonnendecker responded that “[t]hey very well could,” but “it all depends on where the actual antennas” for the different carriers were placed. Id. “If they're placed in different locations, then we might see something like we're looking at here. It's possible in some instances, where Sprint and AT&T might have antennas in the same physical place, so you would see what looks like they're co-located.” Id. He explained that some carriers may share a tower and both have antennas on the tower. Id. “[B]ut the networks and the infrastructure are different” because they are “specific to their own network.” Id. Counsel asked if the dots shown in slide 4 which are “a little bit apart from each other” are consistent with two people near each other but who have different carriers. Id. at 50. Agent Sonnendecker testified that “they could have been near each other and accessing different sites from different carriers[.]” Id. “[G]iven how short of a distance these particular cell sites are from that homicide location, Agent Sonnendecker testified that “[i]t is likely” that those cell sites provided service to that location. Id.

Slide 5 shows the time frame of 6:33 p.m. to 7:30 p.m. on May 15, 2015. Id. The slide shows four different data sets (i.e., phone numbers) that were accessing cell sites in this particular area. Id. at 51. The phone numbers hitting the cell sites reflected in this slide are associated with Samuel Rakestraw, Marcell Gray, Ana Rodriguez, and Jermaine Maxwell. Id.

Slide 6 shows the calendar view that was mentioned earlier. Id. Agent Sonnendecker explained that the underlying data is the same as used for slide 5, but slide 6 shows the “data in a three-dimensional mode so that we can get a better context of time and location.” Id. Slide 6 shows that “cell sites were accessed sort of in the general area of where the yellow icon showing” where the “shop” is located. Id. As time moves along, the phones access cell sites to the east which shows “some probable movement of that device into that area.” Id. at 52. Agent Sonnendecker explained that the antennas depicted in all of the slides cover an area in the direction the antenna is pointed. Id.

The Court asked Agent Sonnnendecker if there could be cell sites between the cell sites noted on this particular slide which he did not note. Id. at 53. He testified that “[t]here may be,” but those were not cell sites that were accessed by any of the phones during this particular time frame. Id. Counsel followed up on the Court's question. Agent Sonnendecker testified that for the records that he received in 2015, he would not be able to tell “anything about the handsets' locations for periods of time where there's no activity on the phone[.]” Id. at 54. The records that he obtained, which are referred to as “call detail records,” show cell sites that were accessed for a call or text message. Id. He further testified (confusingly) that “[t]here are a number of other communications occurring between the handset itself and the network where some of those locations could be captured. However, we do not have those records. What we're looking at only are the call detail records.” Id. As a result, “if [a] phone was in an area where typically it would hit other cell sites, if they exist, between those two but no calls were made, we wouldn't know.” Id. at 55.

There were sixteen phone calls between Ana Rodriguez and F.D. between May 12, 2015 and May 15, 2015. Id. at 86. Slide 7 specifically pertains to calls on May 15, 2015, at 9:50 p.m. Id. at 55. This slide shows phone numbers attributed to Rodriguez and F.D. that accessed cell sites, which gives “an idea of this communication between those two parties.” Id.

Slide 8 also shows phone numbers attributed to Rodriguez and F.D. on May 15, 2015. Id. During the nine-minute time frame between 9:59 p.m. and 10:08 p.m., “there's a series of communications back and forth” between these two individuals and “a few other communications that were occurring on those devices with other parties.” Id. at 56. In contrast to slide 7, slide 8 shows Rodriguez's phone number access a cell site that is “in close proximity” to the cell site accessed by F.D.'s phone number. Id. Rodriguez and F.D. had different carriers, “[s]o even if they were very close together, they would still be hitting different cell sites[.]” Id.

Slide 9 shows five sets of records accessing cell sites “in the general area of the shop” between 10:48 p.m. and 11:34 p.m. on May 15, 2015. Id. at 56-57. The phone numbers are attributable to Samuel Rakestraw, Marcell Gray, Jermaine Maxwell, Michael Williams, and Megan Borges. Id. at 57. Slide 10 uses the same data as slide 9, but in the three-dimensional view to provide “a better idea of the volume of interactions that are happening . . . during this time frame.” Id. The phone attributed to Michael Williams accessed the “same cell site around seven times between” 11:00 p.m. and 11:30 p.m. Id.

With respect to slide 11, it shows the phone number attributed to F.D. accessing cell sites in close proximity between 12:15 a.m. and 12:20 a.m. Id. at 58. “It is possible that there could have been overlapping coverage” for those two events, but Agent Sonnendecker cannot definitely tell by looking at this image. Id. In response to counsel's question of whether the activation of the two cell sites five minutes apart could show the movement of F.D.'s phone, he testified: “It could have been.” Id. One cell site (counsel didn't ask which one) is two-tenths of a mile from the LaQuinta Hotel. Id. at 58-59. Counsel asked Agent Sonnendecker if he would expect the LaQuinta Hotel to be within the coverage area of that cell site. Id. at 59. He testified as follows: “It very well could be. Now, that particular sector could provide coverage into that - - into the LaQuinta, or another sector on that could also provide coverage, too.” Id.

Slide 12 pertains to phone numbers attributable to Michael Williams and Samuel Rakestraw. Id. The closest call to the time that Agent Sonnendecker was “given of the homicide report” is an outgoing call from Rakestraw to Williams at 12:11 a.m. Id. Both phones accessed the same cell site which is two-tenths of a mile from the scene of the homicide. Id. Based on the proximity of the cell site to the homicide scene and the direction of the antenna, Agent Sonnendecker believes that this cell site provided service to the area where the homicide occurred. Id. at 59-60.

Slide 14 reflects the time frame of 12:58 a.m. and 1:20 a.m. on May 15, 2015. Id. at 60. Phone numbers attributable to Michael Williams, Samuel Rakestraw, Shawmaine Moore, and Marcell Gray access cell sites “in this general area” within an hour of the homicide. Id. The phone numbers attributed to Williams and Rakestraw “accessed the cell sites in close proximity to the homicide scene during the relevant time frame.” Id. at 60-61. The defendants' phones do not access this same cell site before or after the homicide. Id. at 61.

Slide 15 shows three different cell sites being accessed between 1:49 a.m. and 2:05 a.m. by phone numbers attributed to Michael Williams, Megan Borges, Shawmaine Moore, and Marcell Gray. Id. at 62. Agent Sonnendecker was informed that the F.D. homicide occurred at around 12:35 a.m. on May 15, 2015. Id. at 63. Records for the two phones attributed to David Williams reflect a period of inactivity near the time of the homicide. Records for one of the phones show an incoming voice call from a phone number attributed to Ana Rodriguez at 11:52 p.m. on May 15, 2015. Id. at 63-64. Records for the other phone show an incoming call at 11:38 p.m. on May 15^th and the next event is an outgoing call at 2:05 a.m. on May 16^th. Id. at 64.

Records for a phone attributed to Shawmaine Moore show a period of inactivity around the time of the homicide on May 15, 2015. Id. at 65. There was a call at 11:36 p.m. on May 15^th and the next event is a call at 1:18 a.m. on May 16^th. Id.

Records for a phone attributed to Marcell Gray show a period of inactivity around the time of the homicide. Id. There was an incoming text message from a phone number attributed to Megan Borges at 11:52 p.m. on May 15^th and the next event is another incoming text message from the same phone at 12:40 a.m. on May 16^th. Id.

Records for a phone attributed to Michael Williams show a period of inactivity around the time of the homicide. Id. at 65-66. There is a text message at 11:32 p.m. on May 15^th and the next event is an incoming call from a phone number attributed to Samuel Rakestraw at 12:12 a.m. on May 16^th. Id. at 66. The next event is an outgoing text message at 1:05 a.m. on May 16^th. Id.

Records for a phone attributable to Jermaine Maxwell show a period of inactivity around the time of the homicide. Id. There is an incoming call at 11:19 p.m. on May 15^thand the next event is an outgoing call at 6:13 a.m. on May 16^th. Id.

Records for a phone attributable to Bhishm Neal show a period of inactivity around the time of the homicide. Id. at 66-67. There is a phone call at 3:20 p.m. on May 15^th and the next event is an outgoing call at 2:23 a.m. on May 16^th. Id. at 67.

Records for a phone number attributed to Reginald Johnson show a period of activity around the time of the homicide. Id. There is an outgoing call at 12:28 a.m. on May 16^th and the next event is an incoming call at 12:52 a.m. on the same day.

Records for a phone number attributed to Jermaine Maxwell show a period of inactivity around the time of the homicide. Id. There is a call routed to voice mail at 10:59 p.m. on May 15^th and the next event is an outgoing call at 7:38 a.m. on May 16^th. Id. at 6768.

Records for a phone number attributed to Megan Borges show a period of inactivity around the time of the homicide. Id. at 68. There is an incoming call at 12:13 a.m. on May 16^th and the next event is an incoming call from a phone attributed to Marcell Gray at 1: 51 a.m. on the same day. Id.

Records for a phone number attributed to Samuel Rakestraw show a period of inactivity around the time of the homicide. Id. There is an outgoing call at 12:11 a.m. and the next event is an outgoing call to a phone number associated to Michael Williams at 1:05 a.m. Id. at 68-69.

Agent Sonnendecker did an analysis of phone records for cellular events on May 16, 2015. Id. at 69. Slide 16 shows events between 11:30 a.m. and 12:20 p.m. on that day. Id. There is a location labelled “car wash” on the slide. Id. During this time period, “there were 50 cellular events across these five different data sets that accessed these cell sites in this general area.” Id.

Slide 17 is the calendar view of the data reflected in slide 16. Id. The phones numbers reflected in these slides are attributable to Michael Williams, Shawmaine Moore, Marcell Gray, Megan Borges, and Ana Rodriguez. Id. at 69-70.

Counsel next turned to an animation which Detective Sonnendecker described as “a summary of the images that we've seen to this point relating to the hours leading up to where the victim was killed.” Id. at 70. Agent Sonnendecker stressed that:

when we see movement from one cell site to the next, that's not a pathway that I'm saying that the actual handset went in. What the program is doing is mapping from cell site to cell site in a direct line. So we see the movement of the cell sites. We don't want to interpret that as the actual movement of [where] the handset was. But by looking at this in this way, we can sort of get a better perspective of how movement of these cell sites relates to time.

Id. As the animation was played, Agent Sonnendecker again testified that it shows cell sites that are being accessed by the handsets attributed to various people. Id. at 72-73.

Defense counsel objected to the use of this animation because the defense had never been provided with it, as well as because it purports to show the movement of the phones. Id. at 70-71. The Court overruled the objection because Agent Sonnendecker made clear that he would not testify at trial about the movement of phones, only the activation of cell towers. Id. at 71. To the extent the demonstrative exhibit suggests the movement of phones, that issue can be dealt with prior to trial if Agent Sonnendecker is permitted to testify as an expert witness.

Agent Sonnendecker also did an analysis of phone data related to the shooting of K.G. on June 12, 2015. Id. at 76. He was provided with information that showed that this shooting occurred on June 12, 2015, at approximately 11:41 p.m. around 5403 East 30^thStreet in Tucson. Id. at 76-77. Slide 19 is related to this incident. Id. at 77. The slide shows events between 11:34 p.m. and 11:39 p.m. for phone numbers attributed to Michael Williams and Sherman Shields. Id. Agent Sonnendecker does not know how close the cell site was to the scene of the shooting, but it is “going to be in that two- to three-tenths of a mile.” Id. However, he added that in a dense city area the coverage area varies and could be a mile to a mile-and-a-half. Id. at 78.

The next slide, which is in calendar view, has similar information with an expanded time frame - 11:00 p.m. to 12:25 a.m. Id. The slide shows the cell sites accessed during this time frame from east to west. Id. This visual “just puts dots on the times that [the pertinent phone numbers] accessed those particular cell sites straight from the records themselves.” Id. During this time frame, “there were cell sites that were accessed [by phone numbers attributed to Michael Williams and Sherman Shields] that are in close proximity to where the shooting occurred, and those cell sites likely provide coverage into that area.” Id. at 79.

Agent Sonnendecker also did an analysis of phone records for May 11, 2015, which is when a drive-by shooting occurred. Id. Slide 22 shows cellular events for phone numbers attributed to Michael Williams and Marcell Gray “during the 9 p.m. hour on May 11^th.” Id. There are five cellular events that occurred between 9:00 p.m. and 9:14 p.m. Id. There is a cellular event that accessed a different cell site at 9:24 p.m. Id. at 80. There are eight cellular events that accessed different cell sites between 9:38 p.m. and 9:43 p.m. Id. There are four more events on a cell site to the east between 9:53 p.m. and 9:54 p.m. Id.

Slide 23 also pertains to May 11, 2015 but reflects cellular events occurring during the 10:00 p.m. hour. Id. Agent Sonnendecker was told that the May 11^th shooting occurred at around 10:43 p.m. Id. The events on this slide are “accessing the cell sites closer in time to the shooting” than the previous slide. Id. These cellular events pertain to a phone number attributed to Marcell Gray. Id. This phone number accessed three different cell sites. Id. at 81. The cell site accessed at 10:07 p.m. is “in the general area” of the shooting. Id. The cell sites appear to have overlapping coverage. Id. Given the direction of the antennas and the density and geography of the area where the cell sites are located, it is likely that these cell sites would service the area where the shooting took place. Id.

The next slide pertains to a phone number attributed to Michael Williams and cellular events that occurred the night of May 11, 2015. Id. This phone number accessed “three different cell sites that are in the general area of where this shooting occurred.” Id. at 82. There are five cellular events (which could have been calls or text messages) between 10:20 p.m. and 10:33 p.m. Id. There is another cellular event at 10:35 p.m. that accesses a different cell site. Id. And there is a different cell site accessed at 10:48 p.m. by this phone number. Id. The phone numbers attributed to Marcell Gray and Michael Williams are “hitting on the same cell sites during this time frame[.]” Id.

Slide 25 pertains to the time frame after the shooting, 10:43 p.m. until almost midnight. Id. The phone numbers attributed to Gray and Williams are not “still accessing the same cell sites that they were accessing” right before the shooting. Id. at 82-83. These new cell sites, which start being accessed at around 10:59 p.m., are about five miles away from the location of the shooting. Id. at 83. Agent Sonnendecker would not expect these later cellular events to have accessed the cell sites near the shooting. Id.

Slide 26 is a calendar view of the same data detailed in slide 25. Id. The slide shows cell sites accessed from 9:00 p.m. until 12:00 a.m. Id. There are cell sites accessed in the area of the shooting and then cell sites further to the west which were accessed after the reported time of the shooting. Id. at 84. Agent Sonnendecker used “directional trails” to help follow “pattern of movement based on this into the general area of where the shooting occurred and then away from it afterwards, based on cell site movement.” Id.

Slide 28 is an animation of the data from slides 25 and 26 which shows “the pattern of cell sites that were accessed” for the phone numbers attributed to Michael Williams and Marcell Gray. Id. On the left-hand side of the animation, “the time line slider” gives perspective on time, and the icons help see the cell sites that are being accessed during points in time. Id. at 85.

Agent Sonnendecker did an analysis of a phone number attributed to Bhishm Neal. Id. at 85. The purpose of the analysis was to identify cell sites accessed by that phone number around the time that Neal was “encountered by an officer on October 9^th of 2015 at 3125 North Alvernon Way[.]” Id. at 85-86. “[T]here were two different cell sites that were accessed during this time frame in the general area of where that scene was located.” Id. at 86.

Agent Sonnendecker was “also asked to do an analysis of where cell phones, including the cell phones of [Rodriguez], David Williams, Mr. Johnson, Ms. Monteen, Mr. Maxwell, Mr. Neal, and Mr. Moore, were on May 13^th of 2015[.]” Id. He testified that the phone numbers attributed to those people accessed cell sites “in the same general proximity” of each other and a WHB shop “at the time of interest when [Rodriguez] made a phone call to Mr. Davis[.]” Id. at 87.

Cross-Examination by Mr. Payson:

Agent Sonnendecker is not an engineer or a scientist. Id. at 89. He is familiar with the software used for drive testing, but not “the actual coding of how it is all pieced together to create the software[.]” Id.

Generally speaking, the goal of the phone carriers is to have “two or maybe even three cell sites that might be able to pick up a call from any location[.]” However, it depends on the area. Id. But Agent Sonnendecker agreed that in an urban area, the more cell sites the better. Id.

With respect to the GeoTime program, Agent Sonnendecker testified that he could take all the data input into the program and manually create exhibits. Id. at 90. However, “[i]t would be extraordinarily voluminous” and would “[t]ake forever in a case like this[.]” Id. Agent Sonnendecker manually checked some of the GeoTime data. Id. He explained that he will do “spot checks of the underlying data.” Id. He did not memorialize the spot checks he performed in this case. Id. He also manually checked some of the exhibits used in the hearing. Id.

With respect to the animation, Agent Sonnendecker again made clear that moving icons do not represent “the movement of the actual underlying handset.” Id. at 91. He testified that “[w]e can't surmise that, hey, this person, whoever had this phone, traveled in this same pathway.” Id. Agent Sonnendecker disagreed with counsel that the animation shows the movement of phones. Id. at 91-92. He testified that “[w]hat we're seeing are known locations, which are the cell sites.” Id. at 92. He again emphasized that he is not “trying to estimate the location of any cell phone at any point in time. What we can tell you are the locations of the cell sites from the underlying records.” Id. Each dot that appears in the animation “represents a cellular event that accessed a cell site.” Id. at 94. Agent Sonnendecker estimates that there are about 100 to 150 cellular events illustrated in the animation. Id. He believes there were about eight or nine cell phones in the animation. Id.

Agent Sonnendecker is familiar with the acronym CSLI, which stands for cell site location information. Id. at 95. He agreed with counsel that “it's not cell phone location information,” and that is “an important distinction.” Id. He agreed with counsel that his opinions are based on call detail reports generated by the cell phone carriers. Id. He also agreed with counsel that the intent of the carriers “in generating this information is not that this information be used in court[.]” Id. at 95-96. Their intent is to use the information for billing purposes and/or to evaluate network efficiency. Id. at 96.

Agent Sonnendecker agreed that “[t]here's not an assumption that the handset would always go to the closest cell site.” Id. at 97. Rather, the general principle is that the “call will go to the strongest tower signal[.]” Id. Agent Sonnendecker testified that his understanding is that “the handset chooses the signal that has the best quality at that moment in time.” Id.

The Court interrupted with a series of questions relating to which cell site a phone will most likely connect to. Id. at 99. Agent Sonnendecker made clear (for the Court) that “a cell phone, by default or design, connects to the cell tower with the strongest signal, as opposed to proximity.” Id. at 99. The Court asked if Agent Sonnendecker determined whether there were other cell sites for a specific carrier in the area where a cell site for the same carrier was accessed by one of the phones already discussed. Id. What the Court was getting at with its imprecise question is whether it was possible that the cell site that was accessed merely had a stronger signal than another cell site that may have been closer to the handset. Agent Sonnendecker did not do that. Specifically, he testified that:

[t]he underlying records identify the cell site that was used for a call or a text message. So I don't have to know anything more than what the phone company is telling us about the location or which cell site it was if we're looking to just do that, just figure out what was the cell site that was accessed for a given call. If there is other information that we need to kind of dig down deeper in, like maybe try to figure out, hey, what is the possible coverage area of this particular cell site, we could look at other surrounding cell sites in an effort to try to get a general idea.

Id.

The Court pointed out that the defense's concern seems to be that “if a cell phone is picking up the strongest signal rather than [the closest in] proximity, then that kind of undercuts, to some extent, the ability to pinpoint a location of a cell phone.” Id. at 99-100. Agent Sonnendecker testified that all the images that he created are based on “what the cell phone company said has occurred.” Id. at 100. The Court asked if there was a way to determine the coverage area of the cell site, in terms of feet or miles, as well as direction. Id. Agent Sonnendecker explained that “[t]he antennas that are assigned to these different cell sites are all designed to cover different areas. Some of those could be a small range; some of it could be a larger range. It's all contingent on a number of factors within the cellular network itself.” Id. In response to the Court's question of whether “only the cell phone company would be able to tell us the coverage area for any specific tower,” Agent Sonnendecker testified: “That or by usage of specialized tools, a drive scanner,” which was not done in this case. Id. at 101.

Before resuming cross-examination after the lunch break, defense counsel noted that the defense did receive the animation in July 2022; as such, counsel withdrew their objection based on late disclosure (but not as to the admissibility of that exhibit). Id. at 103. Agent Sonnendecker does not have statistics about how often a cell call will not go to the closest tower. Id. at 104. He is unaware of a “peer-reviewed article pertaining to how often a call will go to the closest tower[.]” Id. Agent Sonnendecker again testified that cell phones are designed to search for the closest tower. Id. In response to counsel's question of whether “companies design software for their phones to be able to do this,” he testified that “[i]t's a combination of things within hardware and software.” Id. Agent Sonnendecker agreed with counsel that phone companies “design and use proprietary algorithms,” he is not familiar with how they work, and phone companies “don't let other people know what they are[.]” Id. at 105. Agent Sonnendecker also agreed that the “original intent of design” for cell sites is not for tracking purposes. Id.

Cell phones generally are going to connect to cell sites in their own network. Id. Some towers have one network and others have multiple networks. Id. at 106. Counsel turned to the tower listed within opinion eight in the government's Notice. Id. Agent Sonnendecker agreed that he testified on direct that this tower is two-tenths of a mile from the LaQuinta. Id. He has never gone to look at that tower; he looked at it on the GeoTime mapping tool. Id. He does not know if that tower is “the same now as it was back in 2015 when these CDRs were generated[.]” Id. He agreed that repairs could have been made since then as well as “additions.” Id. at 106-107. He does not know the height of this tower. Id. at 107. He agreed that “the height of the tower is one of the factors that will affect its range[.]” Id. He does not know how many cell phone companies use that tower. Id. Different companies would use different antennas. Id. In terms of the topography of the area around this tower, Agent Sonnendecker testified that it is “relatively flat,” but “[s]ome areas are rolling.” Id. There are some higher buildings in that area “but not skyscrapers[.]” Id. He does not know the height of those buildings because “he does not know the ins and outs of that area[.]” Id. at 108. He does not know the angle that “the antennae on that tower [are] tilted at[.]” Id. He agreed that the angle of the antenna would affect the range; but he added that this “information is not normally provided in the cell site keys.” Id. He also agreed that he previously testified that he does not know the range of this particular tower and that the range could be a mile-and-a-half. Id. He further agreed that it is “feasible” that the range could be longer. Id. He testified that “[t]here will be a number of other cell sites” around that tower, but he does not know how many. Id. at 109. He does not know the strength of the signal, or how the strength is measured (e.g., amps). Id.

Counsel turned to the PowerPoint (Gov. Ex. 2). On page two, Agent Sonnendecker placed blue outlines around groupings of cell sites because they were all in the same general area. Id. at 110. Agent Sonnendecker testified that “during this one-hour time frame, the cell sites that were being accessed by the phones within that blue circle . . . is where the cell sites were.” Id. at 111. The cell sites outside of the blue circle which were being accessed “were not exhibiting a similar pattern.” Id.

Counsel pointed Agent Sonnendecker to two purple dots that purportedly represent a phone number attributed to Mr. Rakestraw accessing a cell site. Id. at 112. The dots mean that there were two cellular events, meaning that two calls either originated from or were received by that phone number. Id. at 112-113. If the person with that phone called someone in the area depicted on the exhibit, then that event would be depicted if the phone accessed that particular cell site. Id. at 113. Agent Sonnendecker agreed with counsel that he “can't say necessarily that the device that ends in 1247 [(the last four digits of the number attributed to Rakestraw)] was even in that area at that time.” Id. Because the Court was confused by this statement, it asked the following question: “Let's say someone called Mr. Rakestraw[,]” are you saying his “number could show up on that cell tower even if he was nowhere in the area[?]” Id. at 115. Agent Sonnendecker clarified that if there was an incoming call to Rakestraw's phone and the call accessed this cell site, “the phone would need to be in the area of coverage for that particular cell site in order for that to happen.” Id. at 115-116. So, regardless of whether a call was incoming or outgoing, the phone “would have to be within the coverage area of that site.” Id. at 116.

The azimuth is the direction of the sectors covered by an antenna. Id. at 118. It varies with the carriers in terms of whether that information comes with the CDRs. Id. For example, MetroPCS provided “the latitude-longitude coordinates of the site, but they weren't clearly defining what the azimuth was.” Id. By contrast, AT&T and T-Mobile provide the latitude-longitude coordinates as well as “the orientation or the azimuth for most of the events that occurred.” Id. Agent Sonnendecker agreed with counsel that “cell phone companies carry a separate list of the directions of the site of their antennae.” Id. at 119. They are called “cell site keys.” Id. Some companies provided cell site keys and other times Agent Sonnendecker had to specifically request them. Id.

Counsel turned to page eleven of the PowerPoint which “indicates that at 12:15 a.m. there was a call to the red phone” (which has a number attributed to F.D.) which accessed the cell tower near the LaQuinta. Id. Agent Sonnendecker agreed that he previously testified that it is possible that this call originated at the LaQuinta “[e]ven though it's outside of this sector that [he] delineated.” Id. Agent Sonnendecker explained “it's feasible that an area called the rear lobe of this antenna could have conceivably carried that. We don't, obviously, know where that handset was at that point in time other than to know” that this cell site was accessed at that point in time. Id. at 120. He agreed with counsel that the range of this tower “isn't even actually limited to being inside of these lines that [he's] drawn.” Id. Agent Sonnendecker does not have any information about where F.D. was when he might have made this call. Id.

Five minutes later at 12:20 a.m. there is an incoming text message to the same phone. Id. The shaded area on the PowerPoint page does not indicate the range of the cell site, but rather the direction the antenna is pointed. Id. at 121. Agent Sonnendecker explained that the range is not noted because “[i]t might lead people to think that we're trying to estimate coverage area by showing a shaded area all the way out.” Id. Agent Sonnendecker disagreed with counsel that the shaded area “would suggest to the jury that that's a coverage area[.]” Id. “All it's trying to do is show a direction[.]” Id. However, Agent Sonnendecker agreed that the direction could be shown with an arrow or testimony, and not a shaded area. Id. at 121-122.

Counsel turned to page 13 of the PowerPoint. Id. at 122. Agent Sonnendecker created this image yesterday, “[b]ut the underlying data is the same.” Id. He used a distance measurement tool within the GeoTime program to create this image. Id. Agent Sonnendecker has compared this program against the Google Earth measurement tool and they have been consistent. Id. at 122-123. However, he did not use Google Earth for this image. Id. at 123. He also did not “go out to the scene.” Id.

Agent Sonnendecker testified that he “recently obtained a drive scanner, and we've done a number of scans” in the past year-and-a-half to two years. Id. at 125. He agreed with counsel that drive-scan testing “is the most accurate measure of a tower's coverage.” Id. at 127. He did not do drive-scan testing of the towers in this case. Id. at 128. He agreed that “[a]t this point, seven years later, it wouldn't give [him] valid results anyway[.]” Id.

The furthest “lobe range” - i.e., a cell site connecting to a phone which is not within the direction an antenna is pointed - is “a couple hundred feet.” Id. at 124-125. With respect to a call within the area an antenna is pointed, Agent Sonnendecker agreed that it is feasible that the phone could be further than a mile-and-half from the cell site. Id. at 125-126.

Agent Sonnendecker agreed that GPS technology would be the most accurate method to determine the location of a phone. Id. at 127. However, he testified that “[w]e don't generally receive GPS information from phone companies.” Id. The next most accurate data for locating a phone would be specialized location records. Id. He agreed with counsel that “CSLI is the least accurate of these methods[.]” Id.

Counsel turned to a “controversy with CSLI in Denmark.” Id. at 129. Agent Sonnendecker was not familiar with this controversy until he read the article provided by the defense this morning. Id. He agreed that the article notes that “there was faulty software with the CSLI being used in Denmark for a long period of time,” which resulted in thousands of cases that had inaccurate data. Id. at 130. Although Agent Sonnendecker does not know what the problem was in Denmark, he does not believe that “there's a problem with any of the records that we've been provided in this case.” Id.

Counsel asked Agent Sonnendecker whether the opinions set forth in the government's Notice are the only ones that he will provide at trial. Id. at 130-131. He testified that “the opinions that I would render relate to the movement of particular devices or possibly if a cell phone could be in a given area of coverage of a particular cell site.” Id. at 131. He agreed with counsel that he “intend[s] to testify about movement even though that's not mentioned” in the proffered opinions. Id. Agent Sonnendecker conferred with prosecutors to help draft the Notice and provided his input. Id. The prosecutors told him that they needed to detail all of his opinions. Id. at 132. However, he added that “this process is very different than any other I've been a part of, as far as laying out every single opinion that could conceivably be discussed.” Id. at 131. He also added that things change and “we're evolving with how we're going[.]” Id. at 132.

Agent Sonnendecker agreed that the first proffered opinion says that the CDRs indicate that the cell phones accessed cell sites whose coverage area encompassed a specific location. Id. He also agreed that opinion eight uses different language, specifically that “the CDRs indicate that both phones were in proximity of a cell site.” Id. In response to counsel's question of whether opinion one uses more accurate language, Agent Sonnendecker testified that although the language used is different, the opinions are “saying ultimately the same thing.” Id. at 133. However, he ultimately agreed with counsel that “the language in opinion 1 is more accurate and better reflects what [he's] trying to tell the jury[.]” Id. He also agreed that “[t]he phones might not necessarily be actually very close to the cell tower that they access.” Id. at 135. It's possible that the phones could be “a mile-and-a-half away or even more[.]” Id.

Cross-Examination by Mr. Flores:

Counsel turned back to opinion one and asked if Agent Sonnendecker was going to testify that the phones attributed to David Williams and Megan Borges were “near the same area and accessed cell sites whose coverage area may have encompassed 1800 South Pantano Road.” Id. at 136. Agent Sonnendecker testified that his opinion is that “these two devices accessed cell sites” in the “general area encompassing this 1800 Pantano.” Id. He disagreed with counsel that his testimony would discuss the location of phones. Id. at 137. He testified that he “can't sit here and say. . .both of those phones were at this address.” Id. Agent Sonnendecker agreed with counsel that he “can't say they're in the same area either[.]” Id. What he can say is that “based on the cell sites that were accessed around a similar time frame, the underlying handsets would need to be in that general area in order to access those cell sites in those sectors.” Id. Counsel asked the basis for that opinion given that Agent Sonnendecker does not know the strength of that tower. Id. Agent Sonnendecker testified as follows: “I can't determine how far away the handset is with these records. What I can tell you is the cell site and likely the antenna that it spoke with at a certain point in time.” Id. at 137-138.

Counsel attempted to break down this opinion because it likely applies “to all opinions regarding cell phone coverage and area and location.” Id. at 138. Agent Sonnendecker agreed that David Williams' phone “could be a hundred feet from that tower[.]” He also agreed that Borges' phone “could be three miles away from that tower.” Id. Agent Sonnendecker agreed with counsel that the phrases “in the area'” or “coverage area” are undefinable because he does not know the coverage area of that tower. Id. at 138-139. As a result, without an understanding of “the engineering of that particular tower,” it is feasible that the phones could be more than three miles from the tower. Id. at 139. In addition to understanding the underlying engineering, it is important to know the topography of the area (e.g., mountains, hills, buildings). Id. Density of the population of an area is important as well. Id. Agent Sonnendecker agreed with counsel that “without all those variables defined, we can't identify or we cannot say, with any scientific certainty, what ‘coverage area' means[.]” Id. Counsel returned to opinion one and asked Agent Sonnendecker if it was fair to say that he would be suggesting to the jury that “those phones are somehow near each other” in the area of 1800 South Pantano. Id. at 140. He testified as follows: “My testimony will not be that. It will be that these phones accessed this cell site and this antenna at this point in time.” Id.

Counsel turned to opinion eight, which is that “an analysis of the CDRs indicate that both phones were in the proximity of cell sites near the LaQuinta Hotel at 7001 South Tucson Boulevard.” Id. In response to counsel's question of whether that “opinion suggests that those phones were near the LaQuinta,” Agent Sonnendecker testified: “I can see your interpretation of that. What I would say is that both of those phones accessed that cell site and sector during that time frame, which provides likely coverage into the LaQuinta Hotel.” Id. He is “not going to say that both phones were in a direct proximity of the LaQuinta,” but rather that the phones “accessed the cell sites and the sectors that provided likely coverage into that area.” Id. at 140-141. Agent Sonnendecker agreed with counsel that for consistency it was reasonable to “substitute ‘proximity' for ‘coverage area.'” Id. at 141. Agent Sonnendecker again testified that he “will not be defining any coverage areas.” Id. Once again, he agreed with counsel that it is feasible that any one of the phones could have been “a hundred feet from the tower or miles from the tower.” Id.

Counsel asked Agent Sonnendecker for the foundation for his opinions given that there is no scientific definition of “coverage area” or “proximity” for the cell sites “and all of the different plottings” that he has done in this case. Id. at 142. He testified as follows:

So I think it goes like this: Use the LaQuinta, for example. For this cell site that was accessed, what I'd say is the cell site was accessed at this point, at this date and time, and provides probable coverage into the area of where that LaQuinta is. And I base that upon the hundreds of live time operations that I've conducted, live time tracking, analysis of records, drive scans and understanding of how the networks work.

Id. He agreed with counsel that his opinions are “almost entirely based on [his] experience.” Id.

Counsel turned to the foundation for his experience. Id. Agent Sonnendecker cannot make sure that the phone companies have kept accurate records. Id. at 143. Rather, he relies “on them to provide the records” to him, “along with their business certifications.” Id. He has not asked the phone companies to provide data or their methodology to ensure that the cell site information is accurate. Id. He does not know if the phone companies have ever published that type of material that would give him reason to believe their records are accurate. Id. at 144. He testified that he “know[s] about the accuracy of their records from the years and years of experience using them.” Id. He agreed that he is partially concluding that the records are accurate in this case because they have been accurate in other cases. Id.

Counsel turned to the article on the problems with cell site information in Denmark. Id. Agent Sonnendecker agreed that the article calls “CSLI evidence junk science.” Id. He had not seen this article even though he tries to keep up to date with articles, books, and trainings related to CSLI. Id. He also did not read the study referenced in the article which found that there was a 40 percent error rate between the places where cell towers were supposedly located and where towers were actually located. Id. at 146. Agent Sonnendecker testified that in 2015-2016, he used Google Earth to verify that the location of some of the towers was consistent with the records provided by phone companies. Id. at 147. But he does not recall which towers he looked at or how many of them. Id. Agent Sonnendecker cannot even estimate how many cell towers were referenced in the records provided to him for this case. Id. at 148-149. He did not “plot every single one of them on Google Earth[.]” Id. at 149.

Agent Sonnendecker again testified that a cell phone is designed to connect to a cell site that has the strongest signal. Id. at 150. As a result, he agreed with counsel that when he says that a cell phone is “in the coverage area, that doesn't necessarily mean that that tower. . .is the nearest tower.” Id. He again agreed that it is the signal strength of a tower, and not the proximity of a phone to a tower, that determines which tower a cell phone accesses. Id. at 151. He further agreed with counsel that his testimony “will never suggest to the jury” the phones attributed to Michael Williams and Samuel Rakestraw were near the LaQuinta Hotel. Id. What he will say is that “the handset accessed a cell site, in this instance, two-tenths of a mile away [from the LaQuinta] at this point in time, and that area is a probable coverage area for that particular cell site.” Id. With respect to the “probable coverage area,” Agent Sonnendecker agreed that coverage area is not definable and that the tower could be miles and miles away from the handset. Id.

Counsel asked whether the fact that a handset connects to another cell tower suggests that the phone has moved locations. Id. at 155. Agent Sonnendecker testified that it depends on a lot of factors, such as signal strength, topography, and density. Id. However, Agent Sonnendecker testified that a cell phone accessing different cell towers suggests that the handset has moved into another tower's coverage area. Id. His opinion is based on his “understanding of how these networks are built and how the coverage areas of different antennas are designed.” Id. at 156. But he agreed that towers are designed to cover different ranges, and he does not have information about the particular engineering for each tower in this case. Id. at 156-157. When asked if that information would make his opinion more reliable, Agent Sonnendecker testified: “If I had the propagation, sure, that's always helpful.” Id. at 157. “[T]he methodology that these engineers use to uniquely set up each cell tower” was not provided to Agent Sonnendecker. Id. Agent Sonnendecker agreed with counsel that “[i]t's possible” the tower near the LaQuinta Hotel could “have a range of ten miles,” while “another tower near north Tucson has a range of one mile[.]” Id. at 158. When asked if it was possible that if Michael Williams could have been ten miles from the LaQuinta but his cell phone accessed a tower near that hotel, Agent Sonnendecker testified that “it's an extreme example.” Id.

Agent Sonnendecker agreed that the area around the Tucson Airport has dense housing north of the airport, but “it's pretty desolate land” to the south and east of the airport. Id. at 161. Both the population density and the topography impacts signal strength. Id. The best way to definitively know a cell site's range is to do a drive scan test. Id. at 162. That is currently impossible to do now for cell towers used in 2015 because many of these networks have been retired or towers have been upgraded with better equipment, which will affect the range. Id.

Counsel returned to opinion eight in the government's Notice. Id. Agent Sonnendecker agreed that he testified on direct examination that on May 16, 2015, there was a cellular event 23 minutes before the homicide involving phone numbers attributed to Michael Williams and Samuel Rakestraw that accessed the tower near the LaQuinta. Id. at 162-163. Agent Sonnendecker agreed that in 23 minutes a person “can get pretty far away from the airport[.]” Id. at 164. Nevertheless, Agent Sonnendecker maintains that this event occurred right before the homicide. Id. at 165. He agreed that there is no information to suggest that these phones were stationary and could have merely been in a vehicle driving near that tower. Id. at 164. There is also no information to suggest that the people using the phones were together. Id. at 165.

Counsel turned his questioning to two cell towers accessed by a phone number attributed to Michael Williams. Id. at 167. Agent Sonnendecker did not analyze how many times Williams' phone accessed these two towers. Id. at 169. One tower is at the intersection of Golf Links and Wilmot and the other is near the intersection of Golf Links and Kolb Road. Id. at 166. Defense exhibit 22 shows a cell tower located at 6334 East Golf Links Road. Id. at 168. Michael Williams spent a considerable amount of time at his girlfriend's home located near that tower. Id. For a three-month period, Williams' phone accessed this tower 194 times. Id. at 169. For that same period of time, Williams' phone accessed a tower further from that home 461 times. Id. at 170. Agent Sonnendecker testified that if we assume that all cellular events occurred when the phone was at the girlfriend's home, then this is an “example of how proximity to a cell tower means nothing[.]” Id. at 170-171.

Agent Sonnendecker has never had his methods peer-reviewed in scientific journals. Id. at 173. He explained that as a DOJ employee he cannot release records from a criminal case to peer-reviewed journals. Id. When asked if anyone has ever looked at his methodology used to render opinions, Agent Sonnendecker testified that every time he creates a report, it is turned over to the defense who may have their own experts. Id. at 173-174.

Counsel turned to the movement of phones attributed to Marcell Gray and Michael Williams on May 11, 2015. Id. at 174. Agent Sonnendecker disagreed with counsel's characterization that the animation that he created suggested not only that the phones were near the cell towers but traveled in tandem to the shooting site. Id. at 175. Agent Sonnendecker agreed that he has no idea of whether Williams may have been on the phone at his girlfriend's house and the call was diverted to a cell tower further away. Id. As a result, he does not know where the phone was when it connected to the cell tower; the phone could have been “five miles away, ten miles away, five feet away.” Id. at 175-176. Agent Sonnendecker testified that his animation is not suggesting that the phones are moving; but he then added that “in order for these phones to be up in the north and then move over into the area of where the shooting occurred and then move west, the handset had to have moved.” Id. at 176. Agent Sonnendecker agreed that the “animation is pretty linear, straight from point A to point B.” Id. at 177. He also agreed that he has “no way of verifying whether that mode of travel is even close to what actually happened on May 11^th. Id. More specifically, he cannot tell “what path was taken for the phone to travel from point A to point B.” Id. But he believes the animation is merely reflecting cell sites that were accessed at various points in time in different areas, and not a path of travel. Id.

Agent Sonnendecker used the GeoTime program to create the animation. Id. at 177178. Agent Sonnecker testified that “there's validation efforts that happen on the software end by the company.” Id. at 178. However, he does not “know what their validation efforts were.” Id. at 179.

Because of time constraints, the redirect examination and argument were scheduled for December 14, 2022.

Redirect Examination by Government Counsel:

Prior to redirect examination, defense counsel objected to the use of new exhibits just handed to them this morning. 12/14/22 Tr. at 5. Counsel noted that they had not had the chance to review the exhibits, let alone have their expert review them. Id. Counsel argued that the government was trying to expand Agent Sonnendecker's testimony, which is improper on redirect examination. Id.

Government counsel represented that the exhibits were disclosed on November 18, 2022 and were created “in direct response to defense's questioning.” Id. The government pointed out that the slides change the colors of phones to make the exhibit clearer and adds towers. This is not new information because “the defense has had disclosure of cell site latitude and longitude of towers.” Id.

Defense counsel pointed out that the government admits that the exhibits were created in response to cross-examination; in fact, they were provided three days after Agent Sonnendecker's testimony. Id. at 6. Counsel pointed out that the Daubert hearing has been pending for some time and there has been “four years to get this stuff nailed down.” Id. The government is essentially giving the defense “moving targets” in an effort to improve upon Agent Sonnendecker's ever-changing opinions. Id.

Government counsel responded that the defense has had this information “forever” and redirect “is always to challenge what came up on cross-examination.” Id. at 7. Counsel asserted that the visuals are not different than if a map of Tucson were put on the screen and Agent Sonnendecker “put Xs on where every tower is in Tucson.” Id.

The Court noted that there is a difference between the defense having the raw data and having it summarized in these exhibits. Id. at 8. And the exhibits were created after Agent Sonnendecker's cross-examination. Id. For these reasons, the Court precluded the use of the exhibits on redirect examination. Id. at 13.

Agent Sonnendecker again testified that he cannot write peer-reviewed articles as an agent working for the DOJ. Id. at 14. He has done a number of independent tests in other investigations using cellular records. Id. at 15. He currently does drive scan testing. Id.

The study out of Denmark did not cause any concerns about his analysis in the case at hand. Id. at 16. In that study, “there were some issues with the records of that particular cell phone provider.” Id. In this case, he did “validation across the various sets of records” and did not “see anything that was out of order.” Id.

With respect to the incident on May 11, 2015, Agent Sonnendecker testified that the phone numbers attributed to Marcell Gray and Michael Williams “were closer to the incident location when they hit the towers near the incident location than they were earlier[.]” Id. at 26. He explained that in order to access the cell sites near the incident location, “the handset needs to be in the area of coverage for that particular antenna.” Id. Thus, his opinion is that the handsets were closer to the incident location at the times of those calls near the incident than they were beforehand[.]” Id. at 27. Also, it is his opinion that the handsets accessed cell sites closer to the incident than they did afterwards[.]” Id. In order to do that, the user of the cell phone had to be closer to the incident location than they were when they were hitting towers later on in the evening[.]” Id. Agent Sonnendecker again explained that to “access the cell sites in the general area of where the crime” occurred, the “handset needs to be in that area of coverage for those particular cell sites.” Id. at 28.

In response to a question from the Court, Agent Sonnendecker testified that the cell phone companies provided records regarding the orientation of their antennas in 2015. Id. He explained that the records detailed the azimuth (which is essentially the direction) of the particular antenna. Id. at 29. For example, the exhibit that counsel was questioning him about shows a 120-degree azimuth, which is a sector of coverage. Id. Agent Sonnendecker agreed with the Court that we know the width the antenna covers, but we don't know how far out the coverage area extends within a given sector because the phone companies did not provide those records. Id. at 29-30.

In response to counsel's follow-up question, Agent Sonnendecker testified that the density of towers in an area can change the coverage area of a tower. Id. at 30. Counsel asked Agent Sonnendecker to explain how coverage area works. He testified as follows:

So, generally speaking, you know, a given antenna on a cell site is designed to cover a certain area, and that area is whatever is needed as defined by the network engineers when they're building this. In an urban area, we're going to have more cell sites, and those cell sites are designed to cover less of a distance. In a rural area, you're going to have less cell sites that are designed to cover a further or more of a distance. You know, the records don't tell us definitively what the coverage of each antenna actually is.

Id.

Based on his training and experience and his understanding of how the networks operate, Agent Sonnendecker testified that it is “highly unlikely” that the handsets stayed in the area where they were accessing the towers originally when they later accessed cell sites to the south. Id. at 31. He explained that the antennas to the north were facing an opposite direction than the antennas to the south. Id. Agent Sonnendecker does not know the distance between these two cell sites, but he estimates that it is “multiple miles.” Id.

Counsel turned back to Agent Sonnendecker's review of phone records pertaining to May 16, 2015, the day of the F.D. homicide. Id. at 32. Based on his review of records of a phone number attributed to Michael Williams, Agent Sonnendecker's opinion is that the handset “would have had to be closer to the incident site at the time of this last phone call before the homicide than he was earlier in the evening[.]” Id. at 33-34. He explained that the handset “would need to be closer in order to access this particular antenna” because it is faced in an easterly direction. Id. at 34. Earlier in the day the phone was accessing cell sites much further north. Id. “So in order for this particular antenna to be accessed, the handset needed to be down into the area of this coverage.” Id. After the homicide, this phone was accessing cell sites much further north than it was at the time of the homicide. Id. In order to access the cell site which is at about four-and-a-half miles to the north, “the handset is going to need to be up in the coverage area of those particular antennas.” Id. at 35. And those antennas would have a different coverage area than the antenna near the homicide scene. Id.

With respect to the phone number attributed to Mr. Rakestraw, Agent Sonnendecker's opinion is that the handset was closer to the scene of the homicide than it was before and after the homicide. Id. He again explained that cell sites that were accessed before and after the homicide were four-and-a-half miles to the north of the cell site near the crime scene. Id. at 36. As a result, the handset would have had to have been in the coverage area of the antennas to the north. Id.

The records for the phone numbers attributed to Mr. Rakestraw and Michael Williams span from March 1, 2015 to June 25, 2015. Id. at 37-39. There were 21,188 events for the phone attributed to Williams and 12,724 events for the phone attributed to Rakestraw. Id. at 40-41. During this time frame, each of these phone numbers only accessed the cell site near the homicide scene one time, which was around the time of the homicide on May 16, 2015. Id. at 37-40. However, Agent Sonnendecker clarified that the phone number attributed to Williams did access the same cell site a couple days later on May 20, 2015. Id. at 40. But that phone number never accessed the cell site between March 1, 2015, and the day of the homicide. Id. Also, the phone number attributed to Rakestraw accessed the cell site near the homicide scene on March 16, 2015, “but it was an antenna oriented a different direction.” Id. at 41.

The Court's Examination:

Agent Sonnendecker again explained that drive scan testing involves driving around with “a radio scanner and some other pieces of equipment hooked to a computer and antennas[.]” Id. at 42. Two pieces of information are being captured simultaneously - the GPS location of where Agent Sonnendecker is at a given point in time and “what actual antennas are being communicated with at that moment in time.” Id. The result is that “you can look to see what antennas provide coverage to a certain area that you were in.” Id. Agent Sonnendecker did not do drive scan testing for the towers at issue in this case. Id. at 42-43. The technology existed to do drive scan testing in 2015, but ATF “did not have this equipment at that time frame.” Id. at 43.

The Court asked Agent Sonnendecker if the phone company would have records of the coverage areas of the cell sites as they existed in 2015. Id. He testified as follows: “So I would believe that they do. However, I can tell you that in my tenure with ATF and doing this, I have never on one occasion been able to see those records.” Id. The Court asked if Agent Sonnendecker has ever requested those records and not been provided with them. Id. He testified that “[t]here's been a couple of occasions where we have requested and not received them.” Id. at 44. He added that “it's sort of something that we don't even ask for anymore because we don't believe it will be provided.” Id. Finally, the Court asked if there have been cases where the phone company representative testifies about the coverage area of their cell sites, especially if the phone companies will not provide that information. Id. Agent Sonnendecker testified that “[i]n my experience, I never have had someone from the phone company come testify to that information.” Id.

DISCUSSION

Federal Rule of Evidence 702 allows “a witness who is qualified as an expert by knowledge, skill, experience, training, or education” to provide opinion testimony regarding scientific, technical, or other specialized subject matters if: “the testimony is based on sufficient facts or data; the testimony is the product of reliable principles and methods; and the expert has reliably applied the principles and methods to the facts of the case.” The Ninth Circuit has held that the “admissibility of expert opinion testimony generally turns on the following preliminary questions of law determinations” by the trial judge: (1) whether the expert has appropriate qualifications - i.e., some special knowledge, skill, experience, training, or education on that subject matter; (2) whether the opinion is based on scientific, technical, or other specialized knowledge; (3) whether the testimony is relevant and reliable; (4) whether the expert's opinion would assist the trier of fact in understanding the evidence or determining a fact in issue; and (5) whether the probative value of the expert testimony is substantially outweighed by the risk of unfair prejudice, confusion of issues, or undue consumption of time. United States v. Hankey, 203 F.3d 1160, 1168 (9^th Cir. 2000).

Before turning to the analysis of the factors that the Court must consider to determine whether Agent Sonnendecker may provide expert opinion testimony, the Court must address two threshold issues. The first is whether certain opinions that were not specifically identified in the government's Notice should be precluded solely for that reason. The second is whether some of the proffered opinions are really fact testimony rather than expert opinion testimony.

A. The Sufficiency of the Government's Notice of Expert Opinions.

As mentioned earlier, at the Daubert hearing the defense objected to Agent Sonnendecker testifying at trial about any opinion that was not specifically detailed in the government's Notice. The opinion that the defense argues was not properly noticed is the movement of cellular devices based on accessed cell sites. The defense objects to any opinion being offered at trial regarding the movement of phones because the Notice does not identify the specific cellular devices that will be the subject of that proposed expert testimony. The defense points out that the District Court ordered the government to detail with specificity every opinion that Agent Sonnendecker will testify to at trial. The defense argues that the general statement about “the movement of phones” does not comply with the District Court's Order and should be precluded for that reason.

The government argues it complied with the District Court's Order because the Notice makes clear that Agent Sonnendecker will testify about the movement of phones. Moreover, the fifteen subject areas detailed in the Notice identify the specific phones that accessed cell sites at various points in time. As a result, the defense was on notice that Agent Sonnendecker would testify about the movement of those phones.

The Court finds that the defense's point is well taken. The fifteen subject areas identified in the government's Notice only state that Agent Sonnendecker will testify about phones that accessed certain cell sites at various locations. The Notice does not explicitly state that he will also testify about the movement of any specific phone. However, the government's position is not unreasonable if the two provisions at issue in the Notice are read in conjunction. Specifically, the Notice makes clear that Agent Sonnendecker will testify about the movement of phones, and the fifteen identified subject areas detail the phones in play for Agent Sonnendecker's testimony. In fact, it is hard to imagine what other phones the government would be referring to.

All that said, this Court is ultimately going to punt this issue to the District Court. That court has repeatedly dealt with the defense's arguments that the government has failed to specifically set forth the expert opinions (both of this expert and other experts) that the government wants to introduce at trial. The result was the Order referred to earlier. The District Court is in the best position to determine if the government complied with the letter or at least the spirit of that court's Order.

B. Agent Sonnendecker's Testimony About Cell Phone Numbers That Were In Contact With Other Phone Numbers and the Frequency of Contacts Is Not Opinion Testimony.

As discussed earlier, the government's Notice states that Agent Sonnendecker will testify that there were cellular events - i.e., incoming and outgoing calls and/or text messages - on various dates and times for cell phone numbers attributed to the defendants, co-defendants, a cooperator, and a victim, as well as frequency of those events on any given day and time frame. At the Daubert hearing, Agent Sonnendecker also testified about time periods when there were no cellular events on many of the phones at issue.

The Court finds that this proffered testimony is not expert opinion testimony because it is based solely on phone records which are not subject to any expert interpretation. Agent Sonnendecker merely reviewed phone records (albeit with the use of computer-type technology) to determine which phone number attributed to the people identified above made or received a phone call and/or sent or received a text message. In theory, any person could have gathered that information if s/he went through the laborious task of wading through thousands of pages of phone records. The records are what they are - they reflect that a phone number was in contact with another phone number on a certain date and time. As such, Agent Sonnendecker is not opining that a call was made; the phone record shows it was made, as well as to whom and the duration.² Agent Sonnendecker more is akin to a summary witness who is merely testifying about the cellular events (i.e., calls and texts) that are reflected in the massive amount of phone records to streamline the presentation of the case to the jury. The defense can certainly challenge his summary witness testimony based on the admissibility of the underlying records that Agent Sonnendecker reviewed or the thoroughness and/or accuracy of his review. But neither of those challenges are currently before the Court.

2 At the Daubert hearing, the defense objected to testimony about a call between Michael Williams and Mr. Rakestraw around the time of the Floyd Davis homicide because that testimony was not detailed in the government's Notice. For the same reason, the defense also objected to testimony about time periods when there were no calls or texts on various cell phones, which presumably implies that the defendants and co-defendants stopped communicating after the homicide. The Court finds that this is also fact testimony (i.e., calls/tests were either made or not made based on phone records) that did not need to be set forth in the government's Notice detailing the expert opinions.

With respect to the phone numbers identified in the government's Notice, the Court concludes that Agent Sonnendecker can provide fact testimony about the dates and times calls and texts were made and received and the associated phone numbers, including the frequency of calls and texts, the duration of calls, and the number of texts in a certain time period. He can also testify as to the periods of inactivity on the applicable phones.

C. The Proffered Opinions About the Location and Movement of Phones Are Unreliable Because They Lack Foundation.

As discussed earlier, Agent Sonnendecker plans to testify that the cell phone numbers attributed to various people - i.e., the defendants, co-defendants, a cooperator, and a victim - accessed certain cell sites whose “probable” or “possible” coverage areas encompassed relevant locations (e.g., crime scenes, the shops, a car wash), which indicates that the phone was near the same location. He also plans to testify about the movement of phones based on the cell sites that those phones accessed during certain time periods.

The government argues that Agent Sonnendecker's testimony about what cell site was accessed by a cell phone is fact testimony, not expert testimony, because it is based solely on the phone records. Even if that is true, factual testimony about what cell site was accessed is only probative evidence if the coverage area of the cell sites is known, which is not the case here. Moreover, the government clearly wants to go several steps further than factual testimony about phones accessing cell sites. As discussed below, the government plans to use Agent Sonnendecker's expert testimony to show that certain phones had to be at a crime scene or other relevant location. Moreover, the government wants Agent Sonnendecker to opine that the phones moved toward a crime scene and then away from the scene after the crime was committed. Without knowing the coverage area of any of the cell sites, there is no foundation for expert testimony about the location and movement of phones. Thus, the opinions are not reliable because they are not based on sufficient facts and data. Compare United States v. Morgan, 45 F.4^th 192 (D.C. Cir. 2022) (district court reasonably determined that expert's testimony about the coverage area of cell sites that resulted from drive-scan tests was based on sufficient facts and data).

Agent Sonnendecker repeatedly cautioned that his testimony at trial will be that phone numbers accessed certain cell sites at various points in time, and as a result, the handset had to be within the “probable” or “possible” coverage area of the accessed cell site. However, the “probable” or “possible” coverage area of a cell site means little when the coverage area is undefined and the handset could be miles away from the accessed cell sites and the relevant locations, all of which are in close proximity to one another. Similarly, the foundation for and the probative value of testimony about the movement of cell phones hinges on the same unknown - the coverage area of accessed cell sites.³

3 It is worth noting that Agent Sonnendecker was noncommittal, or perhaps extremely cautious, when responding to government counsel's questions about the movement of phones. For example, in response to counsel's question of whether the activation of two cell sites five minutes apart could show the movement of F.D.'s phone, he testified: “It could have been.” Id.

Additionally, even though Agent Sonnendecker made clear that he will not and cannot testify about the directional path that a phone traveled, there is no doubt that the government wants to use Agent Sonnendecker to establish the path the defendants' phones, and hence the defendants, travelled to get to and from the crime scenes. Specifically, the government plans to use an animation during Agent Sonnendecker's testimony at trial which shows a linear path with handsets moving toward a crime scene, stop at the crime scene, and then move away from a crime scene. However, Agent Sonnendecker admitted that he has “no way of verifying whether mode of travel is even close to what happened[.]” Id. at 177. Nevertheless, the government still maintains that Agent Sonnendecker's testimony is “incredibly relevant” because it shows that the defendants are “speeding toward the scene before it happens and they're speeding away afterwards.” Id. 12/14/22 Tr. at 77. Both the animation and that argument would be incredibly misleading to the jury because there is simply not sufficient foundation to make Agent Sonnendecker's opinions regarding the movement and locations of phones reliable.

The need to ascertain the coverage area of a cell site to establish the general location of a phone is demonstrated in United States v. Ray, 2022 WL 101911, at *6 (S.D.N.Y 2022). In that case, the district court held that the government's Rule 16 expert disclosure was deficient in one respect: the notice did not disclose the substance of the expert's expected testimony regarding the approximate range of distance the wireless devices would be from the cell site location. The court held that “if the witness will testify that a wireless device was within the general location of a cell site or a cell tower, the government should disclose what [it] means by ‘general location.'” Ray, 2022 WL 101911 at *6.

Here, Agent Sonnendecker is clearly opining that the handset was within the “general location” of a cell site, even though his proposed testimony is couched in terms of the handset being within the “probable” or “possible” coverage area of a cell site. As in Ray, the government has not disclosed what it means by the “probable” or “possible” locations of phones in relation to the accessed cell sites. That disclosure has not been made because Agent Sonnendecker cannot do so.

Agent Sonnendecker repeatedly made clear that he does not know the coverage area for any cell site involved in this case. He does not know the specific engineering for any cell site, how they are designed, or the needs for a specific cell site. He testified that coverage area of a cell site is “going to vary based on the cell phone company itself, the technologies that they're trying to deploy in a certain area.” 11/15/22 Tr. at 20. “Generally speaking, each cell site is designed to cover whatever area the network engineers design it to cover.” Id. at 47. As a result, he cannot “say that a particular cell site covers a quarter of a mile or three miles or something to that extent because each one is designed to do something slightly different.” Id. In fact, he admitted that without an understanding of the engineering of a tower, it is feasible that a phone could be more than three miles from the tower it is accessing. Id. at 139.

Agent Sonnendecker testified that a cell phone accesses a cell site based on the strength of the signal and not the proximity of the handset to a cell site. Id. at 97, 99. As a result, he conceded that when he says that a cell phone is “within the coverage area, that doesn't necessarily mean that that tower” is the closet one to the handset. That concession (based on the science used to develop cell sites) coupled with undefined coverage areas undercuts the reliability of Agent Sonnendecker's opinions that the phones at issue had to be in the coverage areas of cell sites closet to the relevant locations.

The reliability of Agent Sonnendecker's opinions is also undercut by the fact his methodology did not consider the external factors that impact the coverage area of the cell sites at issue. In Ray, the district court rejected the defense argument that the cellular expert's methodology was unreliable because it did not take into account the effect of certain characteristics of Manhattan (e.g., surrounding water and the density and complexities in coverage resulting from people and buildings). 2022 WL 101911 at *7. The court found that the expert had considered these external factors when previously testifying as an expert in another case involving cell sites in Manhattan and the Bronx, and there was no reason to believe he would not do so again in this case. Id. In fact, the government represented that the expert's testimony would be “based on a review of the geography, topography, and environmental and man-made factors that might affect the functioning of cell sites in Manhattan and New Jersey.” Id.

By contrast, here, Agent Sonnendecker's testimony is not based upon a review of any of these external factors that affect the coverage area of cell sites. Agent Sonnendecker conceded that in 2015 the coverage area of the cell sites could have been a couple hundred feet or as far as ten miles, depending on the design of the cell site, the height of the cell tower, the angle of the antenna, the population density of an area, and the topography of an area. Id. at 139, 158. Similarly, in response to defense counsel's question of whether the fact that a handset connects to another cell site suggests that the phone moved locations, Agent Sonnendecker testified that it depends on a lot of factors, such as signal strength, topography, and population density. 11/15/22 Tr. at 155. But Agent Sonnendecker did not consider these external factors that affect cell site coverage; in fact, he did not even go to the cell site locations (in 2015 or subsequently) to examine, let alone assess, any external factors. The cell site near the LaQuinta Hotel and the Tucson Airport, which is important to the government's case, is a prime example.

Agent Sonnendecker testified that in a dense urban area the coverage area is generally about a mile to a mile-and-a-half. 11/15/22 Tr. at 48, 78. But in rural areas the coverage area is much greater. 12/14/22 Tr. at 30. He agreed with counsel that the area north of the airport and the La Quinta Hotel has dense housing, but “it's pretty desolate land” to the south and east of the airport. However, there was no testimony about whether the area around the LaQuinta Hotel and the airport would be considered an urban area or a rural area. Agent Sonnendecker merely testified that the area is “relatively flat,” but “[s]ome areas are rolling,” and there are some higher buildings, “but not skyscrapers[.]” 11/15/22 Tr. at 107. However, these vague descriptions of the geography and topography could apply to both urban and rural areas. Again, Agent Sonnendecker never went to the area around this cell site to assess the geography and topography.⁴ In fact, he testified that he does not know the “ins and outs of that area.” Id. at 108. Thus, not only is the coverage area for the cell site around the LaQuinta Hotel undefined, the coverage area cannot even be narrowed down to the mile-and-a-half coverage area that is generally seen in urban areas.

4 Agent Sonnendecker looked at the area around the cell site near the LaQuinta Hotel using Google Earth. However, that program provides a top-down view of the area which is not going to show the height of the cell tower or the buildings or hills in that area that could affect signal strength.

The uncertainty of the coverage area of the cell site near the LaQuinta Hotel is further demonstrated by Agent Sonnendecker's noncommittal testimony at the Daubert hearing. When government counsel asked Agent Sonnendecker if he would expect the LaQuinta Hotel to be within the coverage area of the cell site which is about two-tenths of a mile away, he hedged his bet. He testified: “It very well could be. Now, that particular sector could provide coverage into . . . the LaQuinta, or another sector on that could also provide coverage, too.” 11/15/22 Tr. at 59. Similarly, on cross-examination, he testified that coverage area of the cell site near the LaQuinta Hotel “provided likely coverage into that area.” 11/15/22 Tr. at 140. Testimony of what “could be” or is “likely” is simply not helpful to the jury.

Additionally, and perhaps most importantly, both the cell sites and the relevant locations (e.g., crime scenes, the “shops,” a car wash) where the government wants to place the phones appear to be well within five square miles of each other.⁵ When the coverage area is unknown, the proximity of these cell sites and locations weakens the probative value of evidence about phones accessing a cell site because a phone could be located anywhere within these five square miles. An example best illustrates why the probative value and reliability of Agent Sonnendecker's testimony is especially low when the coverage area is undefined for closely situated cell sites.

5 The Court is providing its best estimate based solely on the government's exhibits because Agent Sonnendecker did not testify about the distance between each of the various cell sites, the distance between the relevant locations and the accessed cell sites, or the distance between each of the relevant locations. Government's Exhibit Two contains various maps depicting the locations of cell sites. Each map notes the mileage north/south and east/west represented on the map, as well as cell sites located within the resultant area depicted on each map. However, it is difficult to pinpoint the exact distance between various cell sites, and there was no testimony on that point. But, as noted above, they appear to be within, at most, five miles of each other. For example, the map on page two notes a north/south distance of 12.2 miles and an east/west distance of 20.6 miles. However, the cell sites are closely clustered within the center of the map and nowhere near the outer distances noted on the map. The map on page three notes a north/south distance of 5.84 miles and an east/west distance of 9.85 miles. But, again, no cell comes close to the outer distance noted on this map. The map on page four notes a north/south distance of 3.25 miles and an east/west distance of 5.66 miles. The map on page five notes a north/south distance of 3.8 miles and an east/west distance of 6.19 miles. The map on page seven notes a north/south distance of 8.62 miles and an east/west distance of 5.29 miles; but the two cell sites also noted appear to be well within five miles of each other. The map on page eight notes a north/south distance of 3.08 miles and an east/west distance of 5.38 miles. The map on page fifteen notes a north/south distance of 2.65 miles and an east/west distance of 4.62 miles. The map on page sixteen notes a north/south distance of 4.02 miles and an east/west distance of 7.01 miles; the north/south distance between cell sites is close to four miles, but the east/west distance between cell sites is less than that.

If a phone number attributed to a defendant who lives in the Phoenix area is accessing a cell site located in Tucson near a crime scene at or near the time of a crime, then that evidence would be highly probative as to the location of a phone even if the coverage area is undefined. Based on Agent Sonnendecker's testimony, there is no way that the coverage area of a Tucson cell site could have extended to Phoenix such that the handset could have been in the Phoenix area. Rather, the handset had to be in the general coverage area of the cell site in Tucson, regardless of whether the coverage area is 500 feet or ten miles. In that scenario, a phone number accessing of a cell site in the Tucson area would be extremely probative and reliable evidence that a defendant's phone, and likely the defendant, was not in Phoenix, but rather in the general vicinity (whether feet or miles) of a crime scene in Tucson at the time of the crime.

The Court's example is what occurred in Morgan, where the defendant lived in Maryland and the minor victim lived in Washington, D.C. 45 F. 4^th at 196-197. Cell site information showed that the defendant's phone accessed a cell site “whose coverage area was entirely within D.C.” around the time the defendant picked up the minor victim at a bus stop in D.C. Id. at 199. Cell site information also reflected that the defendant's phone accessed a cell site in Maryland whose coverage area included the defendant's home at around the time the minor victim was at the defendant's home. Id. Thus, the cell site information which, unlike in the case at hand, identified the coverage areas of the cell sites in D.C. and Maryland, was probative to show that the defendant's phone had to be in the D.C. area at the time the minor victim was picked up and had to be in the area around his home in Maryland when the minor victim was at his house.⁶

6 See also United States v. Penbrook, 119 F.Supp.3d 577, 597 (E.D. Mich. 2015) (proposed expert testimony was reliable to show that defendants' phones numbers initially accessed cell sites in Philadelphia, Pennsylvania, then accessed cell sites in the Grand Rapids and West Bloomfield areas of Michigan at around the time of the robberies, and thereafter accessed cell sites in Philadelphia); United States v. Reynolds, 626 Fed. App'x 610, 617 (6^th Cir. 2015) (court noted that it did not need to resolve the circuit split on the reliability of using historical cell site analysis to determine if a phone is in a specific cellsector, because the cell site analysis in this case was used to identify a cell-sector “in which callers were not.”); United States v. Jones, 2022 WL 17884450, at *7 (W.D. Ky. 2022) (although expert testimony was not reliable to put the defendant's phone near the robberies, it was reliable to show that the defendant's “phone was at least in Louisville.”).

The government argues that a recent Ninth Circuit decision (issued after the Daubert hearing and oral argument) is directly on point to the issue at hand. See United States v. Baker, 2023 WL 1095359 (9^th Cir. 2023). The government points out that the Ninth Circuit held that the district court properly admitted cell site evidence to show the movement of the defendant's phone to and from the crime scene. Doc. 2298. The defense argues that Baker does not control the outcome in the case at hand, in part, because of the scant discussion of the foundation that supported the expert's opinions. Doc. 2311. This Court agrees with the defense.

In Baker, the defense argued that the district court erred in admitting the testimony of the government's cell data mapping expert because the expert did not circumscribe the import of historical cell site data by explaining that this data shows a phone's location within a range rather than pinpointing it exactly. 2023 WL 1095359, at *11. The expert testified at trial that cell site data can show where a phone was in a general sense at a given time and described how he used that data to conclude that the defendant's phone moved from one area of Los Angles to the general area of the crime scene at the approximate time the crime occurred. Id. The Ninth Circuit held that the district court did not err in admitting this testimony because the jury was adequately informed of the limitations of cell site data. Id.

Baker is of little value to the case at hand because of unknowns in that case. In fact, that case raises more questions than it provides answers. The Ninth Circuit never discussed the foundational facts that supported the expert testimony regarding coverage areas. For instance, the court never mentions if the coverage area of cell sites was ascertained and/or if the expert took steps to at least narrow down the likely coverage area based on topography and population density. Relatedly, the court's reference to the movement of the phone from “one area of Los Angeles” to the “general area of the crime scene” gives no indication of the distance the phone moved within this sprawling and densely populated city. Based on Agent Sonnendecker's testimony, the coverage area for a cell site in a densely populated city like Los Angeles would be very limited. Thus, if the cell sites activated “in the Los Angeles area” were many miles apart, then this testimony on the movement of phones was probative even if the coverage area is undefined.

By contrast, the same is not true here because the accessed cell sites and relevant locations are clustered within five square miles. As a result, the coverage area of the cell sites is essential to establish the general location of a phone. To be clear, unlike the defendant in Baker, the defense is not arguing that a phone's location must be established with pinpoint accuracy, and this Court is not requiring pinpoint accuracy. Rather, the Court is requiring a defined coverage area for cell sites so the general location of a phone can be established. Because the coverage area is unknown, there is not sufficient foundation for Agent Sonnendecker's expert testimony about the general location of phones.

Simply stated, the unknown coverage area of cell sites coupled with the proximity of the cell sites render it impossible to come up with a reliable opinion that a handset was within the “general area” or “proximity” of a pertinent location. Agent Sonnendecker conceded that a handset could have been stationary and accessed a different cell site for any number of reasons - e.g., high activity on a closer cell site or signal strength at any given point in time. Once again, he testified that a lot of factors, such as signal strength, topography, and population density, factor into the analysis of whether the movement of a phone is demonstrated by the phone accessing a different cell site. Because those factors are unknown, if the handset was in fact moving, it could have moved to the east or west and not south toward the crime scene, even though the cell site near a crime scene was accessed by handset. In fact, because signal strength and not proximity determines which cell site is accessed, even a slight movement of the handset could cause it to access a different cell site.

Additionally, no evidence was presented about whether there were other cell sites for the relevant phone carriers within this five-mile area that were not accessed at the time of any given cellular event in 2015. That information may have helped establish, by a process of elimination, that a handset was more likely to be near the accessed cell site based on the direction of an antenna of the cell site that was not accessed. Agent Sonnendecker testified there may have been other cell sites in 2015, but he would have had to “dig down deeper” to try to figure out the possible coverage area of an accessed cell site based on other surrounding cell sites that were not accessed. But he did not do that or take any other steps to ascertain the coverage area of the cell sites.⁷

7 Agent Sonnendecker also does not know if cell sites at issue have been upgraded, if antennas were added to cell towers, and/or if the coverage area increased or decreased since 2015, all of which would impact the distance of the coverage area of the cell sites in 2015. See Morgan, 45 F.4^th at 203-204 (No error in admitting cell site information about the location of the defendant's phone because the district court “required the agents to make sure that there were no changes to the tower or cellular network between the time of the relevant incident and the time of the [drive scan] test.”). Agent Sonnendecker certainly knew that those improvements could be made because when he worked as a drive tester the engineers would “either adjust the antenna itself or possibly increase the strength of the radio signal that was being sent . . .in order to better the network” in a given area.” 11/15/22 Tr. at 13.

Sufficient foundation for Agent Sonnendecker's opinions that phones were within a coverage area of cell sites and moved into a coverage area could also have been established in several (easier) ways. The technology existed in 2015 for Agent Sonnendecker to do drive scan testing of the cell sites to determine their coverage area. See Morgan, 45 F.4^th at 203-204 (district court correctly concluded that drive-scan tests were reliable because the court “assessed each step of the drive-test methodology in substantial detail” to ensure that the government had “adequately assessed potential sources of error in the testing methodology.”); United States v. Frazier, 2016 WL 4994956, at *3 (D. Nev. 2016) (reliability and relevance requirements of Daubert where satisfied by the expert's use of historical cell site analysis and “drive test” methodology to approximate the location of the defendant's phone).⁸ Although ATF did not have that equipment in 2015, he testified that he had borrowed that equipment from other law enforcement agencies in the past. 11/15/22 Tr. at 125.

8 In Morgan, the court explained although drive tests are used primarily by wireless telephone companies, “law enforcement personnel also conduct drive tests, typically to determine the approximate coverage area of a particular tower to which a cell phone of interest connected during a relevant time period. By determining a tower's coverage area, law enforcement can narrow down the cell phone's location when it connected to that tower.” 45 F. 4^th at 197-198 (internal citations omitted).

Additionally, the government could have requested records from the phone companies that specified the coverage area for the cell sites as they existed in 2015 and/or called representatives from cell phone companies to testify about coverage areas. Agent Sonnendecker admitted that the phone companies have those records. 12/14/22 Tr. at 43. However, he testified that during his tenure with ATF he has never been able to see those records. Id. at 44. In fact, on a couple occasions those records were requested but not provided; as a result, “it's sort of something that we don't even ask for anymore because we don't believe it will be provided.” Id. The Court is not convinced by that explanation. The cell phone companies provided Agent Sonnendecker with documents to establish the direction of an antenna and its azimuth, which is basically the width of the coverage area of a cell site sector. Thus, it is hard to believe that a phone company would refuse to comply with a grand jury subpoena for documents reflecting the length (hundreds of feet or miles) of the coverage area within a sector. The phone companies may well have resisted producing proprietary information; but the government did not need that information.

Agent Sonnendecker similarly testified he does not generally receive GPS information from phone companies even though GPS technology would be the most accurate method to determine the location of a phone. 11/15/22 Tr. at 127. However, he did not make clear if he has ever requested GPS records from phone companies and they refused to provide them. He also testified that the next most accurate data for locating a phone would be specialized location records. Id. However, once again, he did not elaborate on whether he has ever requested those records and the phone companies did not provide them.

Agent Sonnendecker also testified that if we wanted map out the exact “coverage area a given cell site” or a given sector, we could “try to obtain RF propagation reports from the cell phone companies.” Id. at 47-48. But, again, it is unclear whether Agent Sonnendecker has requested RF propagation reports but they were not produced. It is also unclear whether the RF propagation reports are yet another category of records that could have been obtained to determine cell site coverage, or these reports are the same as other records previously discussed.

What is clear is that the government never took any of the many steps available to ascertain cell site coverage to establish a foundation for the methodology for Agent Sonnendecker's opinions. The bottom line is that there are more unknowns about the coverage area of the cell sites than knowns. As a result, Agent Sonnendecker's opinions about phones being within a coverage area of a cell site or moving to a cell site are not reliable.

The government argues that defense counsel can cross-examine Agent Sonnendecker about the undefined coverage area and that signal strength, not proximity, determines which cell site is accessed. The defense can then argue to the jury that the phones could have been ten miles from the accessed cell site and the jury would ultimately decide how close the phones were to the accessed cell site.

The Court is wary of whether the cross-examination proposed by the government would be effective or sufficient to deal with the uncertainly of the coverage areas of the cell sites. Expert testimony is always powerful evidence. In the case at hand, the jury would be hearing testimony from an agent who has worked with cellular data and information basically his entire career. While that experience shows that Agent Sonnendecker is qualified to testify as an expert on cellular data, it also creates a risk that the jury will defer to his opinions based on that experience, and not pick up on the nuances about the lack of foundation for his opinions that the defense will try point out on cross-examination.⁹

9 Also, if Agent Sonnendecker were permitted to testify as an expert, the defense will presumably call the expert that they have retained to testify in an effort to undercut the impact of Agent Sonnendecker's testimony. The result is a trial within the trial that likely would have been unnecessary if the government had obtained evidence to define the coverage area of the cell sites at issue.

More importantly, the government's argument that the defense can cross-examine Agent Sonnendecker about the variables that may have impacted the coverage area of the cell sites ignores the foundational requirements for expert testimony. The government has the burden to establish proper foundation for expert testimony. However, the government is essentially trying to skirt its duty to provide the requisite foundation for expert testimony by shifting the burden to the defense to present evidence at trial that impacts the reliability of Agent Sonnendecker's opinions. Again, the reliability of expert testimony turns on the evidence that government has presented to establish sufficient foundation for that testimony; the defense bears no burden in proving that threshold requirement. As discussed at length above, there is not sufficient foundation for Agent Sonnnendecker's proposed expert testimony.

In summary, the purpose of expert testimony is to aid the jury in understanding scientific or technical evidence that is not common knowledge. The jury is not aided by Agent Sonnendecker's testimony because he cannot explain the science behind or the technology used for cell sites at issue in this case. All the jury is told that the coverage of a cell site could be several hundred feet or many miles, which does not help them in understanding or evaluating the evidence in this case. To the contrary, Agent Sonnendecker's expert testimony creates the risk of misleading the jury and prejudicing the defendants.

CONCLUSION

The Court finds that Agent Sonnendecker's testimony is not based on sufficient facts and data because he did not obtain information and data to establish the coverage are for the cell sites at issue. As a result, his opinions on the location and movement of cell phones are not based on reliable principles and methods which can be applied reliably to the facts of this case. Therefore, the opinions would not assist the jury in understanding this cell phone evidence and would mislead the jury and create of risk of unfair prejudice.

For these reasons, it is recommended that the District Court preclude Agent Sonnendecker from providing expert opinion testimony.

Pursuant to Federal Rule of Criminal Procedure 59(b)(2), any party may serve and file written objections within 14 days of being served with a copy of this Report and Recommendation. A party may respond to the other party's objections within fourteen days. No reply brief shall be filed on objections unless leave is granted by the district court. If any objections are filed, this action should be designated case number: CR 18-01695-TUC-JAS. Failure to timely file objections to any factual or legal determination of the Magistrate Judge may be considered a waiver of a party's right to de novo consideration of the issues. See United States v. Reyna-Tapia, 328 F.3d 1114, 1121 (9th Cir. 2003) (en banc).