1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 7 GTE MOBILNET OF CALIFORNIA Case No. 20-cv-05460-DMR LIMITED PARTNERSHIP, 8 Plaintiff, FINDINGS OF FACT AND 9 CONCLUSIONS OF LAW v. 10 CITY OF BERKELEY, 11 Defendant. 12 13 I. INTRODUCTION 14 This is an action under the Telecommunications Act of 1996 (“TCA”), 47 U.S.C. § 15 332(c)(7). Plaintiff GTE Mobilnet of California (“Verizon”) alleges Defendant City of Berkeley 16 (“Berkeley” or “the City”) unlawfully denied Verizon’s application to construct a personal 17 wireless service facility in Berkeley, California. The court determined on cross motions for 18 summary judgment by Verizon and Intervenor-Defendants Berryman Reservoir Neighbors 19 (“BRN”) that Verizon was entitled to summary judgment on the second prong of its claim that 20 Berkeley’s denial of the application had the effect of prohibiting Verizon from providing personal 21 wireless services in violation of 47 U.S.C. § 332(c)(7)(B)(i)(II) (the “effective prohibition” claim). 22 It also determined that Berkeley failed to act on Verizon’s application “within a reasonable period 23 of time” after it was filed in violation of 47 U.S.C. § 332(c)(7)(B)(ii) (the “failure-to-act claim). 24 GTE Mobilnet of Cal. Ltd. P’ship v. City of Berkeley, No. 20-CV-05460-DMR, 2023 WL 25 2648197, at *18, 28 (N.D. Cal. Mar. 27, 2023).1 26
27 1 The court granted BRN’s motion for summary judgment on Verizon’s claim that the denial of 1 The remaining issues are (1) the first prong of Verizon’s effective prohibition claim, which 2 requires it to show that a “significant gap” in service coverage existed at the time of its application 3 to build a personal wireless service facility, and (2) the appropriate remedy for the failure-to-act 4 claim. See id. at *29. The court held a bench trial on the significant gap issue at which Verizon’s 5 expert, Richard Conroy, and BRN’s expert, Robert Beegle, testified. Pursuant to Federal Rule of 6 Civil Procedure 52(a), the court makes the following findings of fact and conclusions of law. 7 II. FINDINGS OF FACT 8 A. Verizon’s Application & ZAB’s Denial 9 On December 17, 2018, Verizon submitted an application for land use permits (“the 10 Application”) to Berkeley’s Planning and Development Department, Land Use Planning Division 11 to build a wireless telecommunications facility at the East Bay Municipal Utility District 12 (“EBMUD”) Berryman Reservoir in Berkeley, California. [Docket No. 126 (Jt. Pretrial 13 Statement) 4-5 (Joint Statement of Undisputed Facts, “JSUF”), A.] The Application proposed 14 installation of a 50-foot monopole disguised as an evergreen tree along with six antennas, six 15 remote radio units, and related cables and equipment mounted on the monopole, with supporting 16 equipment installed on the ground. Administrative Record, “A.R.” 1-55. 17 In a section of the Application entitled, “Description of the Coverage area,” Verizon wrote 18 that “Verizon’s objective is to improve coverage in the Berkeley Hills area, especially along 19 Euclid Avenue north of EBMUD’s Berryman Reservoir. This area is primarily residential and 20 wooded, and provides few traditional locations for cell sites.” A.R. 5. This was the sole 21 description of the “coverage area” in the Application. 22 In a section of the Application entitled, “Description of Services,” Verizon wrote, “Verizon 23 is proposing to provide LTE [long term evolution] service from this facility.” A.R. 6. The 24 proposed wireless facility would provide 4G LTE service, which commingles voice and data 25 services. See Trial Transcript Vol. I (“TTI”) 75-77. 4G LTE travels over four frequencies: 700 26 MHz, 850 MHz, 1900 MHz, and 2100 MHz. The 700 MHz and 850 MHz frequencies are the two 27 “low band” frequencies, while 1900 MHz and 2100 MHz are the “high band” frequencies. TTI 1 band frequencies support greater usage but travel shorter distances than the low band frequencies. 2 TTI 87. 3 In a section of the Application entitled, “Statements Related to Need,” Verizon wrote that 4 “Verizon’s coverage objectives for this project are to improve service in the area described above, 5 and to offload traffic from other nearby sites that are often at or exceeding capacity,” but the 6 Application did not specify any neighboring sites that were at or exceeding capacity, nor did it 7 provide data supporting this claim. A.R. 5. “Coverage” refers to Verizon’s “ability to provide a 8 reliable signal to an area,” while “capacity” refers to users’ ability to access the network. TTI 85.2 9 Berkeley’s Zoning Adjustments Board (“ZAB”) scheduled a public hearing on the 10 Application for June 27, 2019. A.R. 122. Prior to the hearing, Berkeley’s technical planning staff 11 submitted a report to the ZAB analyzing the Application and recommending that it be approved. 12 A.R. 129-37; JSUF B. The ZAB reviewed the Application at a public hearing on June 27, 2019 13 and voted to deny the Application. JSUF C. 14 B. Verizon’s Appeal & the Kharaba Report 15 Verizon appealed the ZAB’s denial to the Berkeley City Council on July 16, 2019. JSUF 16 D. On March 5, 2020, Verizon submitted a report by Amr Kharaba, a Verizon radiofrequency 17 design engineer, to the City in support of its appeal (the “Kharaba Report”). JSUF E; A.R. 2976, 18 1582-91 (Kharaba Report). Kharaba wrote that Verizon “has identified a significant gap in its [4G 19 LTE] service in the north Berkeley hills residential neighborhoods” based on “inadequate 4G LTE 20 service coverage” from five nearby facilities as well as “capacity exhaustion of the Verizon 21 Wireless facility that provides the most service to the gap area.” A.R. 1582. According to 22 Kharaba, the “coverage gap” and “capacity gap” “have resulted in the Significant Gap in Verizon 23
24 2 Verizon submitted two coverage maps with the Application labeled “Before Site RSRP@700” and “After Site RSRP@700.” A.R. 17, 18. As discussed below, “RSRP” stands for “Reference 25 Signal Received Power.” RSRP is a parameter that “depicts the signal strength” and is “used to define coverage in LTE networks.” TTI 106-07. The two coverage maps purport to show by 26 color “In-Building Coverage,” “In-Vehicle Coverage,” and “On-Street Coverage” for the 700 MHz band. The maps do not contain a scale, identify a metric, or provide a description of the 27 methodology used to create them. A.R. 17, 18. Verizon does not rely on the coverage maps to 1 Wireless 4G LTE coverage and network capacity in the north Berkeley hills.” A.R. 1590. 2 With respect to the purported “coverage gap,” Kharaba wrote, “[t]his area currently 3 receives inadequate 4G LTE service coverage from the existing Verizon Wireless Kensington 4 Circle facility 1.1 miles northwest of the proposed facility, the Shattuck North facility 0.5 miles 5 southwest, the Lower University facility 1.6 miles southwest, the Berkeley Bekins facility 1.8 6 miles southwest, and the Gilman Street facility 1.9 miles west.” A.R. 1582. He continued, “[a]s a 7 result of the distance from existing facilities, there is a gap in 4G LTE in-building and in-vehicle 8 service coverage in the north Berkeley hills, and areas lacking outdoor coverage.” Id. 9 The Kharaba Report cited results of a December 2019 drive test and coverage maps as 10 evidence of the alleged coverage gap. A.R. 1583. With respect to the drive test, Kharaba stated 11 the test was performed by Gerald Kinney, Principal System Performance Engineer, from 12:07 12 p.m.-3:18 p.m. on December 17, 2019 using two devices, a Samsung Galaxy S7 and a Samsung 13 Galaxy S8. Kharaba asserted that Kinney used “DMAT,” a Verizon Device Monitoring and 14 Analysis Tool, to perform the analysis and that Kinney scanned the 700 MHz, 850 MHz, 1900 15 MHz, and 2100 MHz bands. A.R. 1583. 16 The Kharaba Report presented the drive test results in a map displaying one metric, RSRP. 17 A.R. 1583-84. As noted, RSRP is a measure of signal strength; it does not vary based on usage. 18 TTI 107, 191, 195. The measurement of RSRP uses negative values, where a higher negative 19 number corresponds to a stronger power level, e.g., -85 dBm is stronger than -95 dBm. TTI 107. 20 Additionally, the dBm scale is logarithmic, whereby -95 dBm is 10 times weaker than -85 dBm. 21 TTI 138. The drive test map, entitled “Drive Test Result—4G LTE Service Levels,” plots 22 colored dots corresponding to RSRP levels onto a map of the area (the “Kharaba drive test map”). 23 A.R. 1584. 24 25 26 27 1 Drive Test Result - 4G LTE Service Levels □ rt fal esas en 4 Me |_| Invehicle >=-95 dBm 3 Sy Quidoor >=-105 dBm F fe rt Marginal >=-110 dBm □□ 4 Sere, None <==-110 dBm 4 = a vs i. ee er a OM \ 6 Hy NORE 3 Ps ) } 5 “1 7 \ « 8 ~ fo * . ta 8 ‘ A rroill □□ AbrtH Bot Xe Ia ww BD \ et aye: 9 ccm rm hi 10 : me “ i . : ll —
= 12 || Kharaba drive test map.
& 13 Verizon’s December 2019 drive test used -85 dBm as the standard required for in-building
v 14 || coverage, -95 dBm as the standard for in-vehicle coverage, and -105 dBm for outdoor coverage.
15 TTI 107-08; A.R. 1584.3 Kharaba concluded, “[t]his [drive test] map shows a lack of in-building A 16 || coverage in the north Berkeley hills and barely any in-vehicle coverage” and stated, “only outdoor
2 17 coverage is available, with pockets receiving marginal or no coverage.” A.R. 1584. Kharaba 2 18 || wrote, “the Proposed Facility will provide new reliable 4G LTE service coverage to the gap area, 19 || including areas near Euclid Avenue between Keith Avenue and Vine Lane,” and “will provide 20 || improved service coverage to an area of 1.6 square miles and a population of 2,420 residents.” 21 A.R. 1585. Kharaba did not otherwise specify the boundaries of the 1.6 square miles that 22 || comprise the alleged gap area, explain the basis for the number of residents in the alleged gap 23 area, or state how many of the 2,420 residents are wireless users. 24 29 | 3 The Kharaba Report also stated that Verizon “uses PCS and AWS to provide over 70 percent of 26 || its 4G LTE service capacity throughout Berkeley,” where PCS refers to the 1900 MHz band and AWS refers to the 2100 MHz band. A.R. 1584. The Kharaba Report included two coverage maps 97 || purporting to depict current and proposed 4G LTE AWS RSRP coverage of the area. A.R. 1586. Other than Conroy’s testimony that Verizon used a “common radio propagation tool” called 28 ATOLL to create the maps, TTI 111-12, Verizon did not otherwise offer evidence about how the maps were created, who created them, or the data supporting them.
1 As to the “capacity gap,” the Kharaba Report included a map labeled “Drive Test Result— 2 Best Serving Facility” that purports to show that 12 facilities provide signal to customer handsets 3 in the alleged gap area, and that the Lower University and Kensington Circle facilities provide 4 55.4% and 22.2% of service in the area, respectively. A.R. 1587. According to Kharaba, “the 5 Lower University facility that serves more than half of the gap area has reached capacity 6 exhaustion.” A.R. 1588. In support of this claim, Kharaba discussed one Key Performance 7 Indicator (“KPI”) that measures network performance, Average Scheduler Eligibility Usage 8 (“ASEU”). A.R. 1588-89. ASEU is “a capacity congestion key performance indicator.” TTI 78. 9 The Kharaba Report included “ASEU Capacity Charts” for the Lower University facility that 10 purportedly depict “how increased demand has already outstripped the capacity of the Lower 11 University facility’s antenna sectors that serve the gap area” and “the increased usage of [the 12 Lower University] facility through January 2020 [and] predicted usage through late 2020.” A.R. 13 1588-89. The ASEU Capacity Charts plot “Actual Usage,” “Normalized Usage,” “Usage Trend,” 14 and “Capacity Exhaustion” using color coding for two east-facing antenna sectors in the Lower 15 University facility. A.R. 1589 (ASEU Capacity Charts). The chart for the Alpha Antenna Sector 16 plots usage for six dates, including two future projections. The chart for the Delta Antenna Sector 17 plots usage for four dates, two of which were future projections. A.R. 1589. 18 19 20 21 22 23 24 25 26 27 1 ASEU Capacity Charts 2 Lower University Facility East-Facing Antenna Sectors 3 — _ Actual Usage Normalized Usage Usage Trend 4 ——— Capacity Exhaustion 5 Alpha Antenna Sector (80 degrees from north) ad Alpha 6 Sector 60— 4 r4o0— 8 9 20— 0 . . I 20-4 ' 1/1/2018 7/1/2018 §=-1/1/2019 7/1/2019 11/2020 71/2020 12
413 Delta Antenna Sector (101 degrees from north) aww Delta 14 Sector © 200 — 15 Sam 150 — 16 & 100 — = 17 50 — —— Z 18 □□ tT o— 19 -50 — 20 1/1/2019 7/1/2019 1/1/2020 7/1/2020 21 ASEU Capacity Charts. Neither chart cites a source for the usage data, includes any information 22 about how the past usage data was obtained, or states the time of day usage was measured. 23 Kharaba concluded that the “Lower University facility antenna sectors reached capacity 24 || exhaustion over one year ago. Capacity exhaustion severely compromises the Verizon Wireless 25 network in the entire area served by the exhausted antenna sectors, leading to call failures and 26 slow data speeds, as well as poor call quality and reliability over Verizon Wireless VoLTE 27 28
1 technology (the ‘Capacity Gap’).” A.R. 1589.4 The Kharaba Report did not include any data 2 about call failures, slow data speeds, call quality, or reliability. 3 On May 7, 2020, Telecom Law Firm, P.C. (“TLF”), an independent consultant retained by 4 Berkeley to analyze the Application, submitted a report examining Verizon’s claims regarding 5 service within the claimed gap area. A.R. 1718-48; JSUF F. TLF stated that it reviewed the drive 6 test data and coverage maps in the Kharaba Report, A.R. 1743-48, and noted that with respect to 7 the RSRP coverage standards, including the use of -85 dBm as the standard required for in- 8 building coverage, “the levels and descriptions shown [in the drive test results] are solely selected 9 by Verizon. They are not set or otherwise regulated by the FCC or any State of California agency, 10 and each wireless company is free to select its own levels and descriptions.” A.R. 1744. TLF 11 concluded that “there are areas within the claimed gap that have adequate Verizon signal strength 12 to provide personal wireless services, but other areas that do not,” and that failure to add another 13 cell site to serve the “claimed gap area” could lead to “lower data throughput speeds and 14 potentially some undisclosed degradation(s) on the ability of Verizon customers . . . to make and 15 receive calls in the claimed gap area.” A.R. 1748. 16 On July 7, 2020, the City Council heard Verizon’s appeal and voted to adopt a resolution 17 affirming the ZAB’s denial of the Application (the “Denial Resolution”). A.R. 2665-67; JSUF H. 18 Berkeley mailed written notice of its denial of the Application and the Denial Resolution to 19 Verizon on September 8, 2020. A.R. 2746-48, 2707-10; JSUF I. 20 C. Verizon’s Lawsuit 21 Verizon timely filed the original complaint against Berkeley on August 6, 2020. In 22 October 2020, the court granted BRN’s motion for leave to intervene. Verizon filed the first 23 4 At the June 27, 2019 hearing, Kharaba informed the ZAB that the gap area was currently served 24 by the Shattuck North and Downtown Berkeley facilities and stated, “these sites are already exceeding their capacity limitations.” A.R. 2859-60. There is no record evidence about the 25 capacity of the Shattuck North or Downtown Berkeley facilities. The March 2020 Kharaba Report identified the Lower University and Kensington Circle facilities as the two “best serving 26 facilit[ies]” and asserted that the Lower University facility “reached capacity exhaustion over one year” prior to the report. A.R. 1587, 1589. Verizon’s expert Conroy was unable to explain this 27 discrepancy at trial, see Trial Transcript Vol. II (“TTII”) 275-77, although he testified on redirect 1 amended and supplemental complaint (“FASC”), which is the operative complaint, on February 1, 2 2021. The FAC states three claims for relief: 1) failure to act on the Application “within a 3 reasonable period of time” after it was filed, in violation of 47 U.S.C. § 32(c)(7)(B)(ii) (the 4 “failure-to-act claim”); 2) denial not based on substantial evidence in violation of 47 U.S.C. § 5 332(c)(7)(B)(iii) (the “substantial evidence claim”); and 3) unlawful prohibition of service in 6 violation of 47 U.S.C. § 332(c)(7)(B)(i)(II) (the “effective prohibition claim”). 7 On cross motions for summary judgment by Verizon and BRN, the court found in 8 Verizon’s favor on the failure-to-act claim; in Berkeley’s favor on the substantial evidence claim; 9 and in Verizon’s favor on the second prong of the effective prohibition claim.5 The court denied 10 summary judgment on the first prong of the effective prohibition claim. Two issues remain: 1) the 11 first prong of the effective prohibition claim, i.e., the showing of a “significant gap” in service 12 coverage as of 2020, when Berkeley denied the Application; and 2) the appropriate remedy for the 13 failure-to-act claim. GTE Mobilnet, 2023 WL 2648197, at *18, 23, 28, 29. 14 D. The Parties’ Evidence Regarding Coverage 15 At trial, Verizon did not provide testimony from Kharaba or Kinney, the individual who 16 performed the 2019 drive test. Instead, Verizon offered the testimony of Richard Conroy, an 17 expert in the field of radiofrequency engineering and network design. Conroy testified that he 18 reviewed the administrative record, including the Application and the Kharaba Report; obtained 19 the “raw data” from the 2019 drive test; and obtained “additional key performance indicator data 20 from Verizon.” He also visited the “subject area” and performed his own drive test in September 21 2023. TTI 84. 22 With respect to the 2019 data, Conroy testified that he had access to the RSRP drive test 23 data; “different maps by frequency band and also by site, by server, by frequency band by server, 24 by site”; and one KPI, the ASEU data. TTII 300. Conroy testified about his understanding of how 25 the 2019 drive test was performed. According to Conroy, the tester used two phones “inside the 26
27 5 The second prong of the effective prohibition claim requires the provider to show “the lack of 1 vehicles, and then the vehicles were driven around the drive route” and the phones collected 2 signals. The drive test “collect[ed] information on the cell sites . . . providing coverage,” the 3 “frequencies . . . providing the coverage, the signal strengths, power levels, and the latitude and 4 longitude of each of the data points.” TTI 110. According to Conroy, the drive test “mirrors 5 what’s going on in the network” and reproduces the customer experience while inside a vehicle. 6 TTI 110. 7 Conroy testified that he received the raw data from the 2019 drive test in excel files. TTI 8 110. Specifically, he testified that he received an excel table that included “latitude, longitude, the 9 signal strengths, the cell I.D.S., where the site signals were coming from.” TTI 185. Conroy then 10 prepared a map using the 2019 drive test raw data on which he plotted signal levels using colored 11 dots that correspond to RSRP levels. TTI 104-06, Trial Ex. 13C. Conroy testified that he used the 12 2019 drive test raw data and “create[d] a virtually identical map” to the Kharaba drive test map, 13 “but clearer,” TTI 105, and that the 2019 drive test data he received from Verizon “match[ed]” the 14 drive test data shown in the Kharaba Report. TTI 186, TTII 306. 15 Several points undermine Conroy’s validation of the 2019 drive test data. First, Conroy 16 knew little about the phones used for the 2019 drive test, including how old they were at the time 17 of the test. TTI 192. He testified that “these were perfectly fine test phones when they were 18 utilized” but there is no evidence that he inspected or reviewed the test phones. TTI 187. Conroy 19 also admitted that he did not speak with Kinney, the person who performed the 2019 drive test. 20 TTI 177. When asked how he knew the test phones were “perfectly fine . . . when they were 21 utilized,” Conroy answered, “[f]rom the Verizon report and information they use and the data they 22 provided to me” without offering any specifics, TTI 187, and there is no evidence that Verizon 23 turned over the phones or data regarding the operation of the phones to the defense. Thus, there is 24 no evidence in the record about the age or condition of the phones used in the 2019 drive test. 25 Next, even though the Kharaba Report stated that Kinney used two phones to collect the 26 2019 drive test data, A.R. 1583, Conroy testified that Verizon provided him with data from only 27 one phone. TTI 185, 189. When asked about the data from the second phone, Conroy initially 1 all the data cumulative,” TTI 185-86, but stated that “it seemed to match the data that was in the 2 report. It seemed to be all I needed.” TTI 188. On the second day of his testimony, Conroy 3 offered a slightly different explanation, speculating that the data “could have been combined data 4 between the two phones.” TTII 229. Conroy subsequently admitted that he had met with 5 Verizon’s counsel the morning before he resumed his testimony on the second day of trial and that 6 he and counsel had discussed that “maybe the data was the combined combination of the two 7 phones.” He could not recall other specifics of their conversation. TTII 230-31. The court finds 8 that Conroy’s testimony that the 2019 drive test data could have been combined data from both 9 phones is not credible and affords it no weight. 10 Finally, the Kharaba Report claimed that there was a coverage gap in the north Berkeley 11 hills based on only one metric, RSRP. See A.R. 1583-84. Robert Beegle, BRN’s expert in the 12 fields of forensic cell phone analysis, cell tower coverage analysis, radio frequency propagation 13 mapping, and drive testing, TTII 356, 365-66, testified that using RSRP alone to determine 14 coverage is not “a reliable method in and of itself to be able to determine coverage issues or lack 15 of coverage issues.” According to Beegle, RSRP should instead be “coupled with drive tests” and 16 be considered with “noise factors, interference factors,” and speed testing “to be more 17 demonstrative of the actual user experience in the field,” per FCC guidance. TTII 380. For his 18 part, although Conroy testified that “the significant gap is identified through the RSRP,” TTI 179, 19 he also testified that RSRP is “one of the parameters” that “is used to define coverage in LTE 20 networks,” suggesting that other parameters also play a role in determining network coverage. 21 TTI 106-07. Thus, Verizon did not establish that considering RSRP alone is sufficient to assess 22 network coverage. 23 In addition to the problems with the 2019 drive test results described above, Conroy’s own 24 testing and analysis of Verizon’s network conditions in 2023 provide little, if any, support for the 25 conclusions in the Kharaba Report.6 With respect to the Kharaba Report’s analysis of the 26 purported capacity gap, it examined only one KPI—ASEU—and the ASEU Capacity Charts in the 27 1 Kharaba Report do not cite sources for their usage data or include information about how or when 2 the data was obtained. Although 2023 ASEU data was available to him, Conroy chose not to 3 analyze it or compare it to Kharaba’s January 2020 ASEU data. TTII 324-25. Instead, Conroy 4 obtained four other categories of KPIs for a roughly three-month period between July and October 5 2023: dropped call rate, access failure rate, downlink utilization rate, and downlink throughput 6 utilization rate. TTII 302-03; Trial Exs. 13K. Conroy was not able to analyze or compare similar 7 KPI data for 2019 or 2020 because Verizon did not retain it. TTII 247-50. He testified that 2023 8 dropped call and access failure rates exceeded carriers’ standards and showed poor network 9 performance, TTI 150-57, but made no attempt to offer explanation or analysis as to how the 10 network’s performance in 2023 is probative of its performance in early 2020. 11 Conroy also testified that he performed his own drive test in 2023 to collect data and 12 “validate” the results of the 2019 drive test “under [his] own controlled set of circumstances,” TTI 13 84, 115-16, and that his 2023 drive test confirmed the reliability of the 2019 drive test. TTI 189- 14 90. However, the tests used different methods of data collection. Specifically, Conroy used a 15 scanner with two antennae mounted on top of a vehicle to perform his test, whereas the 2019 drive 16 test used two cell phones inside a vehicle. TTI 190, TTII 211, 217, 262.7 Conroy also testified 17 that the signal measured outside a vehicle is stronger than one measured from inside a vehicle. 18 TTI 217. Conroy did not provide testimony explaining how to compare data collected at different 19 times by different equipment. As with the 2023 KPIs, Conroy was unable to provide any 20 substantive analysis regarding how the 2023 drive test is probative of whether there was a gap in 21 Verizon’s coverage at the time the Application was denied in 2020. See TTI 125-32.8 22 7 Conroy testified that he “should have photographs” documenting the two antennas mounted on 23 the rooftop of the vehicle because he “always photograph[s] the test set-up” before a drive test. TTII 263. BRN’s counsel represented that no such photos were provided to the defense. TTII 24 263.
25 8 Conroy’s reports did not offer opinions about how the 2023 drive test results corroborate the 2019 drive test results, see TTI 116-17, 125-31, even though the court instructed the parties that it 26 “will consider evidence that post-dates the denial of [the Application] only to the extent that it is probative of that issue, i.e., whether there was a significant gap in service coverage at the time of 27 the denial of [the Application].” [See Docket No. 121.] Conroy’s opening report stated only that 1 Accordingly, Verizon’s evidence about network conditions in the gap area in 2023, including 2 Conroy’s drive test results and analysis of the KPIs, is of minimal evidentiary value in 3 determining whether a significant gap existed in 2020. 4 Verizon also failed to justify its use of -85 dBm signal strength as the necessary standard 5 for in-building coverage, -95 dBm for in-vehicle coverage, and -105 dBm for outdoor coverage in 6 the 2019 drive test. The signal strength standard is selected by the carrier; it is not mandated by 7 the FCC or by law. TTII 220, 331. Conroy testified that wireless carriers use a “link budget” to 8 determine the standard and explained that a link budget is “the mathematical calculation of how 9 much power you have between the transmitting tower and the receiver.” TTII 219. The Kharaba 10 Report described the parameters of the link budget as follows:
11 Verizon Wireless uses a 4G LTE RF link budget to calculate the maximum allowable path loss (MAPL). The link budget takes into 12 account free space loss, fading and interference margins, and equipment receiver sensitivity to calculate the maximum allowable 13 MAPL. A combination of the transmit power out of the antennas and the MAPL determine the receive signal threshold required for outdoor 14 coverage. Adding vehicle body losses to the calculation determines the receive signal threshold value for in-vehicle coverage. Similarly, 15 adding the building penetration losses to the calculation determines the receive signal threshold required for in-building coverage. 16 17 A.R. 1585. However, the Kharaba Report did not include any actual figures used by Verizon to 18 perform the foregoing link budget calculation. 19 Beegle testified that the -85 dBm in-building standard “was a new standard that [he] hadn’t 20 seen before,” and that typically -95 dBm is used as the standard for in-building coverage. TTII 21 378. Conroy similarly testified that “the typical design level that you’ll see for residential and 22 building in like a suburban environment usually is around a minus 95 dBm” and that Verizon has 23 used the -95 dBm standard in other cases, but he also testified that in areas with “high levels of 24
25 the area surrounding the Berkeley Hills area since the 2020 decision.” [Docket No. 115 (Verizon’s Expert Witness Designation and Disclosure) at ¶ 35.] Accordingly, the court sustained 26 BRN’s objections to testimony about the significance of the 2023 analysis beyond what was disclosed in Conroy’s report. TTI 131 (“I’m not going to allow [Verizon] to go into any 27 explanations that are not in this report . . . he cannot testify beyond what is in his report, which is 1 interference, you may need to have stronger signal.” TTII 220, 332-33, 335. 2 Conroy testified that Verizon informed him that the -85 dBm standard was necessary due 3 to “additional interference” in the area and that it did not give him the link budget information it 4 used to determine the -85 dBm in-building standard for the 2019 drive test. TTII 227-28, 323. 5 Conroy testified that he calculated the link budget for the area using “typical values for link 6 budgets” to evaluate whether the -85 dBm standard was appropriate in this case and concluded 7 that “what [Verizon is] suggesting is reasonable and it’s within the typical link budget standards.” 8 TTII 323. However, Conroy admitted that he had not turned over to Verizon (and thus to BRN 9 and Berkeley) the figures he used to perform the link budget calculation. TTII 323-24. His failure 10 to turn over data supporting his calculation violated Federal Rule of Civil Procedure 11 26(a)(2)(B)(ii), which provides that an expert witness report “must contain . . . the facts or data 12 considered by the witness in forming” their opinions. As BRN and Berkeley were not provided 13 with the basis for Conroy’s calculations, they were deprived of the opportunity to challenge his 14 opinions on the reasonableness of Verizon’s standard. Accordingly, the court will not consider 15 Conroy’s testimony that he independently confirmed that the -85 dBm standard for in-building 16 coverage was reasonable or his opinion that it was “within the typical link budget standards.” See 17 Fed. R. Civ. P. 37(c)(1) (“[i]f a party fails to provide information or identify a witness as required 18 by Rule 26(a) . . . the party is not allowed to use that information or witness to supply evidence . . . 19 at a trial, unless the failure was substantially justified or is harmless.”). 20 Conroy also testified that even if Verizon used -95 dBm as the standard for in-building 21 coverage, “there still would be a gap in in-building service.” However, Conroy cited only the 22 existing coverage maps as evidence supporting his opinion. He did not create a different coverage 23 map using the lower standard and Verizon did not present data supporting his conclusion. TTII 24 334-36. For his part, Beegle testified that “the complexion of the maps” would change if a -95 25 dBm standard was used for in-building coverage due to the existence of “[a]reas . . . on the 26 threshold somewhere between negative 85 and negative 95”; such areas “might go from a I don’t 27 have service or I don’t have good service to I have good service or I have excellent service or I 1 the use of -95 dBm as the standard for in-building coverage instead of -85 dBm would support a 2 finding that there was a gap in in-building service. 3 III. CONCLUSIONS OF LAW 4 A. Effective Prohibition Claim 5 The TCA states that “[t]he regulation of the placement, construction, and modification of 6 personal wireless service facilities by any State or local government or instrumentality thereof . . . 7 shall not prohibit or have the effect of prohibiting the provision of personal wireless services.” 47 8 U.S.C. § 332(c)(7)(B)(i)(II). The statute defines the term “personal wireless services” to mean 9 “commercial mobile services . . . and common carrier wireless exchange access services.” 10 “Personal wireless service facilities” means “facilities for the provision of personal wireless 11 services.” 47 U.S.C. § 332(c)(7)(C)(i), (ii). 12 The Ninth Circuit uses a two-pronged standard to evaluate an effective prohibition claim. 13 Under this analysis, the provider must show (1) “a ‘significant gap’ in service coverage” and (2) 14 “the lack of available and technologically feasible alternatives.” T-Mobile USA, Inc. v. City of 15 Anacortes, 572 F.3d 987, 995-96 (9th Cir. 2009). When evaluating an effective prohibition claim, 16 courts review the administrative record de novo without “deference to local findings.” Los 17 Angeles SMSA Ltd. P’ship v. City of Los Angeles, No. 2:16-cv-04954-FLA (SKx), 2021 WL 18 3741539, at *4 (C.D. Cal. Aug. 24, 2021) (collecting cases). As noted, the court previously 19 granted summary judgment in Verizon’s favor on the second prong of the effective prohibition 20 test, the lack of available and feasible alternatives. GTE Mobilnet, 2023 WL 2648197, at *28. 21 Accordingly, the sole issue for this claim is whether a significant gap in coverage existed in 2020, 22 at the time of the denial of the Application. 23 “The significant gap prong is satisfied ‘whenever a provider is prevented from filling a 24 significant gap in its own service coverage.’” Am. Tower Corp. v. City of San Diego, 763 F.3d 25 1035, 1056 (9th Cir. 2014) (emphasis in original) (quoting MetroPCS, Inc. v. City & Cnty. of San 26 Francisco, 400 F.3d 715, 733 (9th Cir. 2005), abrogated on other grounds in T-Mobile S., LLC v. 27 City of Roswell, Ga., 574 U.S. 293 (2015)). “[T]he relevant service gap must be truly 1 ‘dead spots’ . . .” Metro PCS, 400 F.3d at 733 & n.10. 2 “‘[S]ignificant gap’ determinations are extremely fact-specific inquiries that defy any 3 bright-line legal rule.” Sprint PCS Assets, L.L.C. v. City of Palos Verdes Ests., 583 F.3d 716, 727 4 (9th Cir. 2009) (quoting MetroPCS, 400 F.3d at 733). “[C]ourts consider a ‘wide range of 5 context-specific factors’ in assessing the significance of alleged coverage gaps, including the 6 nature and character of the area, the number of potential users in that area who may be affected by 7 the alleged lack of service, whether facilities are needed to improve weak signals or to fill a 8 complete void in coverage, whether gaps pose public safety risks, and the effects of gaps on roads, 9 highways, railways, and commercial districts.” Los Angeles SMSA Ltd. P’ship v. City of Los 10 Angeles, No. 2:16-CV-04954-FLA (SKX), 2021 WL 3741539, at *4 (C.D. Cal. Aug. 24, 2021) 11 (citing Palos Verdes Ests., 583 F.3d at 727). Other factors include whether the alleged “gap 12 covers well traveled roads on which customers lack roaming capabilities” and “drive tests.” Palos 13 Verdes Ests., 583 F.3d at 727. “Inadequate or unreliable in-building service can be sufficient to 14 show the existence of a significant gap in coverage,” Los Angeles SMSA, 2021 WL 3741539, at *4 15 (citing T-Mobile W. Corp. v. City of Huntington Beach, No. 2:10-cv-02835-CAS, 2012 WL 16 4867775, at *4 C.D. Cal. Oct. 10, 2012)), as can “a gap in a provider’s in-home coverage that 17 consists of more than a few isolated pockets of inadequate in-home coverage.” T-Mobile W. Corp. 18 v. City of Agoura Hills, No. 2:09-cv-09077-DSF (PJWx), 2010 WL 5313398, at *8 (C.D. Cal. 19 Dec. 20, 2010). 20 Verizon’s Application described the “nature and character of the area” in the Berkeley hills 21 as “primarily residential and wooded,” but the precise location of the alleged gap area has been a 22 moving target. Similarly, Verizon has not clearly identified the “number of potential users in that 23 area who may be affected by the alleged lack of service.” The Application described the gap area 24 as “the Berkeley Hills area, especially along Euclid Avenue north of EBMUD’s Berryman 25 Reservoir.” A.R. 5. While the Kharaba Report provided slightly more detail, it was still 26 somewhat imprecise in describing the gap area as “areas near Euclid Avenue between Keith 27 Avenue and Vine Lane.” A.R. 1585. Kharaba also asserted that the proposed facility would 1 residents.” A.R. 1585. As noted above, Kharaba did not provide the underlying factual basis for 2 these statements. In contrast, Conroy testified that the polygon-shaped gap encompasses over 1.1 3 square miles9 and includes approximately 7,346 people based on 2010 U.S. Census data. TTI 112- 4 13, 158. Verizon did not offer evidence addressing or explaining the discrepancies between these 5 estimates of the size of the gap area and population therein. It also did not offer evidence about 6 the number of Verizon or wireless users located in the area who might be affected by the alleged 7 coverage gap. 8 With respect to whether the proposed facility is “needed to improve weak signals or to fill 9 a complete void in coverage,” Verizon has never taken the position that there is a “complete void 10 in coverage” in the gap area. Instead, the Application described the objective of “improv[ing] 11 coverage in the Berkeley Hills area.” A.R. 5. Verizon presented no evidence that the purported 12 gap poses public safety risks. It also presented no evidence about the daily traffic through the 13 area, even though Conroy testified that he has included such information in other reports. TTII 14 254-55. In fact, there is no evidence that the gap area is a high-traffic area; it is largely made up of 15 single-family residential properties and open park space. It has no commercial development and 16 lacks any well-traveled commuter highways or railways. A.R. 132, 133, TTII 255. 17 Drive tests are another factor courts consider in evaluating the existence of a significant 18 gap. There are numerous problems with the 2019 drive test, which purports to show “inadequate 19 4G LTE service coverage” in the gap area. A.R. 1582. Verizon did not present testimony by 20 Kinney about how he performed the 2019 drive test, nor did it submit evidence regarding his 21 experience or qualifications to perform the test. Further, it is not clear that Verizon produced all 22 of the 2019 drive test data. As discussed, even though the 2019 drive test data was collected using 23 two phones, Conroy initially testified that he received data from only one phone and could not 24 explain the missing data from the second phone. Although Conroy later offered a new explanation 25 that the data he received “could have been combined data between the two phones,” the court 26 9 Conroy’s opening and rebuttal reports did not provide the basis for his 1.1 square miles 27 calculation. Verizon attempted to introduce an exhibit at trial that substantiated the calculation to 1 affords this testimony no weight given the circumstances under which it was given and the fact 2 that it is entirely speculative. 3 Additionally, the Kharaba Report based its determination that there was “inadequate 4G 4 LTE service coverage” on RSRP only, and Verizon did not establish that consideration of RSRP 5 alone is sufficient to assess network coverage. As discussed, Conroy testified that RSRP is “one 6 of the parameters” that “is used to define coverage in LTE networks,” TTI 106-07, and Beegle 7 testified that RSRP is not “a reliable method in and of itself to be able to determine coverage 8 issues or lack of coverage issues” and should instead be combined with other factors “to be more 9 demonstrative of the actual user experience in the field.” TTII 380. 10 Verizon also did not adequately support its use of -85 dBm signal strength as the standard 11 for assessing in-building coverage. There is evidence that Verizon has used -95 dBm as the 12 standard for in-building coverage in other cases; this is 10 times lower than the -85 dBM standard 13 Verizon used to support its Application in this case. Moreover, Verizon failed to produce the 14 underlying link budget information to provide factual justification for the -85 dBm standard. With 15 respect to Conroy’s testimony that “there would still be a gap in in-building service” even if 16 Verizon used -95 dBm as the standard, he did not create a separate map depicting the effect of a 17 different standard or otherwise present data supporting his opinion. 18 Ultimately, Verizon rested on the 2019 drive testing and Conroy’s 2023 analysis to prove 19 the existence of a significant gap in coverage in 2020, but failed to establish that the 2019 test 20 results were reliable or that the 2023 analysis is probative of the conditions in 2020. It did not 21 offer other credible evidence supporting the existence of a significant gap. Accordingly, the court 22 concludes that Verizon has failed to prove by a preponderance of the evidence that there was a 23 significant gap in coverage in the relevant area in 2020. Accordingly, the court grants judgment in 24 BRN and Berkeley’s favor on the effective prohibition claim. 25 B. Remedy for the Failure-to-Act Claim 26 The sole remaining issue is Verizon’s remedy for the City’s failure to act on Verizon’s 27 application “within a reasonable period of time” after it was filed, in violation of 47 U.S.C. § 1 question “until the effective prohibition claim is resolved, because resolution of the effective 2 || prohibition claim in Verizon's favor would effectively moot the relief requested on the failure-to- 3 act claim.” GTE Mobilnet, 2023 WL 2648197, at *29. As the effective resolution claim has now 4 || been resolved in the City and BRN’s favor, the parties shall brief the issue of the appropriate 5 || remedy for the City’s failure to timely act on Verizon’s application. Verizon shall file a brief 6 addressing this issue that does not exceed five pages by no later than November 6, 2024. The City 7 and BRN shall each file a brief in response that does not exceed five pages (or may file one 8 combined brief that does not exceed five pages) by no later than November 15, 2024. The court 9 will take the matter under submission upon completion of the briefing. 10 || Iv. CONCLUSION 11 For the foregoing reasons, Berkeley and BRN are entitled to judgment in their favor on 12 || Verizon’s effective prohibition claim. Verizon’s five-page brief regarding the appropriate remedy
13 || for the failure-to-act claim is due by November 6, 2024. The City and BRN’s five-page brief(s) in
14 || response is due by November 15, 2024. KS DISTR I> 15 Sy CS 2 IT IS SO ORDERED ey □□ 17 || Dated: October 25, 2024 Ss \" is □□ J = Z\ \ Sood Ry. BYES 19 O ne eC D> 0 © vp iPENETo Oe Ry oy □□□ 20 □□ LOY gS 71 DISTRICS 22 23 24 25 26 27 28