ATR-2016-01708_For_Release_Final_(4).pdf

AEROSPACE REPORT NO.

ATR-2016-01708

Radio Frequency Interference Monitoring System Non-Recurring Engineering Task

August 11, 2016

Brian R. Castello1, Matthew A. Clark1, Jinyoung Jang1, Peter G. Kim1, Andrew M. McAllister1, James P. Roberts2, Konstantin N. Tarasov3, Louis H. Sacks3, Lan Xu1, and Laurence F. Zapanta1 1Communication and Network Architectures Subdivision, Communications and Cyber Division 2System Development, Operations and Protection Directorate, NASA and Civil Space Division 3Communication Systems Implementation Subdivision, Communications and Cyber Division

Prepared for:

National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite, Data, and Information Service 1325 East-West Highway Silver Spring, MD 20910

Contract No. DG-133E-13-CQ-0009

Authorized by: Vaeros

The content of this report has been altered from its original delivery state. Portions of the report that were not relevant to the RFIMS procurement have been removed by OSGS so that this report could be released as Government Furnished Information (GFI). Apart from the redacted content, this report has not been further altered.

http://pages.aero.org/aeroreports/author-overview/ http://pages.aero.org/aeroreports/author-overview/ i

Acknowledgments

The authors of this study would like to take the opportunity to extend our thanks and appreciation to the following people for their support during the course of this task: Steven Grippando, the RFIMS Program Manager; Jim “Chili” Lindsay, Steve’s right-hand man; Girija Metha, a member of the RFIMS program team for her positive, optimistic attitude, and insightful questions; and Neil Wyse, Systems Director, Ground Engineering Directorate, The Aerospace Corporation, for his guidance and support navigating both the technical and financial waters involved in accomplishing this task.

ii

Abstract

This non-recurring engineering (NRE) task was commissioned by the National Environmental Satellite Data Information Service, Office of Satellite Ground Services in support of the Department of Commerce (DOC), Radio Frequency Management Department (RFMD). As a result of the Advanced Wireless System 3 auction, there is now a requirement for NOAA to share the 1695–1710 MHz band with wireless carriers. As a part of the actions necessary to support successful sharing, a monitoring capability must be implemented at specific NOAA and National Weather Service (NWS) locations across the United States and its protectorates. This task examines the existing radio frequency (RF) environment surrounding three of the protected NOAA facilities.

iii

Executive Summary

“Presidential Memorandum: Unleashing the Wireless Broadband Revolution,” dated June 28, 2010, laid out a plan to make available a total of 500 MHz of federal and nonfederal spectrum for both mobile and fixed wireless broadband use over the next 10 years. A section of L band (1695–1710 MHz) was identified for potential sharing with the commercial broadband industry where National Oceanic and Atmospheric Administration (NOAA) was the primary user of this L band for downlinking data from polar operational environmental satellites (POES). Advanced Wireless Services 3 (AWS-3) that closed on January 29, 2015, led NOAA to share the 1695–1710 MHz band with commercial wireless 4th generation–long-term evolution (4G-LTE) carriers. This band was subdivided into two unpaired blocks: A1 block (1695–1700 MHz) and B1 block (1700–1710 MHz), where carriers could perform uplink (user equipment to base station) operations. However, spectrum-sharing is an uncovered ground and presents many challenges. In order to recommend a viable solution, there needs to be a starting point or a “baseline.” Phase 1 of this study will focus on creating a baseline reference.

The Aerospace Corporation, under the direction of NOAA/National Environmental Satellite, Data, and Information service (NESDIS)/Office of Satellite Ground Services (OSGS), developed a test plan to create a “baseline” for these impacted sites. There are a total of 17 NOAA sites/facilities within 15 Protection Zones that were identified as being potentially impacted by the AWS-3 auction. Facilities within two Protection Zones were chosen to effectively represent the entire range of sites/facilities. The Wallops Command and Data Acquisition Station (WCDAS) located in Wallops Island, Virginia, was chosen to represent the NOAA command and data acquisition stations. The Atlantic Oceanographic and Meteorological Laboratory (AOML) and National Hurricane Center (NHC) were chosen to represent smaller sites that are scattered throughout the United States. The Aerospace Corporation conducted two site surveys to create a baseline reference.

iv

Contents

1. Introduction

1.1 Background

1.2 Objectives

1.3 Rationale to Select Candidate Sites

1.4 Task Status

1.5 AWS-3 and AWS-1 Description

1.6 The Protection Zones as Submitted in the Department of Commerce Transition Plan

1.6.1 Sites That Capture POES Downlink Data Only

1.6.2 Sites That Capture GOES Downlink Data Only and/or Provide Backup

Capability

1.6.3 Dual Capability Sites

1.6.4 Andersen AFB, Guam

2. SOAP Analysis

2.1 Introduction

2.2 SOAP Runs for POES Visibility at Each Site

2.3 Actual Average Number of POES Passes per Day at Each Site

2.4 Passes Per Day in Each of the Auctioned Bands

2.4.1 AWS-3 A1 Block (1695–1700 MHz)

2.4.2 AWS-3 B1 Block (1700–1710 MHz)

3. Recap of Commerce Spectrum Management Advisory Committee (CSMAC) Working Group 1 (WG-1) Results

3.1 Assumptions/Simulation

3.2 Interference Protection Criteria (IPC) and Protection Zones

3.2.1 Wallops Island, Virginia, Results

3.2.2 Miami AOML, Florida, Results

3.2.3 50/50 ITM Model

4. Site Surveys and Data Analysis

4.1 Test Plan Summary

4.1.1 Test Plan

4.1.2 Data Collection Times

4.1.3 Equipment Configuration

4.1.4 Data Collection Summary

4.2 Data Analysis

4.2.1 Approaches for Analysis

4.2.2 POES

4.2.3 GOES

4.3 Interference

4.3.1 Miami Interference

4.3.2 AWS-1 Interference

4.4 Noise Level Summary

5. Summary Observation

5.1 Operational Observations/Operations Concepts

5.1.1 Wallops Island, Virginia

5.1.2 National Hurricane Center (NHC) (Miami, Florida)

5.1.3 AOML (Miami, Florida)

5.2 Environmental Observations

vi

Figures

Figure 1. Timeline for RFIMS Phase 1 site survey Figure 2. EA market map of Wallops Island, Virginia, and Miami, Florida [3] Figure 3. Protection zone and EA regions for Wallops Island, Virginia Figure 4. Protection zone and EA regions for Miami, Florida Figure 5. CMA market map of Wallops Island, Virginia, and Miami, Florida [2] Figure 6. Total visibility at 5° vs. time of day (Miami, Florida, and Wallops Island, Virginia) Figure 7. Actual passes for Miami (October 22 to November 9) Figure 8. A1 Visibility vs. time of day (Miami AOML, Florida, and Wallops Island, Virginia) Figure 9. B1 Visibility vs. time of day (Miami AOML, Florida, and Wallops Island, Virginia) Figure 10. Data collection times Figure 11. Equipment configuration Figure 12. Example signal-to-noise power measurements Figure 13. Results of steps 9 and 11 through 13 Figure 14. Average noise level for Miami, Florida Figure 15. NOAA-15 signals collected on November 17, 2015, in Wallops Island, Virginia Figure 16. Wallops NOAA 19 signal margin vs. EL/AZ Figure 17. GOES West signal collected on December 7, 2015, in Miami, Florida Figure 18. GOES East PDR Eb/No margin Figure 19. Spectrum capture 1665–1730 MHz Figure 20. Range of interference scan for Miami Figure 21. Eb/No margins for NOAA-19 pass at AOML Figure 22. AWS-1 spectrum vii

Tables

Table 1. Location and Site Capability Table 2. Deliverables and Schedule Table 3. AWS-3 Summary [2] Table 4. AWS-3 EA Licensee Information for Wallops and Miami [5] Table 5. AWS-1 Summary [2] Table 6. AWS-1 CMA Licensee Information for Wallops and Miami [5] Table 7. List of Protected Sites from the Transition Plan Table 8. Equipment List for Data Collection Table 9. POES Collection (Passes) Table 10. GOES Collection (Data Sets) Table 11. POES, GOES, and MetOp Characteristics Table 12. Details of Downlinked Signals Table 13. Required Eb/No for Different Signals Table 14. Highest Noise Levels Predicted vs. Simulation vs. Measured [1]

1. Introduction

1.1 Background

One of the National Telecommunications and Information Administration (NTIA) Commerce Spectrum Management Advisory Committee (CSMAC) Working Group 1 (WG-1)’s final report recommendations stipulated that procedures for implementing an on-going realtime monitoring capability to ensure interference power spectral density (IPSD) limits are not exceeded and that commercial operations can be adjusted immediately if they are exceeded. Per the report, “significant progress was made to refine interference analysis and develop a deeper understanding of the issues and options available for maximizing access to the spectrum for commercial services while protecting incumbent federal operations in the 1695–1710 MHz and the adjacent 1675–1695 MHz bands” [1]. In order to achieve this, a robust monitor capability was determined to be necessary. The system would be required to prevent harmful interference received by National Oceanic and Atmospheric Administration (NOAA) ground stations, emanating from commercial wireless mobile system transmitters (e.g., user equipment [UEs] in 3GPP LTE networks) in the 1695–1710 MHz band (the downlink band for numerous NOAA ground stations) pursuant to a prospective spectrum licensing of wireless mobile users in this band. For brevity, the wireless mobile transmitters will be referred to as UEs and service providers as wireless carriers or carriers. The monitor capability was approached from the position it must operate 24/7/365 for 10 years at each of the identified NOAA ground stations in the United States and Possessions (US&P). Effective monitoring and enforcement mechanisms are critical to sharing in the 1695–1710 MHz band.

1.2 Objectives

Phase 1 objectives are to (1) perform a comprehensive study of all aspects that affect the successful continued operation of the NOAA downlinks in 1695–1710 MHz band as a result of spectrum-sharing with wireless carriers.

1.3 Rationale to Select Candidate Sites

There are a total of 15 locations with the possibility of 17 NOAA/National Weather Service (NWS) facilities being impacted. Three facilities were selected to be representative of all 17 sites. Wallops Command and Data Acquisition Station (WCDAS) (Wallops Island, Virginia) represents other large data acquisition sites such as Fairbanks, Alaska, and Suitland, Maryland. The National Hurricane Center (NHC) and Atmospheric Oceanographic Meteorological Laboratory (AOML) in Miami, Florida, represent the remaining sites located throughout the United States. In addition, the 2010 census population also affected the decision. Miami had the second largest population in contrast to Wallops Island, where the population was less dense. The details regarding 2010 census population are provided in the appendices.

1.4 Task Status

A test plan was developed to facilitate an efficient data capture of POES and geostationary operational environmental satellite (GOES) in order to create a baseline of the RF environment in the geographic area surrounding the NOAA sites. Table 1 shows the locations and the systems that were available for each particular site for data collection.

Table 1. Location and Site Capability

Site Name Location GOES POES Wallops Command and Data Acquisition Station (WCDAS) Wallops Island, VA X X Atlantic Oceanographic and Meteorological Laboratory (AOML) Miami, FL X National Hurricane Center (NHC) Miami, FL X

Table 2 shows the deliverables and due dates for the Phase 1 tasks.

Table 2. Deliverables and Schedule

Deliverable Due Date Wallops Preliminary Briefing December 4, 2015 Miami Preliminary Briefing December 18, 2015 Preliminary Final Out Brief January 26, 2016 Final Out Brief for Phase 1 Phase II Initial Out Brief

April 27, 2016 April 27, 2016

Aerospace Technical Report June 30, 2016

Figure 1 shows the overall timeline of the site survey performed for Phase 1 of RFIMS.

Figure 1. Timeline for RFIMS Phase 1 site survey.

1.5 AWS-3 and AWS-1 Description

The Advanced Wireless Service 3 (AWS-3) auction began on November 13, 2014, and ended on January 29, 2015, to allow commercial users access to frequency bands that had previously been allocated for government use only. Table 3 is a list of the frequency bands auctioned during the AWS-3, which is used for POES in-band analysis. The AWS-3 band in general, 1695–1710 MHz and 1755–1780 MHz, is used for uplink, while 2155–2180 MHz is used for downlink.

Test Plan Development (8/23/2015 – 10/25/2015)

Wallops Site Coordination

(10/26/15)

Miami Site Coordination

NHC

(11/12/15)

AOML

(11/13/15)

Wallops Site Survey and

Measurement (11/17/15 – 11/19/15)

Miami Site Survey and

Measurement

NHC

(12/7/15)

AOML

(12/8/15 – 12/10/15)

Wallops Preliminary

Briefing (12/4/15)

Miami Preliminary

Briefing (12/18/15)

Preliminary Phase I Out

Brief (1/26/16)

Table 3. AWS-3 Summary [2]

Block Frequencies Bandwidth Pairing Geographic Area Type

Number of Licenses

A1 1695–1700 MHz 5 MHz Unpaired EA 176 B1 G H I J

1700–1710 MHz 1755–1760/2155–2160 MHz 1760–1765/2160–2165 MHz 1765–1770/2165–2170 MHz 1770–1780/2170–2180 MHz

10 MHz 10 MHz 10 MHz 10 MHz 20 MHz

Unpaired 2 x 5 MHz 2 x 5 MHz 2 x 5 MHz 2 x 10 MHz

EA

CMA

EA

EA

EA

Figure 2. EA market map of Wallops Island, Virginia, and Miami, Florida [3].

Licensee information for the site survey locations for AWS-3 economic areas (EA) are shown in Table 4 where Northstar Wireless is one of Dish Network’s affiliates. EA is defined as “The Economic Area service areas are based on the Economic Areas delineated by the Regional Economic Analysis Division, Bureau of Economic Analysis, U.S. Department of Commerce February 1995 (1-172).” [4]

Table 4. AWS-3 EA Licensee Information for Wallops and Miami [5]

Market Licensee Channel Block Associated

Frequencies (MHz) BEA014 Northstar Wireless, LLC A1 1695–1700

BEA014

BEA031

BEA031

Northstar Wireless, LLC Northstar Wireless, LLC

2014 AWS Spectrum Bidco Corporation

B1 A1 B1

1700–1710 1695–1700 1700–1710

The relationship between Protection Zones and individual sites are shown in the following figures. For NOAA sites that have GOES and POES in different facilities, the Protection Zone is centered at the POES facility and two Protection Zones are combined into one larger Protection Zone. For example, Miami AOML and Miami NHC is such a case and both are at BEA014, while Wallops Island is located at

BEA031.

Figure 3. Protection zone and EA regions for Wallops Island, Virginia.

Figure 4. Protection zone and EA regions for Miami, Florida.

The advanced wireless services (AWS-1) auction began on August 9, 2006, and ended on September 18, 2006. For AWS-1, 1710–1755 MHz is used for uplink, while 2100–2155 MHz is used for downlink; more details are shown in Table 5. Regional Economic Area Groupings (REAG) is defined as “The Regional Economic Areas (REA) also known as the REAG for Wireless Communication Service (WCS) which were created by Commission staff are an aggregation of Major Economic Areas (MEA) into 12 regions.”

[4]

Table 5. AWS-1 Summary [2]

Block Frequencies Bandwidth Pairing Geographic Area Type

Number of Licenses

A 1710–1720 and 2110–2120 MHz 20 MHz 2 x 10 MHz CMA 734 B C D E F

1720–1730 and 2120–2130 MHz 1730–1735 and 2130–2135 MHz 1735–1740 and 2135–2140 MHz 1740–1745 and 2140–2145 MHz 1745–1755 and 2145–2155 MHz

20 MHz 10 MHz 10 MHz 10 MHz 20 MHz

2 x 10 MHz 2 x 5 MHz 2 x 5 MHz 2 x 5 MHz 2 x 10 MHz

EA

EA

REAG

REAG

REAG

The AWS-1 A1 block is above the AWS-3 B1 block and is considered out of band for this study. Out of band services can also have an effect on POES in-band and, therefore, an examination and nominal analysis regarding AWS-1 were made. AWS-1 services for the two selected NOAA sites were identified using the Federal Communications Commission (FCC) website. “The Cellular Market Areas (CMAs) were created from the Metropolitan Statistical Areas (MSAs) defined by the Office of Management and Budget (1-305)” [4] and CMA map indicates what cellular market NOAA’s sites are located in and from Figure 6 below, WCDAS is located at CMA 692 Virginia 12 – Caroline, while Miami NHC and Miami AOML are located at CMA 12 Miami – Fort Lauderdale – Hollywood, Florida.

Figure 5. CMA market map of Wallops Island, Virginia, and Miami, Florida [2].

The license for CMA in Wallops Island is owned by New Cingular Wireless PCS, LLC, subsidiary of AT&T, and the license for the CMA in Miami is owned by Cellco Partnership, which is the legal name for Verizon. However, user equipment (UEs) from other licensee/carrier can utilize roaming capabilities and/or other network services if roaming agreements exist among the licensees. Licensee information for the site survey locations for AWS-1 CMA are shown in Table 6.

Table 6. AWS-1 CMA Licensee Information for Wallops and Miami [5]

Market Licensee Channel Block Associated Frequencies (MHz) CMA692 New Cingular Wireless PCS, LLC A 1710–1720 2110–2120 CMA012 Cellco Partnership A 1710–1720 2110–2120

1.6 The Protection Zones as Submitted in the Department of Commerce Transition Plan

Fifteen Protection Zones were identified around NOAA/National Environmental Satellite, Data, and Information Service (NESDIS) and NWS facilities that capture downlink data from either GOES or POES. The respective downlink frequencies are 1675–1695 MHz for the GOES downlink, and 1695– 1710 MHz for the POES downlink. The Protection Zones (PZs) and associated sites/facilities are identified in the Coordination Procedures in the 1695–1710 MHz and 1755–1780 MHz Bands, GN Docket No. 13-185, DA 14-1023, July 18, 2014 [6], as depicted in Table 7 below.

Table 7. List of Protected Sites from the Transition Plan

State Facility Latitude Longitude Radius km

(up to 20 dBm) Radius km

(above 20 dBm) AK Barrow 71° 19’ 22” 156° 36’ 41” 35 60 AK Elmendorf AFB 61° 14’ 08” 149° 55’ 31” 98 129 AK Fairbanks 64° 58’ 22” 147° 30’ 02” 20 45 CA Monterey 36° 35’ 34” 121° 51’ 20” 76 101 CO Boulder 39° 59’ 26” 105° 15’ 51” 2 27 FL Miami/AOML/NHC 25° 44’ 05” 080° 09’ 45” 51 76 GU Andersen AFB 13° 34’ 52” 144° 55’ 28” 42 67 HI Hickham AFB 21° 19’ 18” 157° 57’ 30” 28 53

MD Suitland/Greenbelt 38° 51’ 07” 076° 56’ 12” 98 123 MO Kansas City 39° 16’ 40” 094° 39’ 44” 40 65 MS Stennis 30° 21’ 23” 089° 36’ 41” 57 82 OK Norman 35° 10’ 52” 097° 26’ 21” 3 28 PRI Guaynabo 18° 25’ 26” 066° 06’ 50” 48 73 VA Wallops Island 37° 56’ 45” 075° 27’ 45” 30 55 WV Fairmont 39° 26’ 02” 080° 11’ 33” 4 29

1.6.1 Sites That Capture POES Downlink Data Only

The sites/facilities that supported the POES downlink capture are:

• Barrow, Alaska

• Monterey, California

• Ford Island, Hawaii (falls within the Hickham AFB PZ)

• Miami, Florida (AOML)

Barrow is scheduled by the NOAA Fairbanks Command and Data Acquisition Station (CDAS).

Monterey, Ford Island and Miami sites are considered direct broadcast sites and fall under NESDIS purview. Guaynabo, Puerto Rico, is a future POES capability. The sites operate fairly autonomously. In the case of the AOML, the equipment is monitored and maintained by a technician remotely. The technician does not monitor the system in real-time and can only detect problems such as data degradation after the fact. The operational aspects of the Monterey, California, and Ford Island, Hawaii, are thought to be similar to the AOML.

1.6.2 Sites That Capture GOES Downlink Data Only and/or Provide Backup Capability

The GOES-only sites are primarily NWS facilities that house the end users of the GOES downlink data captured at the respective sites or GOES backup command and data acquisition facilities.

• Boulder, Colorado (GOES backup)

• Miami, Florida (NHC)

• Greenbelt, Maryland (GOES backup)

• Kansas City, Missouri (Aviation Weather Center)

• Stennis, Mississippi (National Data Buoy Center)

• Norman, Oklahoma (National Storm Prediction Center)

• Fairmont, West Virginia (GOES backup)

1.6.3 Dual Capability Sites

Dual capability sites are those locations that can capture both POES and GOES downlink data. They include:

• Anchorage, Alaska (NWS; within the Elmendorf Air Force Base [AFB] PZ)

• Fairbanks, Alaska (CDAS)

• Suitland, Maryland (NOAA Satellite Operations Facility [NSOF]; POES capability not in use, awaiting upgrade)

• Wallops Island, Virginia (CDAS)

1.6.4 Andersen AFB, Guam

Andersen AFB is identified in the Department of Commerce (DOC) transition plan as a joint DOC/Department of Defense (DOD) site. The Guam NWS office recently decommissioned its GOES antenna and no longer captures GOES downlink data. Currently, a new antenna to capture the Himawari-8 downlink is being installed and tested for this new capability.

2. SOAP Analysis

2.1 Introduction

This section explains the impact of the user equipment (UEs) on the National Oceanic and Atmospheric Administration’s (NOAA’s) ability to collect downlink satellite data. One of the major factors in determining how the UEs will impact the NOAA downlink is how often satellites are in view of the respective ground stations and how often the stations are collecting downlinked data. This section describes the process for determining the impact of satellite visibility on overall satellite operations.

The Aerospace Corporation’s Satellite Orbit Analysis Program (SOAP) was used to assess this factor.

NOAA satellite’s two-line element (TLE) sets were used to provide an accurate prediction of the NOAA satellite locations and orbit durations. Custom Python software was used for plotting and visualization.

While predicting the satellite visibility for any given site provides information on the number of times it could be supported in a day, the average number of actual supports is usually based upon the need for specific data and resource limitations.

As mentioned in section 1, the spectrum was auctioned in two blocks. A1 Block occupies 1695– 1700 MHz and B1 Block occupies 1700–1710 MHz. To simulate the potential impact periods of the long-term evolution (LTE) interference on the NOAA downlinks, Aerospace assessed the blocks of frequencies independently. NOAA-19 would be impacted by carrier operations in the A1 block, while NOAA-15, NOAA-18, MetOp-A and MetOp-B would be impacted by carrier operations in the B1 block.

During the SOAP analysis, the separate blocks were assessed and the potential impact to NOAA downlinks was defined. Also, when the frequency blocks are separated it becomes possible to allocate periods of access that are predictable and finite. These allocated time periods are helpful in assisting carriers with defining periods of reduced transmit power from the UEs or enabling the UEs to operate within the Protection Zones (PZs) while ensuring the quality of the NOAA downlink data.

2.2 SOAP Runs for POES Visibility at Each Site

As shown in Figure 6, pass times are periodic and predictable. Figure 6 also shows the average pass times in a given day. These passes were collected over a period of 30 simulated days. Miami and Wallops Island have similar periods of collection. This is due to the close longitudes and latitudes of each location.

As seen in the Appendix A, locations in different longitudes (same latitudes) will have similar periods but different times, while different latitudes (same longitudes) will have different periods but similar times.

Figure 6. Total visibility at 5° vs. time of day (Miami, Florida, and Wallops Island, Virginia).

2.3 Actual Average Number of POES Passes per Day at Each Site

When observing operations at Wallops and Miami, data from multiple satellite passes was taken from the NOAA and Meterological Operational Satellite Program (MetOp) polar orbiters, and from other satellites such as JASON-2. Figure 7 shows the number of passes (NOAA and MetOp) that the Atlantic Oceanographic and Meterological Laboratory (AOML) in Miami, Florida, had visibility of and could collect from during the period from October 22 to November 9, 2015. Each satellite (represented in a separate color) follows a periodic and predictable pattern. For example, NOAA-19 (orange) passes over Miami everyday around 15:00. On average, the Miami AOML supports approximately 14 passes a day.

Figure 7. Actual passes for Miami (October 22 to November 9).

2.4 Passes Per Day in Each of the Auctioned Bands

Potential LTE interference can be further reduced if the analysis is performed specifically addressing the separate frequency blocks. Although the different NOAA and MetOp satellites may utilize one or more frequencies during a pass (e.g., a different frequency for the high resolution picture transmission [HRPT], the local area coverage [LAC], and/or global area coverage [GAC]). Any radio frequency interference (RFI) in one of the downlink signals can be determined by the schedule and addressed based upon the block within which it lies (i.e., the A1 or B1 block).

2.4.1 AWS-3 A1 Block (1695–1700 MHz)

Figure 8 shows that if A1 block is independently assessed, the potential impact of the LTE interference would be limited to the blocks of time. In the example below, limiting time of impact to approximately 1:00 a.m. to 5:00 a.m. and 11:00 a.m. to 2:00 p.m. greatly reduces the periods of time when the NOAA downlinks could be affected by LTE interference.

Figure 8. A1 Visibility vs. time of day (Miami AOML, Florida, and Wallops Island, Virginia).

2.4.2 AWS-3 B1 Block (1700–1710 MHz)

B1 block has more satellites; however, it is still possible to predict likely times of interference. For example, Figure 9 for Miami AOML and Wallops Island POES passes, the time for the greatest probability of impact is from 4:00 a.m. to 11:00 a.m. and 3:00 p.m. to midnight.

Figure 9. B1 Visibility vs. time of day (Miami AOML, Florida, and Wallops Island, Virginia).

3. Recap of Commerce Spectrum Management Advisory Committee (CSMAC) Working Group 1 (WG-1) Results

3.1 Assumptions/Simulation

CSMAC WG-1 was tasked with developing recommendations for use of the 1695–1710 MHz band for commercial services while protecting Federal Meteorological Ground stations from harmful interference.

WG-1 provided refined long-term evolution (LTE) system parameters that more accurately reflect real world deployment scenarios and reviewed operating parameters of federal systems affected by commercial operations in the 1695–1710 MHz band. The work of this committee resulted in agreed LTE technical parameters for analysis that more accurately depict the real-world operation of LTE networks and how to apply the parameters to interference analysis.

The output of the technical WG includes refined mobile operating parameters that more closely represent real operations including:

• Cumulative distribution function (CDF) of total effective isotropic radiated power (EIRP) per scheduled mobile.

- The dynamic range of mobile power is from -30 dBm to 20 dBm, and the 42 types of power levels are defined for urban/suburban and rural topologies.

• Assumed number of scheduled (transmitting) mobile per sector—it was based on the LTE channel bandwidth.

- 3 user equipment (UEs)/5 MHz, 6 UEs/10 MHz, 9 UEs/15 MHz

• Assumed inter-site distance (ISD)

- 1.732 km for urban and suburban

- 7 km for rural

• Antenna parameters

- Wallops Islands, Virginia: UE antenna height 1.5 m, National Oceanic and Atmospheric Administration (NOAA) receiver antenna height 17 m, gain 43.1 dBi, minimum operating elevation angle 14 degree

- Miami Atlantic Oceanographic and Meterological Laboratory (AOML), Florida: UE antenna height 1.5 m, NOAA receiver antenna height 15 m, gain 29.5 dBi, minimum operating elevation angle 5 degrees

• Requirements for unwanted emissions—out-of-band (OOB) emission and spectrum emission mask (SEM)

• 50/50 irregular terrain model (ITM) propagation model and related parameters

- Refer to the appendices for more information on ITM 50/50 model

• Ground station antenna modeling including pattern, gain, etc. (ITU-R F.1245-1 model)

The following common and site-specific parameters are used to calculate the Protection Zone distance.

• The common parameters are:

- Operating frequency

- NOAA ground station’s antenna pattern

- Propagation model

- Same topology (urban/suburban/rural)

- LTE system-related parameters such as UE distribution in suburban/urban and rural, number of UEs per base station, and UE power curves in urban/suburban and rural per site

• The site-specific parameters are:

- Interference protection criteria (IPC) level per site

- NOAA receiver’s antenna height, gain, elevation angle, azimuth angle per site. The worst-case azimuth was selected based on population density and radio frequency (RF) path loss.

Additional details are provided in the appendices.

3.2 Interference Protection Criteria (IPC) and Protection Zones

IPC was defined from National Telecommunications and Information Administration (NTIA) as “A relative or absolute interfering signal level defined at the receiver input, under specified conditions, such that allowable performance degradation is not exceeded. This is usually defined as an absolute interference level, interference-to-noise power ratio (I/N), or carrier-to-interfering signal power ratio (C/I).” [7] Allowable performance degradation is defined as, “Agreed upon degradation in the performance of a radio communication system due to interference that is at a sufficiently low level such that the performance capabilities are not significantly compromised. This is often defined in terms of a percentage change in a key performance measure such as voice circuit noise, outage time, or target detection probability.” [7] Equation and details for IPC can be found in appendix 7 of the CSMAC WG-1 final report.

Protection zones are defined for UE transmitters at or below 20 dBm and for UE transmitters above 20 dBm and up to a maximum of 30 dBm. Federal Communications Commission (FCC) Public Notice DA 14-1023, “Coordination Procedures in the 1695–1710 MHz and 1755–1780 MHz Bands,” dated July 18, 2014, defined the Protection Zone radii and guidance for formal coordination that shall be initiated by the AWS-3 licensee.

3.2.1 Wallops Island, Virginia, Results

The CSMAC WG-1 final report stated an IPC level of -120.6 dBm/MHz, and worst azimuth was determined to be at 0 degrees relative to true North [1]. Protection zone distance was set to 30 km (29 km/5 MHz, 30 km/10 MHz, 30 km/15 MHz) for up to 20 dBm EIRP [7] and 55 km for power levels above 20 dBm but below 30 dBm EIRP [6].

3.2.2 Miami AOML, Florida, Results

The CSMAC WG-1 final report stated an IPC level of -124.1 dBm/MHz and worst azimuth was determined to be at 335 degrees relative to true North [1]. Protection zone distance was set to 51 km for up to 20 dBm EIRP and 76 km for power levels above 20 dBm but below 30 dBm EIRP [6]. (For different bandwidths, the Protection Zone is 46 km/5 MHz, 46 km/10 MHz, 46 km/15 MHz, and combining sites that have overlapping Protection Zones of Miami GOES site result in 51 km.)

3.2.3 50/50 ITM Model

The appropriate propagation model to be used in the aggregate interference analysis to compute the protection distances was discussed within the CSMAC working group. The industry representatives presented several propagation models such as Okumura-Hata and COST-231 models, but these models particularly overestimate propagation loss at small time percentages, and they are not appropriate for interference calculations. The WG did agree to use the ITM in the point-to-point mode for the aggregate compatibility analysis associated with the meteorological-satellite receivers. Since point-to-point mode retrieves a terrain profile based on latitude and longitude values for the transmitter and receiver, it provides a better estimate of the propagation loss. The statistical and environmental parameters used with the actual terrain profiles in calculating propagation loss are provided in the appendices.

ITM is based on electromagnetic theory and on statistical analyses of both terrain features and radio measurements for frequency bands between 20 MHz and 20 GHz, and predicts the median attenuation of a radio signal as a function of distance and the variability of the signal in time and in space. There are two modes in the ITM: ITM area mode and point-to-point mode.

ITM defines four modes of variability. The mode selected determines the meaning of the reliability and confidence values used in the model. The mode of variability can be considered the point of view for considering the meaning of reliability and confidence in the calculations. The defined modes of variability are:

1. Single message mode

2. Individual mode (selected in the WG-1)

3. Mobile mode

4. Broadcast mode

The effect of percentage of time, locations, and situations depends on the mode selected. Three dimensions of variability are:

1. Reliability (time) variability. Accounts for the fraction of time during which actual received signal strength is expected to be equal to or higher than the hourly median value. The actual signal strength at the receiver location would be expected to be above that value for 30 minutes each hour and below that value for 30 minutes of each hour. The time variability for the calculation is expressed as a percentage from 0.1 percent to 99.9 percent. This variable allows a user to specify how to handle the time variability of changing atmospherics. Entering higher percentage reliability values effectively reduces signal strength predicted and results in the reduction of Protection Zone distance.

2. Confidence (situation or probability) variability. Accounts for a probability measure imposed on the collection of all possible or conceivable propagation paths and all possible or conceivable moments of time since the model is based on many sets of propagation measurement. The situation variability for the calculation is expressed as a percentage from 0.1 percent to

99.9 percent. Similarly, entering higher percentage confidence values effectively reduces the signal strength. The predicted signal strength will be lower and results in reduction of Protection Zone distance. It is observed that this variability is more sensitive than the time (reliability) variability as described in terms of protection distance.

3. Location variability: (N/A in the point-to-point mode). Accounts for the fraction of locations where actual received signal strength is expected to be equal to or higher than the median value.

The location variability for the calculation is expressed as a percentage from 0.1 percent to

99.9 percent.

The WG selected the individual mode for interference evaluation that determines the meaning of the reliability and confidence values such as Reliability equal to Time, Confidence equal to Situation and Location (use terrain information). The 50/50 ITM means “with probability (confidence) 50 percent, the signal attenuation will not exceed the hourly median value for at least 50 percent of the time.” Note that 10/10 ITM is more conservative than 50/50 ITM in terms of interference protection criteria, so it will effectively provide larger Protection Zone distance.

4. Site Surveys and Data Analysis

4.1 Test Plan Summary

4.1.1 Test Plan

Two locations were selected to baseline the radio frequency (RF) environment and understand the operations at each location. In order to create a baseline reference, Aerospace performed RF data collection to characterize the RF environment at the selected National Oceanic and Atmospheric Administration (NOAA) facilities. The operational environment was assessed by talking to the operators, observing day-to-day operations and activities and taking part in the normal activities associated with collecting geosynchronous operational environmental satellite (GOES) or polar operational environmental satellite (POES) data.

The two locations selected were Wallops Island, Virginia, and Miami, Florida. Each location was selected in part due to its population density around the facility, the size of operation, and the facilities’ mission.

For Wallops Island, both the GOES and POES systems were housed in the same building. Miami had two separate facilities for GOES and POES. GOES data was collected at the National Hurricane Center (NHC). POES was collected at the Atlantic Oceanographic Meteorological Laboratory (AOML).

4.1.2 Data Collection Times

Data collection was set up to collect at peak and off-peak hours. In general, wireless carrier off-peak times were from 9:01 p.m. to 5:59 a.m., and peak times were from 6:00 a.m. to 9:00 p.m. Ideal minimum traffic time was determined to be 4:00 a.m. to 5:00 a.m. and maximum traffic hours was determined to be 6:00 p.m. to 9:00 p.m. In addition to collecting data to document the RF environment during off-peak and peak times, Aerospace observed the effect of peak and off-peak hours for AWS-1, and data collection times were chosen to overlap with the nominal peak and off-peak hours. Measurements for POES downlink were taken when possible to overlap the peak and off-peak times. Figure 10 shows the duration and the times that the data collection began and ended.

Wireless System’s Traffic Wallops Island, VA Miami, FL (AOML) NHC Miami, FL Non-peak hours (10:00 p.m. – 6:00 a.m.)

Min. Traffic hours (4:00 a.m. – 5:00 a.m.)

GOES

• Nov. 17, 4:50 –6:57 p.m.

• Nov. 18, 9:39 a.m.–

Nov. 19, 10:46 a.m.

POES Manual Testing

• Nov. 19, 9:26–9:42 a.m.

POES Passes

• Dec. 9, 2:01–5:45 a.m.

• Dec. 9, 10:25–11:01 p.m.

• Dec. 10, 1:50–5:30 a.m.

GOES East

• Dec. 7, 10:14–1:08 a.m.

• Dec. 7, 1:36–2:28 p.m.

GOES West

• Dec. 7, 11:25–12:52 p.m.

• Dec. 7, 2:56–3:48 p.m.

Peak hours (6:00 a.m.–10:00 p.m.)

Max. Traffic hours (6:00 p.m.–9:00 p.m.)

POES Manual Testing

• Nov. 18, 6:07–6:17 p.m.

• Nov. 18, 5:12–5:22 p.m.

POES Passes

• Nov. 17, 2:00–5:30 p.m.

• Nov. 18, 10:34 a.m.–5:05 p.m.

• Nov. 19, 8:33 a.m.–4:24 p.m.

POES Passes

• Dec. 8, 8:46 a.m.–10:26 p.m.

• Dec. 9, 6:57 a.m.–7:45 p.m.

• Dec. 10, 6:31 a.m.–3:00 p.m.

POES Manual Testing

• Dec. 8, 2:04–2:24 p.m.

• Dec. 9, 12:43–1:48 p.m.

Figure 10. Data collection times.

4.1.3 Equipment Configuration

The typical equipment configuration utilized during the data collection is shown in Figure 11. The equipment was connected through standard network interfaces. Communication between the laptop and spectrum analyzer was accomplished through Standard Commands for Programmable Instruments (SCPI). Data was stored on both the network hard drive and laptop during data collection. The RF input of the spectrum analyzer was connected directly to a patch panel, a matrix switch, or a test output port on the receiver, usually at intermediate frequency (IF) of 70 MHz or 120 MHz. This allowed for the signal to be routed to the spectrum analyzer without interrupting the operational signal flow.

Figure 11. Equipment configuration.

The equipment used is listed in Table 8. The network hub and hard drive are generic off-the-shelf components. The laptop required at least 1 gigabyte of memory and 1 terabyte of hard drive storage capacity. A Python integrated development environment (IDE) was also installed.

Table 8. Equipment List for Data Collection

Equipment Type Details Spectrum Analyzer Rhode & Schwartz FSV40-N

With the following additional options:

• B1 – ruggedized housing

• B4 – OCXO, precision reference frequency

• K40 – Phase noise measurements (not used during testing)

• K70 – Vector signal analysis (not used during testing)

Network Hub Netgear switch Network Hard Drive Seagate network hard drive Laptop Hewlett Packard (HP) laptop

4.1.4 Data Collection Summary

4.1.4.1 POES

In addition to the POES data collected during realtime events as shown in Table 9, the engineers also collected specific azimuth and elevation data between passes to understand the spectrum environment for key antenna orientations. The key antenna orientations were selected based upon the large population centers or the orientations utilized by the Commerce Spectrum Management Advisory Committee (CSMAC) Working Group 1 (WG-1). For Wallops Island, Virginia, the azimuths selected were 0°, 30° and 308°. For Miami, Florida, the azimuths selected were 300°, 310°, 315°, 320° and 330°. For each elevation at Wallops Island, data was collected at 0°, 5°, and 10°. For each elevation at Miami, data was collected at 0°, 5°, 10°, 15°, 20°, 25°, and 30°.

Table 9. POES Collection (Passes)

Satellite Wallops Miami Total

NOAA-15 3 8 9

NOAA-18 1 6 7

NOAA-19 5 10 15

MetOp-A 1 10 11 MetOp-B Note 1 8 8 Total: 57

Note 1: No Meteorological Operational Satellite Program (MetOp) B data collected at Wallops Island, Virginia.

4.1.4.2 GOES

Specific frequency ranges were collected for GOES. During the Wallops and Miami data collections, data was also collected from the 1.67 GHz to 1.72 GHz band. Data sets were typically one- to two-hour long collection periods.

Table 10. GOES Collection (Data Sets)

Satellite/Signal Wallops Miami Total (data sets) GOES-13 (GOES East)/PDR 4 5 9 GOES-15 (GOES West)/PDR 3 10 13 GOES-13 (GOES East)/LRIT 6 3 9 GOES-15 (GOES West)/LRIT 1 9 10 Total: 41

4.2 Data Analysis

4.2.1 Approaches for Analysis

Analysis was accomplished using MATLAB® for both plotting of the spectrum and providing an estimate of the available margin. Note that the Eb/No margin estimate computes the downlink section only. For some GOES signals, the downlink is assumed to be the limiting case and not the uplink.

The following steps illustrate the process for determining margin:

1. Load the collected data into MATLAB and plot signal power vs. frequency.

a. If measuring noise level, go to step 14.

b. If signal collected is at IF, convert it to RF.

2. Refer to Table 11 to find the center frequency and bandwidth of the signal. Measure the in-band signal plus noise power, which is shown in red box of Figure 12.

Table 11. POES, GOES, and MetOp Characteristics

Figure 12. Example signal-to-noise power measurements.

3. If the collected signal is from NOAA-15, -18 and -19, remove carrier power 0.71 dB from signal plus noise power.

4. Estimate noise level using the side channel as shown in the green box of Figure 12.

5. Calculate the energy per symbol to noise power spectral density ratio based on the formula:

Es/No = (Signal power + Noise Power)/ Noise Power

6. Calculate the energy per bit to noise power spectral density ratio based on formula:

Eb/No= Es/No * BW / Data Rate

Where:

BW is the Nyquist bandwidth or symbol rate (SR).

Data Rate (DR) is information rate in Table 12.

𝐵𝐵𝐵𝐵 = 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝑅𝑅𝐷𝐷𝐷𝐷𝑅𝑅 ∗

𝐵𝐵𝐵𝐵𝐷𝐷𝐵𝐵

𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆

𝐶𝐶𝑆𝑆𝐶𝐶𝑅𝑅 𝑅𝑅𝐷𝐷𝐷𝐷𝑅𝑅

Table 12. Details of Downlinked Signals

NOAA-15, -18, -19

(HRPT) MetOp GVAR LRIT

Modulation Split phase BPSK QPSK BPSK BPSK Bit/Symbol 1 2 1 1 Coding Rate None FEC ¾ code None ½ convolution + reed

Solomon (255,223) FEC Code

BW (Symbol Rate) 1 DR 4/3/2 DR = 2/3 DR 1 DR 2*255/233*DR

7. Required energy per bit to noise power spectral density Ratio (Eb/No) based on each signal modulation type, coding and bit error rate (BER) requirements are shown in Table 13.

Table 13. Required Eb/No for Different Signals

Signal Required Eb/No (dB)

LRIT (GOES) 9

GVAR (GOES) 13.5

HRPT (POES) 13.5

MetOp A/B 9

8. Calculate Eb/No margin based on the following formula:

Eb/No margin = Eb/No calculated – Eb/No required -3dB (other losses)

9. Repeat steps 1 through 8 for each time step. Get Eb/No margin vs. time. See the top left subplot shown in Figure 13.

a. If it is a GOES signal, go to step 10.

b. If it is a POES signal, go to steps 11 through 13.

10. The GOES signal has a fixed azimuth (AZ) and elevation angle (EL) over time. The average value from step 9 is the average Eb/No margin estimation for each path.

11. If possible, obtain the corresponding AZ/EL data for that satellite pass from NOAA operators, which is shown in the lower left subplot of Figure 13.

12. Map each Eb/No margin with the EL angle. See the top right subplot of Figure 13.

13. Map each Eb/No margin with AZ angle. See the lower right subplot of Figure 13.

Figure 13. Results of steps 9 and 11 through 13.

14. For manual data collection, plot the average noise power level at three discrete frequency locations (1698, 1702.5, and 1707 MHz) vs. azimuth and elevation. An example from Miami is shown in Figure 14.

Figure 14. Average noise level for Miami, Florida.

4.2.2 POES

The POES system offers the advantage of daily global coverage by making nearly polar orbits 14 times per day approximately 520 miles above the surface of the Earth. The Earth’s rotation allows the satellite to see a different view with each orbit, and each satellite provides two complete views of weather around the world each day [8].

4.2.2.1 Spectrum View

Figure 15 shows a sample of the POES signals collected at Wallops Island, Virginia. The different colors represent signal power vs. frequency at different times during a pass. During one pass period, elevation and azimuth will change resulting in the peak received power changing as shown in the figure. These collections are used later in the data analysis to understand the available margin.

Figure 15. NOAA-15 signals collected on November 17, 2015, in Wallops Island, Virginia.

4.2.2.2 Margin Estimates

Figure 16 shows a sample of the POES NOAA-19 available margin estimates. The margin estimates improved our ability to understand the effects of interference on the downlinked signal. More information is available in section 4.3. Signal margin varies as EL/AZ varies. Refer to the appendices for the rest of the margin estimates.

Figure 16. Wallops NOAA 19 signal margin vs. EL/AZ.

4.2.3 GOES

GOES provide the kind of continuous monitoring necessary for intensive data analysis. GOES circle the Earth in a geosynchronous orbit, which means these satellites orbit the equatorial plane of the Earth at a speed equivalent to the Earth’s rotation. This allows them to hover continuously over one position on the surface. The geosynchronous plane is about 35,800 km (22,300 miles) above the Earth, high enough to allow the satellites a full-disc view of the Earth [9].

4.2.3.1 Spectrum View

Figure 17 shows a sample of the GOES signals collected at Miami, Florida. Throughout the data collection at both locations, the GOES signals were stable with amplitude variation less than a few decibels over a period of collection. Similar to the analysis done on the POES signals, the spectrum plots were used to provide the margin estimates. The spectrum plots for the other locations can be found in the appendices.

Figure 17. GOES West signal collected on December 7, 2015, in Miami, Florida.

4.2.3.2 Margin Estimates

Since GOES is geostationary satellite, the signal margin was calculated based on signal (processed data relay [PDR] or GVAR data format or LRIT and position [East or West]). Figure 18 is an example of the GOES East PDR estimate margin over five sets of collected data. The margin estimates for each location were stable over the data collection periods. The data for all the locations were then compared to see if the variation between locations match expectations. The site comparison and the margin estimates for the other sites can be found in the appendices.

SD

PDR

(GVAR)

LRIT

EMWIN

DCS

(DCPR)

MDL

Figure 18. GOES East PDR Eb/No margin.

4.3 Interference

During data collection, Aerospace wanted to understand what characteristics a long-term evolution (LTE) signal might have when carriers are using frequencies in the 1695–1710 MHz band. Aerospace used the AWS-1 frequency band (1710–1720 MHz) as a representative LTE interference. Figure 19 shows a series of manual collections from Miami.

Figure 19. Spectrum capture 1665–1730 MHz.

4.3.1 Miami Interference

In Figure 19, the orange box highlights an out-of-band emissions (OOBE) interference source originating from downtown Miami. This interference source was measured around -55 dBm/Hz during data collection in Miami. The OOBE from the interference source did interfere with the POES signals.

Figure 20 shows the impact the interfering signal has on a NOAA-19 downlink captured at Miami by reducing the available margin below zero at low elevations near 330 degrees.

Figure 20. Range of interference scan for Miami.

The green box in Figure 19 shows the AWS-1 frequencies. The rise in the noise level right after 1710 MHz is believed to be the LTE signal from AWS-1 carriers. The first major observation based on the data collection at Wallops and Miami is that the LTE signals from AWS-1 have not caused significant interference to incumbent POES and GOES signals. The second important observation is that because 4G-LTE operates in burst fashion, it will be technically challenging for the RFIMS.

Figure 21 shows detailed analysis of Eb/No margin decrease due to the interference shown in Figure 19.

This data was collected during a NOAA-19 pass at the AOML on December 19, 2016, at 14:58 p.m. The top subgraph is Eb/No margin vs. time. The middle subgraph is elevation vs. time. The bottom subgraph is azimuth vs. time.

The purple line on the left side of the graph depicts the time when EL was equal to 12 degrees and the AZ equal to 200 degrees. At that point there was 3 dB Eb/No margin. The purple line on the right side of the graph depicts the time when the EL was 12 degrees, but the AZ was 330 degrees. At that point there was 0 dB Eb/No margin. Interference was experienced around an azimuth of 330 degrees causing the Eb/No Margin to drop below zero during pass. This observation is consistent with other observations and confirmed that maximum interference occurred at azimuth of approximately 330 degrees.

Figure 21. Eb/No margins for NOAA-19 pass at AOML.

4.3.2 AWS-1 Interference

During the data collection, there did not appear to be any noticeable interference from the wireless carriers in the AWS-1 frequencies at the Wallops Island facility. LTE signal was observed only at the 1715–1720 MHz range when AWS-1 A block for uplink was allocated from 1710–1720 MHz. It could be possible that AWS-1 services are yet to be fully deployed, or an outage occurred during the manual data collection. If the entire band, 1710–1715 MHz, is ultimately utilized, it could potentially pose a problem to the POES system especially for the global area coverage (GAC), local area coverage (LAC), and HRPT downlinks from NOAA-18 satellite, which uses 1707 MHz as the center frequency.

From the Miami data collection, additional insights to the NOAA’s receivers were made.