ISR Support to Small Footprint CT Operations Somalia Yemen ISR Task Force Requirements and Analysis Division February 2013 Overall classi?cation of this document is

Introduction and Background Factors Impacting ISR Support to Operations Significant ISR Contributors Issues and Recommendations

SECRE l'//NOl ORN Study Overview (U) Previous studies* of Counterterrorism (CT) Kill Capture operations in Iraq and Afghanistan have detailed the role and impact of Intelligence, Surveillance, and Reconnaissance (ISR) for those missions (U) This study extends that previous analysis and reports on ISR performance and requirements for CT operations in Somalia and Yemen and should inform ISR planning and investments for potential small footprint operations elsewhere From Jul Oct 2012, the study team interviewed Task Force (TF) 48-4 personnel and collected extensive data on CT operations conducted from Jan 2011 Jun 2012 (U) Analytic results satisfy three linked objectives: Highlight key factors in smaller-footprint operating environments that have the most significant impact on ISR employment needs Identify capabilities that are most effective critical when operating in these environments Describe issues and make recommendations for resourcing and longer term investment Purose: Study ISR contributions to CT operations in the Horn of Africa (HOA) East Africa and Arabian Peninsula to inform ISR planning and investments for potential future small footprint operations 2007 Iraq HVI ISR Study; 2008 Ms. HVI ISR Study; 2012 Global F3EA Study (Studies conducted by ISR TF, 15/12 and 3

TF 48-4 is organized into two main branches: East Africa (EA) in Nairobi, Kenya and Arabian Peninsula (AP) in Sana?a, Yemen TF 48-4 EA and AP are further subdivided geographically into teams TF 48-4 forward support element is at Camp Lemonier, Djibouti TF 48-4 Theater Footprint ISR is based at three regional airfields, supplemented with sea-based 301"" mam? . 3? TF 484.2 AP Scan Eagles or MCI-8 Fire Scouts ?ammo Asmara, San' . . We Ma am. Djibouti (Camp Lemonier) .Simh FWD HQ Arba Munch om Djibouti .1 MolakalMQ-9 TF 48-4 counterterrorlsm 6x Ethiopia ?434.3 operations are focused on violent 'Rumbek 2" ?mad? . . Arba Minch 8X F-ISE Somalia Sweep ex remis organiza Ions Wm Mm GOCO Al Qaida in the Arabian Peninsula .n . . .. a Maritime ISR Al Qaida in East Africa/Al Shaabab Kampalae 9 1x on TF 48-4 CT operations are only a part of a broader whole-of-government no Manda Bay 0 TF48-4 Sites approach to regional security stability om; 0 Subsequent to data collection and interviews for this study, Some; 7; 43.4 Command Brie; flight operations are being shifted from Camp Lemonier 4

Introduction and Background Factors Impacting ISR Support to Operations Significant ISR Contributors Issues and Recommendations 5

Summary of Factors Impacting ISR Support to Ops HVI operations in HOA highlight several key factors impacting ISR support to CT operations in a small-footprint environment These factors differentiate CT in HOA from similar operations in Afghanistan or Iraq Key Factors Description Implications Not an Operations are ?Outside a Defined Theater of Active Armed Con?ict" (ODTAAC) Active Limits footprint, allowable US activities, penetration of comms networks Warzone Impacts operations and intel collection and exploitation activities HVIs are approved for targeting by President of the United States under Authorization to Use Military Force (AUMF) provisions ?322; AUMF process requires significant intel ISR to justify (and maintain) approvals Relatively few, high-level terrorists meet criteria for targeting under the provisions Approved are usually OPSEC and savvy; limits intel and finishing chances Strict A high level of assurance is required before a strike is approved . Must establish Positive Identification (PID) of HVI with "near certainty? . . . . . Assurances Only fInISh In a confirmed low Collateral Damage Envnronment (CDE) ?Near certainty? increases ISR work factor, reduces targeting opportunities Tyranny of Long distances to operating areas complicate the ?fixing? and ?finishing? of Distance Most objectives In Yemen are "500km away, Somalia can be over 1000 km Long transits consume "50% of ISR flight time and complicate strike planning 5

. Introduction and Background Factors Impacting ISR Support to Operations Significant ISR Contributors Issues and Recommendations 7

ISR Contribution to perations FMV and Aerial Precision Geolocation (APG) dominate Find-Fix-Finish part of F3EA APG tips narrow aperture FMV during Find During Fix and Finish, FMV and APG are used together, to maintain HVI location and PID SIGINT including Computer Network Operations (CNO) also contribute to Find and Fix FINISH HOA kill capture operations require input from other sources to drive the next ?find? ?75% of finishes in HOA are kinetic strikes DOMEX So. Very little ?finish-based? intel (DOMEX or a Ops AUMF interrogation) to drive next cycle FIX HVI ops EXPLOITI PROCESS PROCESS NF) In HOA, analysis-intensive 6. intelligence replacestactical site m, explortatlon, disrupting slowmg the cycle Source Ops F) This intelligence also feeds the D: FMV AUMF process which authorizes for D: SIGINT (Aerial Precision Geolocation, SIGINT internals/CNO) kl? C8 th re El= HUMINT Source Operations AUMF approval usually requires several El: (Tactical Interrogation Report, Document/Media Exploitation) months of intel ISR target development 8

?Exploit/Analyze? Replaced by ?All-Source Analysis? 2 an Target Package Sources - HOA 2012 use other HUMINT and SIGINT sources to fill in for missing 7% 3% when building target packages (Baseball Cards SIGINT, often from foreign partners, provides 57% of the BBC references HUMINT, primarily CIA source reporting, I NSA- Serialized SIGINT Report I NSA- Partner SIGINT Report r: TD - CIA Source Report IIR - Intelligence Information Report provides almost all other references IOther These sources .are [je'th'er as tlmely SIGINT, HUMINT is gap-filler for absent nor as focused as tactical Intelligence Target Package Sources - Iraq 2007 Therefore not as immediately relevant to the next cycle of Find, Fix, Finish 100% IIR-Intelligence 80% Information Report In Iraq and Afghanistan, DOMEX 60% and interrogation from Finishing actions INSA-SerializedSIGlNT 'd the bulk of hone numbers Rm? prov' - DOMEX-Document/Media locations and terrorist names . 7 20% Exploitation 0" I TIR - Tactical Interrogation This intelligence fueled the next F3 cycle, Locations Phone Names Report with objectives often actioned within days Numbers Finish-based intel immediately useful to next Find PORN Source: 48-4 Baseball Cards; IBM Analysis

A Find and Fix Operations Q. F) SIGINT provided two-thirds of the ISR Support for ?Find? ?hand-holds? used for HVI Finds Overwhelmingly, single sources were used to successfully Find top FMV and Computer Network Operations ?t (CNO) also contributed to some Finds FMV I APG I VID (Voice ID) FMV support becomes more CNO important in the Fix phase, typically in I OTHER combination with SIGINT Two-thirds of Fixes involved FMV Two-Thirds of Finds are based on SIGINT SIGINT, in various forms, continued to be a Support for ?Fix? dominant contributor in Fix I FMV maintains Fix on initial Find location to enable the TF to keep eyes on the HVI while building up to "near certainty? FMV, especially HD, is also used to build near certainty via identification of distinguishing physical characteristics I FMV APG I FMV COMINT FMV I VID (Voice ID) APG I APG HUMINT Finishes are largely the same-- every Finish was supported by FMV Source: rr 48-4 ISR Logs, SITREPs, Intel Summaries Oct 2011 -$ep 2012,- IBM Analysis 10 Combining FMV and SIGINT is key to successful Fixes

Introduction and Background Factors Impacting ISR Support to Operations Significant ISR Contributors ISSUES 80d Recommendations 11

Summary of Issues and Recommendations Key Findings and Recommendations Finding: Critical shortfall of capabilities providing PID and HVI location information ReqUirementS Recommendation: Continue to develop/field HD FMV and COMINT sensors Findings: Advance Force Operations (AFO) enable ?capture? vice "kill" finishes. National intelligence partners often have the best information and access. This can EprOit/Ana'yze be completed remotely by COMINT and CNO Alternatives . . Recommendations: Leverage AFO to Increase ?capture? operations. Pre-plan for rapid sharing of IC information and increase COMINT and CNO collection ops Findings: Long distances in A0 is a significant planning factor complicated by basing . and over-flight planning issues Long Distances . . Recommendations: Increase range and endurance for all ISR platforms. Consider sea-based ISR as a necessary complement to land-based ISR Finding: A key factor in' Find/Fix failures is the frequent inability to maintain 24/7 persistent stare on active mission areas, especially when ISR is massed for Finishes HOA ISR Orbits Recommendation: Support Combatant Command (CCMD) requnrements for additional ISR orbits to help prevent ?blinking? on HVIs during demand surges F) Also note: Addressing ISR issues will improve rate of operations (and successes), but expectations should be calibrated for realities of HVI ops outside of active war zone Constraining factors mean OPTEMPO will be significantly lower than previously seen in Iraq and Afghanistan 12

Requirements Unsuccessful Find-Fix-Finish sequences most often were due to issues with: Transition from Find to Fix: Primarily due to inability to acquire positive identification (PID) Transition from Fix to Finish: Combination of difficulty maintaining "near certainty? level PID and need to avoiding collateral damage areas Prosecution of Top EAAP HVIs Failure modes for ISR collections 3? FAILED TO N0 25 . - Detected (SIGINT) I SIGINT 2? ?til?? m? I . Uncertain PID 15 I PID: Weather (2), I Not 1o FINISH Sensor Fall, Enemy OPSEC . High CDE To Lost PID Unique HVIs VarIous AUMF FIND FIX FINISH Occurrences I TF Ops 3: Other Killed/Captured I Active Key Finding and Recommendation Finding: Critical shortfall of capabilities such as HD FMV, Voice Identification, Single- Use Determination that provide PID and support HVI location and CDE determination ReqUirememS Recommendation: Continue to develop/field critical ISR sensors such as HD FMV and COMINT systems (Voice ID, Single-Use ID, Geolocation) JP (Jackpot): A con?rmed Objective kill 13 51/ NOFORN Source: TF 48-4 ISR Logs, SITREPS, Intel Summaries Oct 2011 - Aug 2012; IBM Analysis

Alternatives to Exploit/Analyze F) Lack of ?Finish? intel to Exploit/Analyze and FiniShes HOA 2011'? cue ISR is probably the most significant reason for the low rate of finishes . Narrow gap" I. I Reconnecting the F3EA cycle would likely f; yield significant improvements to results was I $lrikC Strike (nonTFl I (OTHER) Efforts should focus on expanding 1 Cl 0( where possible through partner engagement will.) Improved remote airborne collection and I, Fill in gap" with broadened IC engagement can also fill the EA gap collectionand IC partnering Key Finding and Recommendation Finding: Small teams of special force advisors can assist the partner nation under Advance Force Host/partner Operations (AFO) which enables ?capture? vice ?kill? ?nishes Engagement Recommendation: Whenever possible, leverage Advance Force Operations (AFO) to increase ?capture? operations - tying previous finishes to future ?nds by generating more TIR DOMEX Fin?ding: In the reduced access environment, national intelligence partners often have the best [c Engagement information and access Recommendation: Pre-plan for rapid sharing of this information to minimize time delays Finding: Identifying information can often be obtained remotely through COMINT and CNO Recommendation: Increase airborne and remote COMINT and CNO collection and exploitation capabilities to make up for lack of access on the ground 5 ECR 0 F0 Scarce: TF 48?4 Storyboards, Baseball Cards Jan 2011 - Aug 2012; IBM Analysis 14

Long Distance Implications In Iraq 80% of finishing operations occurred within 150km of airfields The equivalent distance is 450km for Yemen and over 1000km for Somalia ISR platforms spend half their mission flight 5" time in transit-generating 38% fewer orbits per sortie than in other theaters sscaerl The issue of distance is magnified when translated to all of northern Africa MFW aircraft with a range of 450km will only reach about 5% of north Africa The range of land-based RPA aircraft allows them to reach 25% of the area 450km range for land- based aircraft . . 450kmrangeforsea- . Sea-basmg allows short-range aircraft to basedaircraft reach 35% of the land mass Transit Ranges from Bases* Key Finding and Recommendation Finding: Long distances from air?elds to operating areas is a signi?cant planning factor LONG ENDURANCE Recommendation: Consider ways to increase mission range and endurance for all ISR platforms; when satisfying airborne ISR requirements, key metric should be ?orbits? not or ?lines? Finding: Political and developmental issues complicate basing and over-?ight planning Recommendation: Even with shorter ranges, sea-based ISR may be a valuable complement to long endurance land-based ISR Some, ARGGIS Dam 59? Dec 2012,. Analysis Includes ISR coverage out of Djibouti, but not Niamey, Arba, or Agadez 15

Additional ISR Will Prevent ?Blinking? r? "i Shortfall in TF 48-4 ISR capacity is a contributor to ?fail to .1. I, 1 . find? and "fail to fixprevent ?blinking? during F3, operations in Yemen 3:321; require a minimum of 3 full orbits of ISR while Somalia operations require at least 1 full orbit Per TF, one orbit is equal to 24/7 on-station coverage .. . Additional capacity would be required to ensure ongoing 3 3? development of one target is not sacri?ced when massing ISR . for Finish of another I 90" Per TTPs, optimal ISR employment is three orbits per actioned 0 Mission Areas objective, which would equate to 3 (Somalia) and 9 (Yemen) orbits .. . r? - When ISR is massed in the Finish, coverage on other HVIs is lost .: . CAO: 301 2012 - at an ISR Actual ISR Req Per SN 3 1. Yemen 2.8 orbits 6.0 orbits 9.0 orbits - . Source: TF 48-4 ISR Logs, SITREPS, Intel Somalia 0.9 orbits 3.0 orbits 3.0 orbits Summaries Oct2011- Aug 2012;I8MAnonsrs Key Findings and Recommendations Finding: A key factor in Find/Fix failures is the frequent inability to maintain 24/7 persistent HOA ISR stare on active mission areas, especially when ISR is massed to support Finishing actions Orbits Recommendation: Support CCMD requirements for additional ISR orbits to help prevent ?blinking? on HVIs during demand surges Approved requirements are now 8 (Yemen) and 5 (Somalia). Actual orbits delivered has also increased to 3.9 (Yemen) and 1.9 (Somalia) 16

Appendix A. Study Background Appendix B. Find-Fix-Finish Appendix C. ISR Orbit Analytics Appendix D. AFRICOM Distances Appendix E. Additional Materials 17

SECRET Study Framework and ISR-TF CT studies are shaped to complement each other. Study teams are coordinating with each other and mission owners to minimize impact on operations SECRET Focus Area) Approach Future NW Africa 17 PACOM 17 Syria 17 Afghanistan 17 team construct Future Vignettes with J5 Case Studies: HOA deep dive Afghanistan BOTH teams C0nd-UCt Case Today AQIM Studies Data Pl? (ISR TF FOCUS Area) Driven AnaIySis with Recent Afghanistan 08 Focus Area) ISR-TF HOA Past Iraq 07/08 AFG ROW LOW Level of Supporting Case Studies CT Ops in highly Small Footprint Large Footprint austere locations CT Ops CT Ops SECRET 18

SURET . Analytic Approach and Methodology Analytic Approach: The Study of ISR Contribution in the Horn of Africa Understand the organization, Qualitativg mission, and map out Direct Observation processes and roles 30" lnterviews Understand the HOA Environment a . Quantitative Analysis Capacity VS. Determine the level ofactuvrty Corremion of capabilitie? 3 - or OPTEMPO to Identify how (assets) to potential capaCIty Capability . (missions) to determine well its achieved throughput I Identify all the causal factors Mixed Analysis . that impacts or limits Quantitative Analysis, Root cause Analys's performance to define the Qualitative Analysis, and "ga p? Stakeholder Coordination Determine how various ISR Quantitative Anamsis . . capabilities perform within Correlation of ISR collection ISR Contnbunon the F3EA cycle and define With Operational application effectiveness based on performance data Id zoo 7- Compare and contrast results .AFG 2008 in HOA with those of AF and Mixed thod Anal sis . Compare HOA, IZ, AF results 0 HOA ?01 1 l2 to understand the bigger and discuss with SMES sm- 20>? context of future operations Final SECRET SECRET 19

UNCLASSIFIED FOUO Study Schedule Oct 12 Nov 12 Dec 12/ 2013 Key Events - - Study Kick-Off 3 I 1. Preliminary Brie?ng 8- Coordination A b. Research/Orient. . Interviews CONUS Visits a. Fort Bragg, NC b. Damneck, VA c. Tampa, FL 4. Theater Visits 3 3. Germany 3 b. Djibouti c. Kenya . . d. Ethiopia (TBD) 5. Data Collection 6. Draft Report Production a. Re?ned Production 3 3 3 A0 Go Study iGck- Off Pie-Trip Prep Pre-Tnp Prep; Pos?Trnp Executive Summary Briefing (CONUS) (OCON US) I 2 Brief (CKONUS) Brie?ngs 7. Brief/Review A A A A A FOUO 20

Appendix A. Study Background Appendix B. Find-Fix-Finish Appendix C. ISR Orbit Analytics Appendix D. AFRICOM Distances Appendix E. Additional Materials 21

Timeline: Objective Peckham Case Study 2009 A FIFIF Find HUMINT SIGINT APG FMV .- Activity Description Summary Iii-2Intelligence collection on target to establish pattern of life lg loo I FIX 2006: 08] Peckham attended ?8th AI-Jinn?, a specialized training 8 FIX ?x I program - including explosives training 2006 - 2009.- OBJ Peckham returned to the UK and provided .2 I financial support to AQ allied elements in East Africa '75 .o . - Oct 2009: OBJ Peckham returned to Somalia after his second :0 I attempt; Travel was coordinated by 08] LOCKHART ?Dec 2009: 08] Peckham coordinated with Kenyan based facilitatar I . 311/201 . 9mm. . tun/:01! to facilitate money, equipment, and ?ghters through the UK to OBJ PECKHAM MOVEMENT Target has been located for kinetic/non-kinetic engagement . - 23 June 2011: Strike failed Problems with Approval Authorities, NAVAF C2, and AC malfunction I . NF) 12 January 2012: Fix during a vehicle follow No Strike "t Kill/Capture or neutralization of an enemy Target ;*a"ed stnke (23 Jun 11) 21 Jan 2012: 081 PECKHAM was eliminated via kinetic strike - 0359: White SUV enters CEL-012 (OBJ PECKHAM NAI) 0502: Adult with heavy strides and slight limp (OBJ PECKHAM) t'ave' no Finish (12 Jan 12) Early Jan 2012 OBJ relocated to Afgooye 0811: Vehicle follow begins - 1039: Full Register/Match (SI) :Lrlilkgeg (SI) 12 Jan OBJ traveled to Merca lzsazcominue to monitorthe scene 14-21 Jan OBJ returned to Af900yei'Ceelasha ECR ll/ NO F0 RN Source: TF 4841 Baseball Cards, Oct 2006 - June 2012; EA ISR Logs, Jan 2011 June 2012; IBM Analysis 22

Timeline: Objective Rhodes Case Study FIFIF HUMINT SIGINT APG FMV Activity Description Summary no no Intelligence collection on target to establish pattern of life I 200 m" I I75 Jan 2009: Yemeni NSB assessed Anjaf as a trusted deputy to I g? isoPOL . I 031 Canton and responsible for transporting extremists I FIX ?mgr F'ms?l - Mar 2009: Classified as one of eight main AQAP facilitators; I pm, . Identified as 828 on ROYG's Most Wanted Terrorist List 2 so I Sep 2010: ROYG assessed Anjaf and 081 Canton preparing useaVBlED against unspecified US interests 9?9r? 441? 3? 0e 0&ch I 2010: Favorable AUMF ruling mby Office of General I- .. .. .. Counsel under Jupiter Garrett CONOP OBJ RHODES MOVEMENT Target has been located for kinetic/non-kinetic engagement . . .4 3 July 2012 3n? - 23 Apr 2012: Unsuccessful Strike I i; 2 Enemy Killed In Action 5. 2 Enemy Wounded In Action (Including OBJ Rhodes) FINISH Kill/Capture or neutralization of an enemy Target - 3 Jul 2012: 081 RHODES was eliminated via kinetic strike 0713: Multiple VID (RHODES) and Geo-located at MAI-064 0825: Vehicle follow begins - 0908: 08] Rhodes correlated to vehicle; Near Certainty established - 1251: Strike - 1301: Continue to monitor the scene - . . 23 Ap, 12 - 1500: 08] buried near MAI-125 's .v Attempth FINISH N0 FORN Source: rr 48-4 Baseball Cards, Oct 2006 - June 2012,- YM ISR Logs, lan 2011 - July 2012,- IBM Analysis 23

Previous analyses show that the Find phase is typically characterized by a mix of APG, FMV vehicle follows, and HUMINT exploitation APG handholds from the Find phase then lead to many hours of HUMINT or FMV surveillance to maintain a fix on a bed-down location In HOA, Find and Fix work somewhat differently Lack of HUMINT puts more demands on SIGINT to cue FMV in Find High CDE at bed-down locations means FMV doesn?t watch to ensure the objective stays, but watches for the objective to leave Despite these differences, FMV and APG continue to be mainstays of the Find and Fix steps L. Afghanistan ?Find? Afghanistan ?Fix? Intel tips to FMV ?vehicle follows? until ?Finish? force arrives FMV ?stares? at bed-down location Find Fix Comparison HOA ?Find? HOA ?Fix? Saurce: TF 48-4 TFO and TF 3-10 TFD Databases; IBM Analysis

_f TF 48-4 SI Geolocation Comparison It} a NF) Unlike in Afghanistan in 2008, delays in target development cannot be attributed to a shortage of SI 2008 Afghanistan HVI operations were characterized by a significant drop in geolocations during the evening, which was attributed to ISR switching to FMV prime to support assaults NF) On the contrary, analysis of SI geolocations and vehicle follows (FMV) in AP shows a high degree of correlation and does not support the hypothesis that target development delays are due to a shortage of (count of logged ISR reports) ISR Activity by Time of Day (L) 1800 ?Orbits - 1600 a 1400 3 . 1200 ?g Vehicle Follows 1000 .Gcc?oc3 1400 :g i 200 ?9 -- 0 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 0100 0200 0300 0400 0500 0600 0700 0800 0900 1000 1100 0000 Source: TF 48-4 TFD, Orbit Tracker; IBM Analysis 25

Target Development Duration Comparison Target Development Duration for Actioned AF Selectors (S) In 2008 Afghanistan HVI ops, 25.0% -- - SECRET actioning of targeted SI selectors (cell '3 200% phone indicators associated with HVI 8 150% took a median of 11 days 100% . anda mean of 30 days ?336longer In AFthan In IZ 0.0% I I I (S) When compared to previous operations, the amount oftime ?0?8?39?90?0required to action objectives is literally orders of magnitude higher MEd'a" ?me to Adm" Objemves Eleven out of fourteen HVIs on the IZ 2007 AF 2008 HOA 2012 AUMF approved HVI list on January Median ?aways until 15t 2012 t'll th - 7 11 were 5 I on IS a amoned end of the stud eriod Me-an of days until 19 30 . . actioned (S) Obvuously, the requnrement for Over 60 days until 0 0 0 FMV time increases substantially actioned 1% 936 under these circumstances 13? days UNK UNK 79% actioned 26 NO F0 RN Source: RTF Summaries and Iraq 2007, Afghanistan 2008 CT Studies; IBM Analysis

SECRET (S) Compared to Iraq and Afghanistan, the pace of ?nishing actions in and Yemen(AP) is extremely low In 07-08, Iraq averaged around 8 finishing actions per day, while Afghanistan averaged around 1.5 Currently, Afghanistan averages around 6 finishing actions per day In AP the average (including mission partners) is around .2 per day (roughly 1 every two weeks) In EA the pace is so low that it cannot be meaningfully measured (only three finishes in the past 18 months) Moving Average of Finishing Operations Tempo 12 3 10 8 8 5 11-12 4 ,8 2 1142 . ?91011121314151617 Month of Study Period SECRET SECRET SourceJuly 2012,- IBM Analysis 27

Appendix A. Study Background Appendix B. Find-Fix-Finish Appendix C. ISR Orbit Analytics Appendix D. AFRICOM Distances Appendix E. Additional Materials 28

ISR Allocation Over the study period, AP has averaged 15 sorties per day, while EA has averaged only four Those sorties have generated an average of 2.9 orbits of ISR for AP and .9 orbits for EA (S) AP has averaged three times the allocation and on station hours compared to EA (S) APs on-station time is larger than EAs allocation N0 FOR Source: HOA Orbit Tracker, as of June 201 2; IBM Analysis Allocated and On-Station Sortie Hours 90,000 60,000 30,000 On Station 0 Allocated unmanned orbits with manned U-28 orbits However, the U-28 is very inefficient at generating an unmanned orbit It requires over 4.5 times the number of sorties to create an orbit Although RPA still deliver more on- station time per sortie, EA shows a more even mix of manned and unmanned platforms This is plausibly due to basing short duration manned aircraft closer to operating areas than the longer legged MQs 29

Requirements Per TF 48-4 AP JZ, AP requires a minimum of 1 manned and 2 unmanned ISR sortie (three total) to prosecute one HVI and conduct TADS related network development Actual SOCOM approved requirement to enable multiple HVI missions is six orbits The TF has mitigated shortfalls in unmanned orbits with the manned U-28 - Referred to as the "Chiclet" line, this practice compounded Djiboutian air control issues As the number of MQ?l?s at DJ has risen from 3 to 4 in Feb 2012, and finally to 6 in May 2012, 48-4 has been able to meet the minimums AP Orbit Meets minimum orbit requirements 4 .. . .. Unmanned Orbits Manned Orbits Orbits Total Orbits Unmanned Reqt - - Manned Reqt - - MinRequired .5124 ECR 0 F0 Source: HOA Orbit Tracker, as of June,2012 IBM Analysis 30

?rd-Kud- EA Minimum Orbit Requirements '7 A. Per TF 48-4 EA JZ, EA requires a minimum of 1 manned and 1 unmanned ISR sortie (two total) to prosecute one HVI Actual SOCOM approved requirement to enable mission is three orbits NF) EA has never been allocated their minimum orbit requirements EA has never consistently meet even half their minimum requirement EA did meet their unmanned requirements in July of 2012 with the addition of GOCO lines at Arba Minch and Fire Scout .. EA orbit AnalYSiS Far below minimum orbit requirements 2.5 Minimrn Betterment. - - 1 Unmanned Orbits Manned - - UnmannedReqt - - MannedR 0.5 c-q - - Min Required "Source: HOA Orbit Tracker, June 2012; IBM Analysis 3 1

Additional ISR Wi llPreve We found a positive correlation between Finishes and ISR clustering - Looked at a one year period (July 2011 June 2012) to see if the clustering patterns of ISR observed during normal operations changed during Finishes 3; Normal operations show about 2 orbits per day in the top cluster, about .85 orbits in the second cluster, and the remainder scattered - During the day of 3 Finish, there are about 2.5 orbits in the top cluster (a gain of .5 orbits) and the second cluster dwindles to about .4 orbits. The day preceding a Finish also shows a bump in not as significant as the day of 3 Finish. The day following 3 Finish shows even more decentralization than the average day This implies that normal Ops for the TF are to put two orbits on the priority objective and one orbit on a secondary objective - As ops go from Find to Fix and then Finish, the TF starts to mass ISR. This massing peaks during the Finish and quickly dissipates When taken in context of actual orbits we conclude that even during normal ops, the TF is ?blinking? a bit on the secondary objective. When ISR is massed in a Finish, there would be a significant loss of SA on a_l other targets, hence the ISR post-Finish is scattered to try to pick up the loose threads Main Effort Secondary Effort Other Efforts Normal 2.0 orbits .85 orbits .35 orbits Finish 2.5 orbits .40 orbits .30 orbits Post-Finish 1.9 orbits .70 orbits .60 orbits 5 0 F0 Scarce: r; 43-4 ISR Logs, SITREPs, Intel Summaries Oct 2011 - Aug 2012,- IBM Analysis Normal Ops Other Efforts ?Blinking? 32

SECRET Orbit Sufficiency 2012 ISR Activity by Day in Somalia 2012 ISR Activity by Day in Yemen TIME 9 . . TIME . . . PorLazea E775 :35. [wan Ratlines 7: - - . . between I I. i I 1. . Ethlopla .- 4 - . -. and l! .274: 25:1 - 5 Somalia 5 Ninjlb?"Mogadishu 'Marka?IMercarouts?dmatf?f?t; . Q. i . .- ISR activity clustering indicates Somalia operations are clustered around one op area at a time, while Yemen operations are routinely dispersed across 3-5 op areas SECRET Source: Somalia -SI Tracker Yemen -Sl mIRCTracker IBM Analysis

?1.4 Appendix A. Study Background Appendix B. Find-Fix-Finish Appendix C. ISR Orbit Analytics Appendix D. AFRICOM Distances Appendix E. Additional Materials 34

- 7?r'f? on. CLDJ Distances to HOA NAls Djibouti airfield to Yemen NAIs NF) Somali NAls are more distant and (km) . . 80 100% more dispersed than Yemeni NAIs 80% Meme" NAB WI 1. 2 so are greater than 500 80% km from Djibouti 60% '3 In Yemen there "6 40 are 225 NAls 40% 20 spread over space with an rt? 69? 59? 9? 9 90? 89? average distance - Djibouti Airfield to Yemen NAIs ?Cumulative 0f km Djibouti Air?eld to Somalia NAI distances 200 (km) 100% . 80% of Somali NAIs 9, In Somalia there are 150 I are greater than 1000 80% 362 NAIs spread over a ?5 mo . km from Djibouti 60% .5 1550 space With an 3 50 20% Average distance 0 - - 0% Of 1065 ?Somalia NAIs ?Cumu ative For Iraq - 80% of OBJs are within 150 km hr round-trip transit for For Afghanistan - 80% of OBJs are within 400 km hr round?trip transit for MQ-1) NOFORN Source: TF 48-4 Data, VI Campaign - Afghanistan ISR Assets, 2008 slide 9; IBM Analysis

SECRET Factors affecting ISR Orbits (S) TF orbits are defined as 24 hour on-station coverage - a reduction in on-station time is a reduction in orbit requirement satisfaction (S) On station time is affected by sortie impacts (weather, maintenance, and transit time) that subtract from sortie duration (S) In a large AOR like Yemen and Somalia, transit time (mu'tip'ied by the number of somesi most significantly impacts on-station time ?On-Station? Calculus Time On-Station Sortie Duration - (Transit Distance Transit Speed) 'sr'cnrr Reduction of On-Station Time per Causal Factori of Mission in Transit or. who 2?0 0 70"; 36,392 hours spent in transit I (almost 3 orbits per day lost) Transit 40v: I ant/[qunp . 30% Weather 20% I Operations 10% I I I Other 0% I MQ u-zs MQ-9 MCI-9 9-3 DJ- Scan rm: - MQ- DJ-AP DJ-AP DJ-AP DJ-EA Eagle Scout AFG '2'07 ?08 Preferred HOW to Swap short-legged aircraft like U-28 with longer Swap out with faster options to mitigate long legged aircraft like airframes like MQ-9 Consider transit Develop bases closer to target areas (slow and Add airframes (impacted by ?costs?? politically challenging process) limited basing) SECRET SECRET Source: Predator Times Data IBM Analysis 35

Long Distance Implications In Iraq 80% of finishing operations occurred within 150km of airfields The equivalent distance is 450km for Yemen and over for Somalia ISR platforms spend half their mission flight time in transit--generating 38% fewer orbits per sortie than in other theaters 3* ssca'cr The issue of distance is magnified when translated to all of northern Africa MFW aircraft with a range of 450km will ago? San-safarme only reach about 5% of north Africa based aircraft . . - The range of land-based RPA aircraft allows Vlosom?ransetr? KW. a '5 - 85:93:: them to reach 25% of the area roirgamgf Sea-basing allows short-range aircraft to based W, 7 I A reaCh 35% 0f the land mass Transit Ranges from Bases Key Finding and Recommendation a; Finding: Long distances from air?elds to operating areas is a signi?cant planning factor LONG ENDURANCE W: Consider ways to increase mission range and endurance for all ISR platforms; when satisfying airborne ISR requirements, key metric should be ?orbits? not or ?lines? Finding: Political and developmental issues complicate basing and over-?ight planning Recommendation: Even with shorter ranges, sea-based ISR may be a valuable complement to long endurance land-based ISR NOFORN Source: rr ISR data 2012,- IBM Analysis 37

Appendix A. Study Background Appendix B. Find-Fix-Finish Appendix C. ISR Orbit Analytics Appendix D. AFRICOM Distances Appendix E. Additional Materials 38

T-v ISR Platforms and Capabilities Current ISR Systems used in HOA Small-Footprint Operations it of Aircraft Armed for Manned (M) Time On Time On DNR Station Station Cruise Max 0' FMV HD-FMV COMINT COMINT APG (hours) (hours) Speed Endurance (35 0f 30 unmanned Mogadishu SanaMedium Fixed Wing (MFW) 4 n/a unk 8 2 U-28 2 -- 3 270 6 Predator 1 6 12 70-90 20 6 MQ-9 Reaper 1 9 10 175 14 4 Scan Eagle 13 n/a 55 15 1 USN Det MC-12 Liberty (Ext?d Range) 1 2(4) 4 (6) 300 6 (8) Red text denotes capabilities not in theater The PlD-providing phenomenologies, HD-FMV and DNR COMINT, are largely absent from ISR systems operating in HOA Not all MQ-95 have HD-FMV MFW platforms currently only fly in Somalia P-3 MS is a low-density/ high-demand platform currently not in Theater F0 RN Source: HOA Orbit Tracker, as of June,2012 aircraft spec sheets IBM Analysis 39

AUMF CONOPS Approval Process Timeline 'mxr "rs: vino" (S) 24 AUMF CONOPS approval times could be fully measured Other CONOPS were either not submitted (14) or were already approved/in staffing (21) Submitted AUMF CONOPS Results i Vt.) Day 54 AveDrzg: 79 Avegzg: 196 Statistic I ly ly Summary p.p times Submitted Disapproved Approved In Staf?ng (24 CONOPS) (4 CONOPS) (1o CONOPS) (1o CONOPS) Max 274 .34. - ?:M55v . . - - . . 120 150 180 210 240 270 300 Med'an 35-5 Days -.. 2. ass-x4: . -- - Luz(S) While it takes ~79 days to approve an AUMF CONOP, over half of the CONOPS were approved within 36 days Removing the single highest data point would reduce the average approval time to 58 days Adding in those CONOPS still in staffing would double or triple these times (S) These times do not account for the effort expended to collect and analyze intelligence to develop these CONOPS NOFO RN Source: rr 48-4 RTF Files; AUG 2011 - ocr 2012,- IBM Analysis 41

HOA HVI Life Cycle (S) HOA require national intelligence resources and years of developmental work (S) Once identified as a target, HOA AUMF approval utilizes ISR intelligence to develop POL and CONOPS (S) Additional time is then required to attain near certainty and low CDE requirements for the process HOA HVI in-process time 4to 8+ 1 Ta'get AUMF Months ?19."qu 1 to 3 . th Target . Target Development 0" 5 18 Months to 6+ Years Data Summary Max 22.2 Years 9 Months 14.2 Months Min 0.4 Year 0.9 Month 1 Month Median 4.8 Years 1.2 Months 7.2 Months Mean 6.0 Years 2.6 Months 8.3 Months . 1. FMV SI Conceptual Intel . 1. National Intel Contribution 1.Natlonal Intel 2. FMV I SI 2. HUMINT 3. SECRETNNOFORN 50urce: TF 48-4 RTF Files; AUG 2011 - ocr 2012; IBM Analysis 42

HD FMV Impact on Fix Successful Fixes Unsuccessful Fixes Detected SIGINT Other 3 2 0 1 Located Successful 11 5 3 3 Fixes Unsuccessful 17 2 15 I Fixes i 45% 72% Involved 88% No HD FMV Involvement FMV in HD FMV is involved in many if not most successful fixes successful fixes involved HD FMV Possible/Probable systems were MQ-9s highly likely to have HD I +Poss HD HD Successful Fixes U'l Most of the unsuccessful fixes did not "anon-o involve HD FMV - In the two failure cases involving HD FMV, cloud cover and bed-down location monitoring likely 0 10 20 decreased its utility Unsuccessful Fixes 43 NOFORN $0urce: rr 48-4 ISR Logs, SITREPs, Intel Summaries Oct 201 1 - Aug 2012; IBM Analysis

Key Ta ke-Aways Strategic Operational Tactical Operations in the small/medium footprint theater are fundamentally different from what we?ve experienced in Afghanistan and Iraq. Political constraints, long distances and ISR limitations make this a challenging future There is a critical shortfall of capabilities providing and HVI location information. We need to continue to develop/field HD FMV and COMINT sensors that provide this information A key factor in Find/Fix failures is the frequent inability to maintain 24/7 persistent stare on active mission areas, especially when ISR is massed to support Finishing actions. Supporting CCMD requirements for additional ISR orbits will help prevent ?blinking? on during demand surges