Conflating the activityof Intelligence, Surveillance, and Reconnaissance (ISR) and the missionofreconnaissance has allowed the Department of the Air Force’s reconnaissance capabilities to atrophy, dangerously.  The Joint Doctrine is clear: ISR includes planning, processing, exploitation, and dissemination (PED). Reconnaissance aims to collect information and data without prescribing its use.  After two decades of low-intensity conflict, the Air Force is making phenomenal advances in ISR.  However, it is making this progress without revising its approach to the mission of reconnaissance.  To meet the challenge of all-domain anti-access/area denial (A2/AD) posed by pacing threats, the Air Force must revitalize reconnaissance by adopting a user-agnostic reconnaissance concept and updating its reconnaissance systems to enable it.

Recce for a Slow-paced Conflict – Not a Dynamic War

              The current reconnaissance model takes an obsolete approach to data logistics, the conventions and architectures through which an organization moves information.  For most of the Global War on Terror (GWOT), reconnaissance assets were not physically threatened.  The electromagnetic spectrum was a permissive environment. Data logistics were largely uncontested, and coalition forces easily gained and maintained physical freedom of maneuver in air and space.  Errors in the application of force were among the greatest threats to mission, and the GWOT adversary could not punish overextended lines of communication in the information environment.  To avoid errors in application of force, ISR activities and supporting reconnaissance missions evolved to value extensive human review.  With data sharing free from attack, Air Force recce professionals optimized technologies and procedures to feed centralized PED.  To do otherwise would have increased risk to mission in conflicts the United States could afford to fight at a comparatively glacial pace.

              The risk assessment must change in peer conflict; the adversary will have both the capability and intent to disrupt centralized processing and attack data logistics.  Even with difficult to target reconnaissance platforms, under the current construct the adversary can achieve reconnaissance mission kills simply by interdicting data.  The PLA’s concept of “System Destruction Warfare” is designed to generate decision paralysis by exploiting exactly this vulnerability.  If Air Force reconnaissance does not conform to a more resilient model, one prioritizing redundant pathways from data collector to data users at the tactical edge and analysis nodes, the Department of the Air Force’s (DAF’s) reconnaissance missions will fail.  

            Due to threat capability, the ability to transmit data from collector to user is more important than ever.  However, for reconnaissance assets, it remains convoluted, human intensive, and fragile.  Fighter and bomber sensors are often constrained by design compromise (e.g., lightweight for maneuverability) and tactical considerations (e.g., low altitude operations). Therefore, they often lack the requisite fidelity or line of sight to accurately sense an adversary employing effective maneuver, camouflage, concealment, and deception.  As a result, locating mobile targets over large areas still requires exquisite sensors and platforms to host them.  However, in contested information environments, Air Force reconnaissance consistently demonstrates that it suffers from its GWOT inheritance.  Data from recce platforms is often only accessible in the battlespace after circling the globe.  As a result, exquisite systems are often unable to provide information on relevant timescales. “The information died in chat” is a discouragingly common statement in exercise debriefs.  When communications are degraded or denied, recce data often does not arrive at all. This state is not a failure of sensing or sense-making. It is a failure of design.

Revisiting Assumptions – Recce Design for the 21^st Century

            The Department of the Air Force must revise its fundamental approach to reconnaissance, making a long-overdue course correction.  In 1935, military leaders intentionally separated strategic and tactical reconnaissance, in part to justify an independent air service.  Tactical reconnaissance focused on informing maneuver forces and became the purview of the Army. Strategic reconnaissance enabled bombing campaigns and would be conducted by the then-incipient air service.  In so doing, the Army Air Forces allowed the means of collect to define data’s use and user. The Air Force entered the Cold War with the conceptual hangover of this bifurcation, which was reinforced by the available technology.  Until the late 20^th century, bulk data was transmitted via physical artifact (e.g., photographs), and there were no feasible alternatives to a centralized, ‘PED first’ model.  Film from RF-4 and SR-71, for example, needed specialists to develop, mensurate, mark, duplicate, and distribute reconnaissance collect.  Film in the belly of an aircraft was useless until system-specialized PED occurred.  Such was the practical design for reconnaissance, given the available tools.  Over decades, the ‘PED first’ model became cemented in organizational constructs and service doctrine, even as technology advanced rapidly.  With no strong incentive to change through the GWOT era, the DAF’s unchecked assumptions about reconnaissance persisted into the 2020s.

            It is time to make a change.  In an era of peer adversaries capable of pushing the pace of conflict, the Air Force must adopt a user-agnostic approach to reconnaissance data.  It must cease tying information’s use to the collecting sensor or platform through a ‘PED first’ model.  Doing so addressees technological realities of the 21^st century and parallels the logic of the Air Force Distributed Common Ground Station’s (DCGS) sensor agnostic paradigm.  Analytic capacity is no longer bound to reconnaissance platforms or missions. Similarly, intelligence processes should not define the utility or use of data obtained through reconnaissance.  In the user-agnostic model, recce data should be as accessible in fighter cockpit as it is on the DGS floor.  Furthermore, reconnaissance systems must be explicitly constructed to streamline data flow to users in fighter and bomber cockpits.  Doing so increases the effectiveness of ISR as a whole by freeing analytic cycles for the complex sense-making tasks that only intelligence professionals can do.  The nation’s premier analysts should be making sense of complex schemes of maneuver, not serving as middlemen for reconnaissance data. 

            The user-agnostic model deletes the insidious assumption that reconnaissance data must first flow to PED nodes before becoming useful.  Supporting the activity of ISR must remain a primary function of reconnaissance systems and personnel (e.g., reconnaissance aircrews).  However, Air Force reconnaissance missions must also make data directly available to those force elements operating at risk in contested areas (e.g., fighter aircraft operating within range of enemy weapons).   In robust A2/AD environments, considerations like emissions control will limit organic fighter and bomber sensing.  For example, an F-15E crew might be unable to generate synthetic aperture radar (SAR) maps of a target using own-ship sensors, denying the crew a tool it routinely uses to complete weapons engagement sequences.  The simplest solution for want of a SAR image is a SAR image, and the DAF does possess multiple systems capable of collecting these images from positions of relative safety.  However, no existing long-look SAR collector can provide images directly to a shooter.  Even basic location and combat identification data, which could be transmitted via voice radio requires reach back to a ground station.  Fighter aircraft have been sharing sensor data across formations for decades.  Doing the same for reconnaissance sensors is a simple logical leap if the DAF sheds outdated assumptions.  

            Some may discount the utility of raw reconnaissance sensor data in fighter and bomber cockpits.  These voices might be surprised at the similarity in sensor output and the degree to which tactical crews are trained in sensor analysis.  A fighter crew capable of interpreting a SAR map, whether it is generated by a U-2 or an F-15E .  Furthermore, today’s reconnaissance data is rarely truly raw. It is normally contextualized as it is collected.  SAR images, moving target detection, and full motion video are tagged with geolocation data.  Electronic intelligence is often processed at the point of sensing for emitter identification.  Advances in artificial intelligence and automated target recognition across a variety of detectable signatures provide further fidelity.  Entrusting tactical crews with reconnaissance sensor data shifts human interpretation to the tactical edge, decreasing timelines and increasing lethality.

            Within a user-agnostic reconnaissance framework, ISR and intelligence preparation remain critical functions.  For tactical aircrews these pre-mission intelligence processes provide context and establish which relevant detectable signatures may be available for sensing. When own-ship sensors prove inadequate, direct data pathways would permit fighter and bomber crews to call for long-range support from exquisite sensors in the same way that ground force commanders call for long-range fires.  To maintain the initiative in future conflicts, an F-35 pilot must be empowered to cue stand-off sensors with a button-push and receive data at machine-to-machine speeds. Joint All Domain Command and Control (JADC2) is intended to provide the backbone for this broad data access and distributed control.  However, if it intends to offer reconnaissance that generates a tactical advantage, the Department of the Air Force must drag its concept of what a reconnaissance platform does out of the last war and into the 21^st Century.

The Relevance Rubric for Reconnaissance Systems

              The Department of the Air Force must adapt its platforms before JADC2 arrives.  As it works towards JADC2’s implementation, the Air Force should leverage existing tactical data links to transmit relevant information while re-baselining reconnaissance tactics, techniques, and procedures to improve direct sensor-to-shooter pathways.  Parsed data such as position, identification, and direction of travel can be transmitted with extremely low bandwidth.  Where existing datalinks are insufficient, voice communications can supplement.  Unfortunately, some reconnaissance crews, even those physically co-located with their sensors, still cannot view their own collect without reach-back, much less publish sensor data at the edge.  In contested environments, these platforms risk irrelevance. To remediate, the Department of the Air Force should adopt a data-centric rubric for reconnaissance systems, mandate compliance, and eliminate systems to free resources where compliance is impossible:

1) Can the system collect, compile, and contextualize information with geolocation and object identification (human or machine) at the edge?

2) Can it host that information for retrieval on-demand and push it to critical users (e.g., both F-35 and analysis teams)? 

3) Does it offer its sensors for use by edge users through cross-platform cuing like that available between fighter aircraft?

4) Can it reliably complete its mission to provide value-added data in a contested physical and information environment? 

            Collection and contextualization need not occur completely autonomously. Human in or on the loop processing is acceptable, and often preferable as in Rivet Joint’s history of operations at the edge.  Fragile data logistics with successive single points of failure must be intolerable.  When enemy action in centralized data transfer layers results in mission failure, whole fleets are held at risk.  The Department of the Air Force must eliminate its dependence on such systems.  Outdated reconnaissance thinking advocates for data security, architecture modernization, and big-data tools as the primary answer to an adversary’s data logistics threat. While hardened reach back remains valuable, betting a sensing mission’s success on a more secure (but not more resilient) version of the current schema equates to building a digital Maginot Line. 

              Some of the nation’s reconnaissance assets already compare favorably to a data-centric rubric. The Rivet Joint and many AFSOC platforms have consistently bridged sensor to shooter gaps.  Platforms like these will continue to add value as forward processing nodes in an age of JADC2.  Others, like U-2 are redeemable. They are presently dependent on centralized processes but could be upgraded in short order and at reasonable expense.  Some systems, however, depend heavily on fragile data architectures. The Department of the Air Force must work to remediate this vulnerability.  In doing so, sunk cost remains an irrational consideration.  It is worth restating that an unmanned or space-based platform is not inherently superior to a manned platform in a peer conflict.  Where placing humans at the point of collect reduces overall risk to force or mission, the Air Force should not hesitate to place reconnaissance crews at risk.

                Success in competition and conflict still requires exquisite sensors on purpose-built platforms.  However, in peer conflict these platforms must be equally capable of supporting both sense-making ISR activities and on-demand sensing for maneuver forces.  Adopting a user-agnostic model permits planners to actively manage the allocation of reconnaissance assets, thereby prioritizing the flow of reconnaissance data to support decision-makers at the most appropriate echelon for current conditions.  This echelon may be an F-22 flight lead one hour and an intelligence cell supporting campaign planning the next.  Failing to adjust the DAF’s current reconnaissance model abdicates proactive decision-making in favor of perpetuating a system designed for a world that no longer exists.  A user-agnostic model and systems compatible with it will provide the decision advantage required to deter and, if necessary, defeat the pacing threat.

Disclaimer: The views expressed are those of the author and do not necessarily reflect the official policy or position of the Department of the Air Force or the U.S. Government.

The author would like to thank Prichard Keely and John Mattson for their support and guidance.