The official position of the US Air Force is that the U-2 can be replaced by satellites. I have discussed this briefly in previous posts, such as this one, which includes a great SAR satellite image of RAF Fairford – a U-2 base – taken by Capella Space during last year’s Royal International Air Tattoo (RIAT).

So let’s take a closer look at the current situation regarding ‘overhead’, a term that is often used by the intelligence community when referring to recon from space. I’ll try to be objective.

Firstly, there are the highly classified satellites operated by the National Reconnaissance Office (NRO). In open forums, before even the existence of the NRO was declassified, they were described as America’s National Technical Means (NTM). They have been orbited with ever more sophisticated payloads since 1960, mainly for imaging and SIGINT. Some are devoted to the relay of reconnaissance data. They are large and have ‘exquisite’ capabilities – that’s another intel euphemism.

But they are enormously expensive and there are thought to be less than 20 of them – not enough to provide constant or frequent coverage of any one area on earth. Their main purpose is to serve the intelligence agencies of the US and four selected partner nations: Australia, Canada, the UK and New Zealand – the so-called Five Eyes (FVEY) alliance. However, their processed ‘take’ was distributed more widely to military users, starting in the 1980s and accelerating in recent years, mostly at a lower classification and usually downgraded and/or with disguised origin.

Imaging of space by commercial satellites started in the mid-1980s. Over the next three decades, spatial resolution and geolocation accuracy increased dramatically, and monochrome coverage was supplemented by multispectral sensing. The commercial companies began to offer a variety of processing and interpretation services, some of them bespoke. The National Geospatial-Intelligence Agency (NGA) started issuing contracts to them. Thereafter, it could interpret and release unclassified imagery to military users with fewer restrictions than NTM, or to the media when required. 

Then came the small satellite (or ‘cubesat’) revolution. The miniaturization trend in electronics led to much smaller payloads. Operating costs reduced as new launch companies such as SpaceX entered the market, and the possibility of sending multiple cubesats to space on the same rocket was exploited. So the companies could afford to build constellations of satellites in low-earth orbit (LEO), thereby increasing the frequency at which they could image any one area or location on earth. 

It was not long before satellite imagery provided by such companies as Maxar and Planet Labs became ubiquitous. 30cm resolution is now routinely offered. Users have the opportunity to task the satellites themselves. They can be maneuvered and rotated for optimum coverage.

In the past few years, new companies have entered the market to specialize in other forms of sensing. For instance, IceEye and Capella Space are doing SAR imaging. The former offers dwell, spot, strip and scan imaging modes at resolutions as good as 50cm. The latter claims industry-leading revisit rates and diverse revisit times.

Hawkeye 360 has led the development of commercial ELINT satellites. Their first application was maritime surveillance, but after its recent acquisition of Maxar’s RF technology and databases, Hawkeye 360 now claims that it can “quickly scan gigahertz of bandwidth to efficiently map active frequencies at a regional level.” Its satellites monitor L-, S- and X-band.

With such capabilities on offer, it’s not surprising that the NRO and the NGA have embraced the commercial satellite companies. It began giving contracts to them in 2019, and the US government now spends over $400 million annually with them. The NGA is even proposing to buy analytical services from the commercial providers, in addition to raw data.

Ten companies now have representatives in the Combined Space Operation Center at Vandenberg, cleared to the top secret SCI level. The NRO Director said recently that “Commercial…has actually taken over part of our mission and allowed us to focus the exquisite capabilities on those things that are needed.”

Moreover, the fact that commercial companies are now launching hundreds of satellites, from ‘hot’ production lines and on rockets that are readily available and (in some cases) reusable, means that the intelligence and military users can be assured of uninterrupted service and those high revisit rates. SpaceX has had a classified contract to provide launch services since 2021. Many commercial satellite companies use SpaceX to launch their own payloads. There are other launch options available to them.

However, there is a major new ISR satellite constellation development that is being entirely funded by the government. It will provide Ground Moving Target Indication (GMTI), a role that the USAF E-8 JSTARS previously performed, along with some other platforms to a lesser extent. (The U-2 has done GMTI in the past, but the shorter antenna of ASARS does not make it an optimum platform for this task.)

It seems that the NRO, the NGA and Space Command have decided broad situational awareness can henceforth be achieved by a combination of commercial EO and SAR imaging from space supplemented by the NRO’s few dozen satellites. But a new GMTI constellation is required, if  ‘tactically responsive’ ISR from space is to be fully realized.

The NRO began serious study of this plan in 2019. Some preliminary contracts were issued. There must have been many design tradeoffs to be considered, not least the power requirements. An analysis of alternatives was done in 2022. The effort has been kept almost completely classified – at the NRO, old habits die hard.

(It’s worth noting here that at least one commercial SAR satellite operator (Umbra) says that operating these satellites in a cluster formation offers distinct phenomenology that could offer some GMTI.)

The NRO and Space Command are jointly funding this effort. The NRO is writing the technical requirements and overseeing the technical development. The ground infrastructure is Space Command’s responsibility.

But the formal funding of a program of record only began in the current fiscal year. The planned spend is nearly $1.75 billion over the next six years. A further nearly $1.5 billion is requested for “auxiliary payloads”, whatever that means. Fielding of the capability is not scheduled to be complete until 2029. (These sums may not include the $1.8 billion contract awarded by the NRO to SpaceX in 2021 that was probably for launch services as part of the GMTI plan. The company has already sent some prototypes into orbit, and on 22nd May this year one of its Falcon 9 boosters launched the first 21 of the new GMTI satellites. Eight of these are believed to be for sensing, with the others being for communications functions in the future constellation.)

So how does the Dragon Lady’s capabilities compare with all this? The version of ASARS currently in operation may no longer offer significant advantage over satellites, in terms of spatial resolution and geolocation accuracy. However, the upgraded ASARS-2B probably redresses that balance, at least for SAR imaging. It probably also includes a better GMTI capability. And an airborne sensor like ASARS can provide a ‘relentless stare’ in order to ‘keep custody’ of targets from an orbit, whereas ISR satellites cannot. Moreover, an airborne platform like the U-2 can field new sensors and sensor modifications more rapidly and cost-effectively than satellites.

On the U-2, SYERS can be substituted for ASARS. It is a very sophisticated ten-band multispectral sensor whose performance can probably not be bettered from space. But SYERS is tasked much less frequently than ASARS, since only radar imaging can overcome cloud cover. (Incidentally, an earlier and less capable version of SYERS went into orbit as long ago as 2011.)

I hesitate to attempt an evaluation of  SIGINT collection from satellites, versus the U-2, because of the high classification. But there are as many as 50 NRO SIGINT satellites, and Hawkeye 360 is one of the ten commercial providers under contract. On the other hand, for reasons of orbital dynamics and distance-to-target, and for the collection of certain types of signals, some SIGINT is probably best done from an airborne platform. That is why the USAF is not planning to retire the RC-135 Rivet Joint, Combat Sent and Cobra Ball aircraft, even though it has described the E-3 AWACS and E-8 JSTARS which are also based on the C-135, as now-vulnerable platforms.

The U-2 flies at twice their altitude, and can therefore ‘hear’ certain signals at a greater distance. But altitude is no longer a refuge for the Dragon Lady. Its mission planners have to consider the ever-increasing threat posed by long-range SAM systems such as the S-400. The ASARS upgrade provides longer-range imaging, to help offset this problem. But the USAF has concluded that the U-2 is no longer survivable against a modern Integrated Air Defence System (IADS). That is why the other part of the U-2 replacement is a stealthy, high-altitude UAV identified by various open sources as the RQ-180, the Penetrating ISR (P-ISR), or The White Bat. I have discussed this still-highly classified system in various previous posts, which you can find here and here.

Timely collection is important. With the SAR satellites’ revisit rates becoming ever more frequent thanks to the large constellations, one advantage of airborne sensors decreases. But what about timely tasking, processing, exploitation and dissemination (TPED)? The Distributed Common Ground System (DCGS) was originally designed around the U-2 to do exactly that, including the ‘fusing’ of the take from different types of sensor. It was later expanded to handle imagery and signals from other platforms, and has undergone continuous improvement.

But as well as embracing commercial space, the NRO began devising new dissemination architectures that could quickly provide warfighters with the products they needed. The creation of Space Command has facilitated that process.

The budget request says that the Space Force “will be working with the other services and the intelligence community to understand the joint tactical-level ISR and warfighting/targeting requirements and the best way to meet those requirements.”

In other words, Space Command will be developing the concept of operations, including consulting with the combatant commands. They will request targeting, and Space Command will process the requests and point the sensors

Does any of this sound like a mature system? It’s worth noting that only last December, the Chief of Space Operations Gen Chance Salzman said that “there are a lot of moving parts that have to be fleshed out.”

Even then, the question remains: can a space-based architecture really provide TPED at the speed of relevance? The USAF says it can, as part of an Advanced Battle Management System (ABMS). The US Army is not convinced. It was the main ‘customer’ for the JSTARS, which the USAF is retiring, and which had a crew of airborne analysts to do the PED. The Army has responded to the JSTARS grounding by creating a whole new programme of record for a fleet of high-flying multisensor platforms based on a business jet.

The NRO claims that it is “developing the most capable, diverse and resilient space-based ISR system the world has ever seen.” What do they mean by resilient? Potential adversaries such as China and Russia have demonstrated their satellite-killing capabilities, and the jamming of datalinks. Presumably the NRO is referring to the redundancy inherent in today’s large constellations, their multiplicity of communications options, and the speed at which their smaller-and-cheaper satellites can be replaced, if some are destroyed by enemy action. But there are other means of disrupting satellite collection, such as denial-of-service type cyber attack, or even kinetic attack on the ground stations.

In comparison, with the increase in jamming and other electronic warfare techniques demonstrated in the Ukraine War, airborne assets may be able to augment precision navigation and timing (PNT) needs. Their datalinks can be narrower and provide more precise pointing in a ‘daisy chain’ arrangement to sources of PNT data that are outside the affected area.

So what would I recommend? First, the U-2 should certainly not be retired before the NRO’s relationship with the commercial providers is mature, the new GMTI satellite constellation is complete and fully functioning, the data relays are proven and offer redundancy, and the military users have developed a robust concept of operations. Based on public statements by NRO and Space Command officials, this checklist cannot possibly be ticked off before October 2025, when the U-2 is due to be retired.

And then? It seems to me that most U-2 imagery collection from the three permanent overseas bases has become unnecessarily routine. Satellites can do much of that mission, and with wider coverage. Whether they can fully replace the U-2 for SIGINT collection is another question.

But the USAF has spent a serious amount of money on developing the U-2 avionics upgrade, the ASARS upgrade, and some still-classified new SIGINT sensors. It has given the Skunk Works a contract to convert the first eight aircraft. I suggest that those eight jets be retained in service, based at Beale, but with the capability of being quickly deployed to overseas locations, whenever their unique collection capabilities are required, or when the tasking of alternative platforms, both airborne and spaceborne, manned and unmanned, exceeds their capacity to deliver. Additional overseas bases should be prepared for this contingency, including the pre-positioning of vital support equipment and the U-2’s unique JP-TS fuel.