Intelligence Retarded

The Army’s Military Intelligence (MI) was poorly prepared for the challenges of a counter-insurgency campaign in Afghanistan and Iraq.  MI had long prepared for a mechanized war of maneuver in Western Europe. The flow of information in such a scenario is largely top down, with technical assets at corps, army, and theater levels gathering, analyzing and dissemination information on the enemies order of battle, location, objectives and probable courses of action. Enemy tank and motorized rifle battalions, regiments, divisions and armies would be detected and tracked by reconnaissance, electronic and signals intelligence, radar mounted on platforms like the E-8 Joint STARS.* Reports from agents, informants, and captured enemy prisoners is important, but usually of a secondary nature. Speed of collection, analysis, and dissemination is critical.

Intelligence in a counter-insurgency, on the other hand, is driven by Human Intelligence (HUMINT), and is a “bottom up” venture. Combat units are not only the primary customer of the final intel product, they are the prime collectors of the raw data that MI uses to generate the analysis. Local agents, members of the indigenous population, and captured insurgents are the primary sources of the raw intelligence data.

It is a slow methodical process to identify persons of interest, finding individual insurgents, and then extrapolating that into locating cells, leadership, communications networks and funding sources. Signals intelligence, electronic intelligence, aerial reconnaissance and other technical means are used to support these efforts, as opposed to being the primary drivers of the collection effort.

After ten years of war, the MI branch and the supported combat units have gotten much better at the challenges of running such an intelligence campaign.

But the Army has to maintain its capability to operated in a high intensity, high tempo operational environment. One possible theater of conflict is, of course, Korea. The North Korean Army is big, and if it isn’t equipped with the latest and greatest, it does have a very large number of tanks and armored personnel carriers. If they can crack through the crust of South Korean defenses on the DMZ, they can use their speed and mobility to overrun Seoul and much of the south. The speed and agility of the US and South Korea will be critical in defeating any North Korean invasion.

Not surprisingly, the US 8th Army and US Forces in Korea train regularly with South Korean forces for just such a scenario.  Recently, a command post exercise, Ulchi Freedom Guardian, was held to game out just a scenario. The results were not encouraging:

Intelligence software that the U.S. would rely on in a war with North Korea froze up repeatedly during a joint military exercise in South Korea in August, hampering the ability of U.S. and South Korean commanders to watch the movements of simulated enemy forces, a senior intelligence official said.

The Distributed Common Ground System-Army (DCGS-A) software is designed to link intelligence analysts to communications intercepts, imagery and radar collections stored in massive databases. When American intelligence analysts tried to use the software to track simulated North Korean troop movements, the screens on their DCGS-A workstations sometimes went black, forcing them to reboot the software, the senior intelligence official said.

We’ve seen an explosion in the use of drones throughout the services, from tiny hand launched short range drones, through the Predator and Reaper armed drones, to the huge RQ-4 Global Hawk to gather intelligence. The challenge is, how do you sort the wheat from the chaff from all the imagery and video collected. Similarly, the massive increase in the use of computers throughout the battlespace means there is exponentially more information that needs to be sorted. The DCGS-A is supposed to be the tool to enable this. But the ability to collect information has far outstripped the ability to collate, analyze and distribute it. Effectively, the DCGS-A was obsolete before it was even fielded.

There are a couple of institutional reasons why this is so.  First, the procurement process is unable to move quickly. Merely defining the requirements for any given system can take years. The end users and the Army’s procurement agency have to decide what they want any platform to accomplish. They then have to justify a program to the Office of the Secretary of Defense, and then convince Congress to fund the program. Then they have to issue a Request for Proposals from industry, select the vendor, and then and only then begin the actual development of the product.

There is also great institutional pressure to have one platform perform as many functions as possible.  One platform means fewer operators needed, a streamlined training pipeline, and streamlined logistical support in terms of spare parts. The need to minimize manning for operators and maintainers (and the training cadre) is critical. The Army’s end strength is set by law, and any soldiers used in one field have to be “bought” at the expense of soldiers elsewhere.

Further, buying and maintaining one platform should be cheaper. Theoretically, a single software platform would also mean simplified software management, but any user of Windows knows that’s not necessarily true.

And so we wind up with a system that is supposed to expedite the intelligence that our command posts need to execute full spectrum warfare hampered by an intelligence platform that is either frozen or rebooting for 10 hours out of a 96 hour exercise.

I’m certain that as the operators learn the ins and outs of the system, they’ll find some workarounds. And eventually, the software/hardware interface will improve to the point where there is some reasonable level of functionality. But the fact of the matter is, we’ve once again put so much stress on procuring an efficient system that the supposed efficiencies are less than we would have by buying three different systems, one for imagery, one for signals intelligence, and one for radar information. We’ve outsmarted ourselves.



*The E-8 JSTARS was a modified 707 radar plane. You’re probably familiar with the E-3 Sentry using its massive radar to track and control the air battlespace for hundreds of miles. The E-8 uses special radar to track movement on the ground.

2 thoughts on “Intelligence Retarded”

  1. Just a couple of points. While I am not sure why DCGS-A is locking up so much, it is not a new system. Been around a while. We also use multiple systems for intelligence analysis, many of which require manually moving intelligence from one system to another. Operations guys use CPOF (Command Post of the Future); companies and battalions use TIGR (Tactical Ground Reporting) while Bns and BCTs use DCGS-A and CIDNE (Coalition Information Dissemination Net-Centric Environment). As a non-MI guy, it is very cumbersome, to say the least. The process of collecting and analyzing raw intelligence and turning it into a finished product is also slow. TIGR and CIDNE were developed strictly for a COIN environment and probably have limited utility in Korea, but they are what intel folks have gotten proficient on lately.
    Lastly, just as an FYI, though your comments on JSTARs are largely on-target, we did use it in Iraq, and got pretty effective at it. We would use trend analysis to select where we wanted JSTARS to look, and then get up either rotary wing attack, or more often fixed-wing Non-Traditional ISR assets up (F15E or F16). When JSTARS detected Moving Target Indicators, then we could direct the fixed wing aircraft to go look at it. So, we had a hunter up, and could divert a killer to that location if necessary. My last ALO actually got an article written about it in some USAF TTP rag.

  2. Wow, I love this topic! I was a 98 series during my five years in (92-97) and while I can’t speak for the software problems (since we were on completely different systems back then) I CAN speak to some of the other items.

    The statement that MI was poorly prepared for the insurgency is both accurate and unfair at the same time. While the Army does have HUMINT assets, they’re strictly the interrogators. Those were just merging with the 98’s when I left service, so I didn’t know a lot of them, but I did know a few from DLI. Most MI (at least when I was in) was some form of SIGINT (98G was voice intercept, 98H was Morse, 98D was direction finding, etc.), and it made sense. When fighting a standing army, HUMINT is almost impossible to obtain (outside of POWs), and imagery was the sole province of DIA. They might pass data on to DoD, but the assets were theirs as was the analysis. So SIGINT was the only real game in town.

    We studied Soviet TO&E extensively. I still remember a lot of FM 100-2-3 even now. Again, it made sense. Most of the US’s potential foes used either the Soviet model, or the British model. And we weren’t particularly worried about the Indians all that much. But one feature of the Soviet model was the centralized Command and Control. While the US gives a lot of leeway for a BN commander to take his objectives, Soviet (and Iraqi, Iranian, North Korean, and indeed most hostile armed forces) gave local commanders no say. Many such commanders feared to move without orders. The advantage for us was twofold. We could get a LOT more data about what they were doing (obviously) and it made them very susceptible to jamming (another MI function). Insurgencies don’t tend to use modern communications in the same way as armies, nor do they tend to have a recognizable TO&E. So yes, MI was singularly unprepared for an insurgency, but by the same token, so was Combat Arms at the time.

    MI operates as Division, Corps and Echelon Above Corps (EAC) assets, because as you said, you only have so many warm bodies allowed, so they must be parceled out. To be fair, Artillery assets work the same way. Sure a Company CO can call for Arty fire, but he can also ask for an MI tasking on a particular area. It’s MI’s job to ID the unit he’s facing, and to get all the data on that unit they can, but it’s really up to Division to make sure that data ends up in that Company Commander’s hands.

    As for needing to keep the current “edge” developed for the insurgencies of Iraq and Afghanistan, I disagree with regards to North Korea. North Korea is an opponent much better suited to the “old school” MI, because they are so VERY heavily dependent on Command and Control. A Nork BN will not move without orders. And those orders will need to be communicated via signals. Their very command structure makes them inordinately vulnerable to SIGINT and their secrecy and closed society makes HUMINT nearly impossible.

Comments are closed.