In her recent keynote address, the new director of the Nation Geospatial-Intelligence Agency (NGA), Letitia Long, unveiled her new vision, which recognizes certain realities of the operating environment and implicitly touches upon profound implications for the GEOINT analytical community. This unveiling suggests the need for a deeper look at the environment, those implications, and the obstacles to successful implementation.
In the presentation of her vision, Long described two key goals for the future of NGA:
Provide online, on-demand access to GEOINT knowledge by giving users access to all GEOINT content, services, expertise and support, while providing tools that allow them to serve themselves; and
Create new value by broadening and deepening NGA’s analytic expertise, through a deeper contextual analysis of places informed not only by the earth’s physical features and imagery intelligence, but also by “human geography.”
Long’s plan would offer users the ability to leverage GEOINT directly online and on-demand, giving them the tools to generate their own products in an automated and individualized capacity. Long intends to move the agency into the 21st century and away from “labor intensive… brute force” production and analytical methods.
In short, she suggests that the imagery products currently provided neither properly leverage the resources that the community will have available to it nor adequately satisfy the needs of the users of the future. The community must find a way to harness the mass volumes of data available, to offer value-added products and services, and to do so on a timeline that meets operational needs.
The battlefield and intelligence operations of the 21st century will continue to evolve, with emerging threats and dynamics. The environment will be flush with contradictions. In population-centric operations, traditional surveillance and reconnaissance missions will continue, but new, focused intelligence questions about grassroots activity will confront the community.
U.S. forces will be required to manage and monitor vast amounts of open, unpopulated, ungoverned space, while operating in a densely active urban environment. Large volumes of data from multiple sources will directly support individual soldiers through unique and specifically developed applications. Dealing with indigenous populations will have large, far-reaching strategic implications, yet the dynamics and terrain will be vastly different for each application. Forming knowledge and developing operational plans will be more decentralized and occur on considerably shorter timelines.
On the future battlefield, adversary decision processes will become more and more decentralized, creating many small, semi-autonomous command and control structures. The urban operating environment will present U.S. forces with the task of discovering, isolating, and countering disparate pockets of militants. Targets will be increasingly mobile.
Friendly locals will be intermixed with adversaries. Dynamics will quickly change and require an unprecedented physical and cognitive agility. Local situational awareness will draw data from multiple sensors and provide a near real-time operating picture for combat forces. Operational windows of opportunity for U.S. troops to get in safely and get out of harm’s way will emerge and disappear quickly.
U.S. forces will continue to integrate sensor-to-shooter concepts into the future battlefield. Further compression of timelines will prioritize process automation over human cognition. The environment will require visualization of vast amounts of data and automated analytics. To enable sensor-to-shooter integration, solutions will link situational awareness to fire control systems. Information systems will increasingly become part of weapons platforms, either in a physical or net-centric context — i.e., finished targeting solutions will be calculated dynamically and autonomously on the weapon platform or sent to it remotely.
This technology will proliferate to each echelon, down to dismounted soldiers. Battle management systems will become more tactical in nature. These solutions will forfeit information completeness for its timeliness — “good enough” information delivered quickly will be more important than perfect information later.
To be successful, these analysts and the other stakeholders within the National System for Geospatial-intelligence (NSG) must consider the full potential of terabytes of data and determine how knowledge is best created and delivered from it. The Intelligence Community has long had a problem of being unable to manage the full breadth and volume of data collected. The data it can handle are largely single-source oriented and therefore limited in value. The future will only compound that problem.
Re-use, re-purposing, and preservation of multi-source data will be commonplace. More sophisticated data from a greater proliferation of sensors will place tremendous demands on the community’s information technology infrastructure.
As an indicator of what is to come, Wide Area Airborne Surveillance missions have increased collection by 2,250% in the last two years alone, creating significant burdens on the theater-based GEOINT infrastructure and overloading the forensics capability of CONUS-based (continental U.S.-based) analysts. Foreshadowing the future, DoD budgets have tapped the industrial base to full capacity in requiring continued manufacture of these aircraft over the next several years.
To make matters more complex, to be valuable to the users, intelligence data will require greater context. Traditional imagery products without context will become a distraction from solving future problems. Users are insisting on spatiotemporal data mining and knowledge discovery from multi-source data streams merged into a visual representation that they can digest. Slowly, they have been educated on the value of non-literal analysis from nontraditional phenomenology, like synthetic aperture RADAR (SAR), and will demand more.
The changes implied by the new vision would fundamentally alter the user’s interaction with GEOINT and the NSG. Proliferation of hand-held devices would give users interaction “with dynamic content and services themselves — if and when they want — through online, on-demand access to global seamless foundation, imagery, product, and activity layers,” according to Long. The NSG would provide users with a direct connection to raw GEOINT data with automated applications that could determine landing zones, allow geo-tagging, or provide access to open-source street maps.
The new vision will fundamentally change the dynamics of today’s analytical process. In this environment, human analysts will do the things that people do well, while machines will do the rest. For example, the campaign for automatic target recognition has proved successful in some situations while not in others. Foundational elements of analysis will occur in near real-time, often decentralized from traditional and core analytical capability of the NSG. The responsibilities of today’s GEOINT analyst will shift to value-added services and products. This shift will be a continuous one within the analytical community, because as technology and demand evolve, new value-added products and services will (and must) become foundational.
In this new world, foundational products may not require a human, while some data can be pushed directly to users. Product development would largely be automated and exposed to users on the web or on deployed hardware, exploited by automated applications. The implications of this alone could free up a significant number of analysts to perform more complex and value-added assessments across the intelligence community. For example, a city map product can burden a single analyst for up to six weeks in production; automation would accomplish this in days.
A responsive online, self-serve capability will free up additional analysts to focus on assisted service and full service support. This will occur through development and provision of deeper knowledge and expertise of analysts. Developing the new expertise of an analyst will require a significant up-front investment in independent research, focused training, and expect-ation of time. It will demand continual learning through the analytical process to understand, appreciate, and apply, as Long calls it, human geography — visually depicted data and information that can be understood spatially and/or temporally and that deepens the knowledge of a specific place. Human geography is about the dynamics of people in certain regions – their history, culture, and patterns as a group. This is different from HUMINT, which in this context is a collection of information about specific individuals or groups of interest – their patterns and intentions.
In the future, GEOINT analysts will be asked to think critically about what they find on a traditional image rather than simply to manufacture the image. Instead of locating two mobile missiles, marking them, stamping the image and sending it on, analysts will use their knowledge of the region to give a deeper understanding of the implications of two missiles in their specific configuration, on that day, in that location.
As analysts are able to provide this new expertise and knowledge, the interface between analysts and users will change. Analysts will answer direct questions from NSG users. They will collaborate with users and other analysts alike on new or complex products. They may simply provide more service than before in solving users’ intelligence problems. In this model, GEOINT analysis will broaden its scope, while becoming timely, customized, and responsive to the users’ specific needs.
However, the community must exercise caution in this respect: analyst-user interaction must not be limited to answering user-defined questions. This would create a myopic, biased view of the external environment. Future analysts must continue to ask, answer, and offer their own questions. They must use their broader understanding of the external environment to pursue alternative paths of logic and to offer competing recommendations when applicable.
Analytical products will enable GEOINT entity development and mapping focused on extraction and classification of features, attributes, or identities. Analytical ability to capture and characterize unexpected linkages, correlations, and patterns through automated data and entity association will be critical. Development of spatiotemporal relationships and signatures, often performed statistically, will produce knowledge within regions, while offering similar patterns of discovery applicable elsewhere.
As an indicator of what is to come, Wide Area Airborne Surveillance missions have increased collection by 2,250% in the last two years alone, creating significant burdens on the theater-based GEOINT infrastructure and overloading the forensics capability of CONUS-based analysts.
Future analysis in this era will move to a more anticipatory posture. GEOINT analysts will prioritize geographic regions based on political trends, economic stability, and transnational issues. Focused GEOINT analysis of each region will provide baseline knowledge of its evolving dynamics long before demand makes it urgent. Users will be able to call upon these products as they want in order to hone their interests and needs.
The volume of analysis required will drive the structured and repetitive portion of this work to automation. Human analysts will then create finely tuned products layered upon this foundation that describe the nuances and details that only a human could understand and produce.
The anticipatory nature of GEOINT analysis will be of great benefit to other analysts across the community. It will allow analysts to reduce search space during crises. Through availability prior to a crisis, more time will be available for analysts to understand the reality of a geographic region and its circumstances. Increased analytical time and the availability of ready-made products will enable the development of a greater number of analytical alternatives, create a better opportunity for competing analysis, and, correspondingly, produce robust modeling and simulation capabilities for analysis within the Intelligence Community.
Collection management processes for other intelligence sources will be better informed and increase the efficiency of those collection platforms and the focus of future analysis. The end result from a greater number of alternatives and increased timelines will be better-informed decisions from policymakers, planners, and battlefield commanders.
Ultimately, this vision should create better knowledge for the end user, but several obstacles must be overcome. Three problems in particular have plagued, and will continue to plague, the movement of the NSG in this direction.
The most significant obstacle is culture. The imagery analysis community must be open and willing to embrace these changes for them to be successful. However, the community has traditionally argued against such changes to their world. The first reason is that some refer to a culture revolving around the “Sanctity of the Seal” as a valid concern that must be addressed. Historically, the analytical community mass-produced products upon which users’ lives depended; the NGA seal symbolized a high level of analytical confidence.
In this culture, the analytical products must be completely accurate, having been verified over time. There will still be a need for dissemination of fully analyzed and high-confidence products, but the new NGA vision moves away from this time-consuming process. The new system will have to accommodate analytical confidence with the user’s need for expediency and application.
The second obstacle is that traditional imagery analysts are comfortable extracting objects and describing activity in an image, a practice known as imagery intelligence (IMINT). The new vision asks analysts to move beyond IMINT to GEOINT and to describe what the presence, activity, and location of an object mean. It encourages analysts to add to the body of knowledge about the hard problems facing the intelligence community. This represents an uncomfortable change in knowledge and activity that will take some time to understand and accept.
The final problem involves processes used to develop analytical tools and applications. Unfortunately, as the Analytical Paradox suggests, as the amount of collected information grows, so grows the complexity of the problem, and, paradoxically, the difficulty of establishing a clear conclusion grows also. One would expect that the more data, the easier finding a solution would be; this is not the case.
Analytical tools and applications must address this difficulty by filtering valueless data and normalizing decision-oriented products. Incorporating a new approach to development can accomplish this. Development and use of products in the future data environment will be driven by collaboration and innovation. Quality development of tools and applications of the future will come from strong relationships between analysts and users, where the consumer in essence becomes the producer.
Tools and applications will be tied to hard problems and recurring tasks, automatically sifting through data with analyst-established, rule-based, predefined indications and warning events. They also will be flexible enough to allow users to ask penetrating questions and to automatically notify users of significant change events requiring action.
Long has set NGA on a necessary path for the future. Implementation of her vision will evolve the organization ahead of the environment and put it in a position to lead the Intelligence Community. However, a great deal must be overcome to succeed. At the core of this change sit the GEOINT analysts. They are being asked, in essence, to lead in a time of change, to do more themselves, and to rise to the challenge, as they have done in the past. The IC can expect no less of them in the future.