Chapters 2 and 3 described the knowledge and skills required for a position in a core and emerging area, and Chapter 4 provided estimates of the number of experts (new graduates and experienced workers) in these areas. This chapter compares these results with information on the National Geospatial-Intelligence Agency’s (NGA’s) needs to identify gaps in the current or future availability of geospatial intelligence expertise (the committee’s Task 2). The committee examined gaps in domain knowledge and skills and where to find them. NGA’s current needs were estimated from information provided by the agency (see Box 1.3) or available on its website. In particular, the job listings1 and occupation descriptions for scientists and analysts (Appendix B) provide a measure of the knowledge and skills the agency is currently seeking, and the schools where NGA recruits potential employees indicate where the agency is looking for this knowledge and skills. The curriculum of the NGA College was assumed to reflect not only what topics are currently important to the agency, but also what knowledge and skills are hard to find in new employees.
Estimating NGA’s needs over the next 20 years is more difficult, in part because trends in hiring may have changed. Moreover, ongoing scientific and technological advances and evolving needs for geospatial intelligence continually change the skill sets needed. In addition, the bimodal age distribution of NGA’s scientists and analysts (Box 5.1) means that junior staff likely have different skills and analysis workflows than those nearing retirement. As these staff move into leadership positions, the agency culture will change, possibly attracting new recruits or accelerating the departure of some staff (see Wilkins and Ouchi, 1983; Carley, 2000; and Cameron and Quinn, 2006, for a discussion of changing organizational cultures). The cultural shift will also change what technologies are used and what skills are sought. Finally, the beginning of the age of big data (Manyika et al., 2011) and ubiquitous geospatial information are driving rapid growth in the geospatial industry as well as creating more competition for graduates with geospatial knowledge and skills (e.g., Gewin, 2004; DiBiase et al., 2006; Solem et al., 2008). The impacts of these changes are difficult to forecast, so the committee estimated NGA’s future needs based on the age distribution of NGA’s current geospatial intelligence workforce and the assumption that future hiring would focus on the core and emerging areas.
The Chapter 4 education and labor analysis yielded estimates of the number of new graduates with education in the core and emerging areas, as well as estimates of the number of experienced workers in closely related occupations. NGA generally hires several hundred people from these two sources each year. Below we compare the education and labor estimates with NGA’s needs for domain knowledge in the core and emerging areas over the next few decades.
The success of recruitment during the years following the September 2001 terrorist attacks in New York and Washington, D.C., led to a bimodal age distribution of NGA scientists and analysts. Compared to other federal agencies, NGA has a relatively young workforce, with only a small fraction of scientists and analysts over 60 years old If current staff retrre at age 65, the fi rst major round of retirements will begin by the end of the decade.
More than half of geospatial intelligence analyst positions at NGA specify degrees or coursework in Geographic Information Systems (GIS), geospatial analysis, geography, or geographic information science (Table B.1, Appendix B). Approximately 189 universities offer relevant degrees, and hundreds of community colleges offer relevant courses (Table A.5, Appendix A). In 2009, 5,404 U.S. citizens and permanent residents received a degree in geography, the instructional program that produces the bulk of expertise in GIS and geospatial analysis (Table C.10, Appendix C). The number of geography graduates far exceeds the number of geography jobs nationwide (1,300 jobs in 2010; see Table D.2, Appendix D) and the field is growing, suggesting that the supply of geographers will be sufficient for NGA’s needs over the next 20 years. On the other hand, GIS applications analysts are in high demand by the private sector, with qualified candidates difficult to find (Mondello et al., 2004, 2008; Solem et al., 2008). Given that the NGA College offers reasonably comprehensive coursework in GIS operations (Box 5.2), it is possible that competition from private companies is already making it difficult to find qualified experts in GIS applications and techniques.
Expertise in remote sensing is also important to NGA: remote sensing appears in the education requirements for nearly half of NGA scientist and analyst occupations (Table B.1, Appendix B), and a few thousand NGA scientists and analysts work on imagery analysis. The supply of remote sensing graduates is likely on the order of hundreds to thousands (Table 4.1). The supply of experienced workers in the most closely related occupation (physical scientists, all others) is 24,690 (Table D.2, Appendix D). Although the supply exceeds the number of NGA positions, the NGA College places heavy emphasis on remote sensing (Box 5.2), suggesting that extensive on-the-job training is already necessary for remote sensing and imagery analysis positions.
Compared to GIS and remote sensing, a relatively small number of NGA positions require specialized knowledge in cartography, geodesy and geophysics, or photogrammetry. A bachelor’s degree in cartography or at least 30 semester hours of cartography coursework is required for NGA analyst positions in cartography and photogrammetry (Table B.1, Appendix B). Only 155 U.S. citizens or permanent residents obtained a degree in cartography in 2009 (Tables C.6 and C.10, Appendix C), but there is a large supply of cartography and photogrammetry professionals (11,670), working mainly in the private sector (Table D.2, Appendix D). The NGA College offers minimal training in cartography (Box 5.2), suggesting that NGA is currently able to find enough qualified candidates. However, the agency is likely to face a shortage (i.e., numbers are too small to give NGA choices or means of meeting sudden demand) in the near future. Employer surveys have identified cartographers as among the most difficult positions to fill (Mondello et al., 2004, 2008; Solem et al., 2008). Moreover, cartography appears to be losing its identity as an academic discipline. Fewer colleges and universities offer degrees or certificates in cartography, and more students are choosing instead to pursue a specialization in geographic information science, remote sensing, or spatial analysis (see Chapter 2).
The situation is worse for photogrammetry, which has nearly disappeared as a field of study in academia. Only 15 universities offer photogrammetry classes (Table A.2, Appendix A), and only 26 U.S. citizens or permanent residents obtained a degree in a closely related field (surveying engineering) in 2009 (Table C.10, Appendix C). A degree in photogrammetry is not required for any NGA position, but coursework in photogrammetry is identified as useful for several occupations, including those related to photogrammetry, cartography, geodesy, and data collection,
The NGA College is an accredited institution housed within NGA that offers approximately 170 courses in geospatial intelligence, leadership, and professional development to government civilians, members of the military, and contractors to NGA and other U.S. defense and intelligence agencies.a The specific training required for new employees depends on the requirements of the position, along with the skills, education, and experience of the individual.b Classes are taught by government employees and contractorsc and typically last between 1 and 5 days. The longest class, basic geographic intelligence, runs about 7 months. About 15,000 students receive training in the college each year.
Nearly 40 percent of the classes offered at the college are related to remote sensing and offer a reasonably comprehensive suite of classes in data collection strategies, image processing, and major remote sensing systems, including infrared, multispectral/hyperspectral, radar/polarimetry, and motion imagery. The treatment of GIS operations using commercial products is also reasonably complete, but there is little coursework in geospatial analysis, such as spatial data analysis, spatial statistical analysis, or spatial optimization. None of the courses focus on geospatial data visualization and information design, even though NGA cartographers and other analysts work with graphics, imagery, movies, and maps.
Classes relevant to other core areas are sparse and introductory in nature. For example, no geophysics classes are offered. A few courses teach basic geodesy concepts; none deal with more advanced concepts, such as platform navigation, charting, Global Navigation Satellite Systems such as the Global Positioning System, or mathematics or statistics. Similarly, the only class offered in photogrammetry is taught at the introductory level, although some photogrammetric concepts, theory, procedures, exploitation techniques, and product quality issues are taught in the remote sensing courses.
Not surprisingly, the emerging areas are poorly covered in the current NGA College curriculum. For example, a few courses touch on methods to visually overlay disparate data, but none cover broader GEOINT fusion concepts such as ontology, the semantic web, schema In tegration, map conflation, or statistical methods of combining different types of evidence. Similarly, a few courses offer basic information useful to visual analytics (e.g., Google Earth and related applications) and to intelligence forecasting or scenario forecasting. Although two courses mention network analysis, the subtopics of strong relevance to NGA (dynamic network analysis and geospatial network analysis) are not covered. No courses discuss the use and limitations of crowdsourcing for creating maps and gathering data, although some of the relevant technologies (e.g., Google Earth, text mining) are covered.
c Presentation to the comm ittee by Mark Pahls, Chief of Learning Integration, NGA College, on May 23, 2011.
as well as to principal and project scientists (Table B.1, Appendix B). The NGA College offers only one introductory course in photogrammetry (Box 5.2), suggesting that qualified candidates are currently available. Much of the stock of trained photogrammetry professionals resides in private companies, including contractors to NGA. There were more than 7,000 jobs in car tography and photogrammetry in the private sector in 2010 (Table D.2, Appendix D). Although this source of experts may be sufficient for NGA’s needs in the short run, the lack of rigorous university training in photogrammetry will eventually yield a shortage of photogrammetrists qualified for a position at NGA.
Geodesy-related positions at NGA require a bachelor’s degree in geodesy, mathematics, physical science, or a related discipline (Table B.1, Appendix B). NGA has no specific positions in geophysics (or courses at the NGA College; Box 5.2), although coursework or experience in geophysics is identified as useful for cartography, geodesy, photogrammetry, and principal and project scientist positions. In 2009, 138 U.S. citizens and permanent residents received a degree in geophysics and seismology, and 26 received a degree in surveying engineering (Table C.10, Appendix C), the instructional programs that produce the most geophysicists and geodesists. Much larger numbers of experts were employed in 2010, including more than 30,000 geoscientists and more than 50,000 surveying and mapping technicians (Table D.2, Appendix D), the most closely related occupations. This supply is large relative to NGA’s current needs. However, the supply of graduates is small (on the order of hundreds) and only about one-third of these have advanced degrees and specialized training in geodesy. The small number of geodesy graduates, coupled with federal agency concerns about a growing deficit of highly skilled geodesists (NRC, 2010c), suggests that NGA may soon have to hire and train professionals from other disciplines. Indeed, the few geodesy-related courses at the NGA College appear to be geared toward analysts trained in other disciplines.
NGA currently has no science or analyst positions in the emerging areas, although some of the knowledge relevant to human geography is needed for NGA analyst positions in political geography, regional geography, regional source, and scientific linguistics. Consequently, any gaps in the supply of expertise in the emerging areas relative to NGA’s needs will occur in the future. It is likely that NGA’s need for expertise in the emerging areas will grow over time. The increasing availability of geospatial data and technology are allowing NGA to tackle increasingly complex intelligence problems, which commonly require interdisciplinary approaches (Box 5.3), such as those embodied in the emerging areas.
By their nature, training in the emerging areas is provided through individual courses often scattered among different university departments. Each program has a unique set of collaborating departments and approach for dealing with the topics, which creates difficulties for finding expertise. For example, different departments tend to explore different aspects of fusion, leading to multiple (and sometimes inconsistent) vocabularies and conceptualizations. Much of the technology development for human geography takes place in computer science, electrical engineering, and physics departments without reference to the large body of theoretical and empirical work in geographic and social science departments, leading to an increasing divergence between theory and methods. The lack of standard curricula, established journals, and even a common language means that graduates from different programs will have different knowledge and skills.
The other major gap associated with the emerging areas is the number of graduates. Fewer than a dozen universities offer specialized training in any emerging area except forecasting, and only a few universities offer a comprehensive degree program (Chapter 3). Anecdotal evidence suggests that many of these graduates are finding jobs quickly,2 so competition, combined with a small supply (tens to hundreds in most emerging areas; see Table 4.1), could lead to shortages in the future availability of expertise in the emerging areas.
A common critique of disciplinary science is that it leads practitioners to look inward and to create numerous subspecialties in what scholars have called the fragmentation of disciplinarity (Strober, 2006) or stovepiping. Countering this tendency is the more recent recognition that scientific breakthroughs often happen at the edges and intersections of disciplines and specialties (Kates, 1987). These intersections occur at a range of scales. Multidisciplinary approaches involve people with different skill and knowledge sets working together, such as a geodesist working with a cartographer as part of a geospatial intelligence team, and they require an infrastructure for information sharing, such as a control room or social network. Interdisciplinary approaches require people to train across multiple fields (e.g., astrobiology). People with interdisciplinary skills may act as catalysts to problem solving, particularly when no approach seems suitable within an existing discipline (e.g., Omenn, 2006). Finally, transdisciplinary research problems are too large and complex to solve by any one discipline (Jantsch, 1972). Examples of transdisciplmary projects include climate change research, mapping the human genome, and testing the laws of physics using the Large Hadron Collider.
Few universities have succeeded in training interdisciplinary students because college and departmental structures often discourage the approach, and only a handful have mastered multidisciplinary approaches. Once created, interdisciplinary programs are hard to maintain because peer-review processes are commonly organized along traditional discipline lines. Most interdisciplinary training takes place at the graduate level. However, undergraduate students can ach1eve these goals by choosing double majors; multiple minors; and interdisciplinary, self-guided, and mixed-mode majors. For example, many students study abroad, create internships, do voluntary work, and seek out accreditation and certificate programs. Such combinations may eventually outnumber more traditional majors.
The distinction between knowledge and skills is not always clear, especially for the geospatial field, which can be viewed as a discipline, a collection of tools, or a profession (DiBiase et al., 2006, 2010). In 2010, the Department of Labor’s Employment and Training Administration issued a geospatial technology competency model to define the scope of disciplines and the training and credentials required to work in the geospatial technology industry. The model lays out tiers of competencies, or capabilities for using sets of knowl edge,
2 The emerging areas can be considered data science jobs— those requiring expertise in multiple technical disciplines, such as computer science, analytics, math, modeling, and statistics. Such jobs are expected to see a shortage of 190,000 data scientists by 2018 (Bertolucci, 2012).
skills, and abilities to successfully perform specific tasks (Figure 5.1). Tiers 1–3 describe general workplace behaviors and knowledge needed in most industries, including personal attributes learned at home (e.g., interpersonal skills, integrity), knowledge and skills learned in academic settings (e.g., geography, communication, basic computer skills), and skills honed in the workplace (e.g., teamwork, creative thinking). Tier 4 describes subjects (e.g., remote sensing, GIS, programming) and background knowledge (e.g., analytical methods, geospatial data) needed by many geospatial professionals in their careers. Tier 5 specifies clusters of subject and background knowledge needed for each of three industry sectors: positioning and geospatial data acquisition; analysis and modeling; and software and application development. Above these tiers are competencies required for specific occupations (e.g., cartographers and photogrammetrists) and managers.
NGA occupation descriptions specify a set of core competencies for all science and analyst positions as
FIGURE 5.1 Geospatial technology competency model. SOURCE: Department of Labor, <http://www.careeronestop.org/competencymodel/pyramid.aspx?GEO=Y>.
well as the skills required for each type of position. The core competencies and skills span all levels of the geospatial technology competency model, although the core competencies stress interpersonal skills, communication, and creative thinking and adaptability, whereas the position-related skills stress working with customers and gathering, analyzing, and disseminating information. The most common skills among NGA science and analysis positions are illustrated in Figure 5.2.
The NGA College offers several courses in interpersonal skills, effective communication, and critical thinking, suggesting that these core competencies are in short supply. These skills are taught in some university programs, and new ways of teaching may also help fill the gap. For example, techniques such as role playing, gaming, and self-assessment favor understanding and conceptual methods, rather than content and memorization.
In the foreseeable future, new questions, as well as the data sets and tools needed to answer them, will continually arise. Dealing with these evolving questions and approaches requires a flexible workforce that is capable of thinking in breadth, rather than depth, through interdisciplinary training and teamwork. Historically, NGA employees acquired the necessary breadth of skills through an undergraduate education in a relevant discipline, internships or service, and/or training through the NGA College. However, the increasing demands of teamwork and of multidisciplinary and interdisciplinary analytical tasks are placing increasing importance on a broader set of skills.
If there were such a thing as an ideal geospatial intelligence analyst, he or she would be well versed and expert at spatial thinking; have considerable interdisciplinary training; be well traveled and knowledgeable of world cultures (and able to use tools such as Google Earth for rapid virtual travel); have some core background in statistics, cartography (coordinates, projections, scale), and computer science (programming principles, operating systems); have a high degree of science literacy; read and write multiple languages; and have a commitment to professional ethics. None of these skills are classified as core competencies of NGA scientists and analysts, and skills in statistics, ethics, cultural analysis, and scientific methods are required only for certain NGA positions. Consequently, it is likely that NGA scientists and analysts are missing skills that will be important for future work in the core and emerging areas.
University departments commonly teach some of these skills. Spatial literacy and spatial reasoning are finding their way into undergraduate and graduate geography curricula nationwide (NRC, 2006). Spatial thinking is highly interdisciplinary, an extension of efforts to bring methods from GIS and spatial analysis into the social sciences and humanities. For example,
FIGURE 5.2 Word cloud illustrating the 25 most common skills identified in job descriptions for NGA scientists and analysts in the five core areas. The most common topics are portrayed by the largest lettering, and an arbitrary color scheme is used to distinguish the various phrases. SOURCE: Generated using <http://www.wordle.net>.
the Center for Spatially Integrated Social Science3 was a 5-year National Science Foundation project designed to expand the knowledge and use of GIS and spatial methods in the social sciences, including demography, sociology, landscape architecture, and other disciplines.
Computer programming skills are needed for many of the core and emerging areas. For example, dealing with big data in geospatial intelligence (GEOINT) fusion, forecasting, visual analytics, and human geography requires skills in database management and construction for large data sets, natural language processing and text mining for large text data streams, social media mining, and streaming image or video processing. These skills are generally learned in computer science, information systems, or information technology programs. Even when data volumes are modest, computer programming skills are needed for writing scripts to encode image analysis and processing steps, implementing algorithms, understanding methods such as tracking and optimization, and communicating effectively with programming staff.
Other skills required for most of the core and emerging areas include statistics, network theory, and advanced mathematics. However, many geography departments, where the cartography and geographic information science specializations are commonly housed, no longer require calculus, statistics, or basic programming, and they have never required network theory. Students in geography do not naturally drift toward coursework in these areas, and it is difficult to teach someone to map residuals, for example, when he or she does not understand means and variance, root-mean-square error, or even the difference between a standard deviation and an interquartile range. Engineering and computer science students have some of this training (particularly in computer programming and advanced mathematics), but they generally have few spatial skills.
Similarly, advanced quantitative skills are required for forecasting, which is based on analog (e.g., similar patterns), analytical (e.g., physical or mathematical), statistical (e.g., deterministic, stochastic), or computational (e.g., numerical models, data-model assimilation) methods. Geospatial forecasting needs to connect components and interactions from physical, social, and cultural systems. However, the majority of GIS or social science students lack adequate mathematical capabilities for geospatial forecasting, although the number of social science students in programs that emphasize statistics, agent-based modeling, and social networks is growing. Students in the physical, environmental, or life sciences generally have better quantitative skills, but they lack the abilities to handle the diverse, uncertain, and culturally and geographically dependent nature of the human dimension.
Other quantitative methods useful to many of the core and emerging areas include visualization and graphics design, modeling and simulation (usually left for graduate school), and the analysis of geospatial data from social media. For example, the suite of software commonly used by students has broadened from standard statistical packages and GIS to include visual analytics, semantic web, content analysis, and others. Standard, often commercial packages have rapidly yielded to extendable “mashups” of open-source software, although few university programs take advantage of this rapid expansion in the type and nature of analytical tools.
Finally, students commonly lack capabilities in the qualitative methods (e.g., interviews, questionnaires, textual content analysis, ethnographic assessment) that are often needed in addition to the quantitative methods discussed above. Few programs teach these methods, despite their importance to many research fields.
Overall, changes in university programs are making some skills needed by NGA scientists and analysts harder to find (e.g., cartographers with math and programming skills) and others easier to find (e.g., geographers with spatial thinking skills). The emergence of interdisciplinary areas such as GEOINT fusion, visual analytics, and human geography is beginning to yield graduates with skills from several university departments (e.g., computer science and spatial skills). However, until these programs develop, individuals with the ideal combination of skills for NGA are likely to remain in short supply.
NGA focuses recruiting on dozens of colleges and universities that are near major NGA facilities (i.e., Springfield, Virginia; Saint Louis, Missouri) or that
have a large population of underrepresented groups (e.g., historically black colleges and universities). Few of these institutions have significant programs in core or emerging areas, although they likely meet other agency goals, such as increasing diversity. About one-third of the schools and universities where NGA recruits are large state universities, and several of these (e.g., George Mason University, Ohio State University, Pennsylvania State University, University of California, Santa Barbara) offer education and training in several core or emerging areas. Extending recruiting to some of the example universities listed in this report (e.g., Tables A.1–A.11, Appendix A) would help NGA find individuals with knowledge and skills in core and emerging areas.
The second task of the committee was to identify gaps in the current or future availability of expertise relative to NGA’s needs. The Chapter 4 analysis showed that the number of new graduates with education in core and emerging areas and the number of experienced workers in closely related occupations far exceeds NGA’s needs for expertise in all core and emerging areas (generally several hundred people a year). However, when other considerations are factored in—including competition from other organizations and the extensive training provided by NGA in some areas—a more nuanced picture emerges. Expertise in geophysics and geospatial analysis is likely sufficient for NGA’s current and future needs. NGA hires only a small fraction of the available experts and offers little or no training in these areas to employees through the NGA College. The supply of experts in cartography, photogrammetry, and geodesy appears adequate for now. The number of professionals working in these areas is substantially higher than the number of NGA job openings, and only minimal training is offered at the NGA College. However, some shortages are likely in the future because photogrammetry, geodesy, and cartography programs produce a small number of graduates, and the number of academic programs in photogrammetry and cartography is shrinking. Moreover, employer surveys suggest that skilled cartographers and geodesists are hard to find. Shortages may already be appearing in GIS and remote sensing, given the extensive training in these fields provided by the NGA College. Although the supply in both fields exceeds NGA’s needs, competition for GIS applications analysts is strong. By definition, NGA has no current positions for experts in emerging areas, but as the agency tackles increasingly complex geospatial intelligence problems, demand for the types of interdisciplinary approaches embodied by the emerging areas is likely to grow.
In addition to domain knowledge and interdisciplinary skills, NGA scientists and analysts need a variety of personal, academic, and workplace skills. The NGA College offers several courses in interpersonal skills, written and oral communication, and critical thinking, suggesting that these skills are currently in short supply. In NGA’s future workforce, which is likely to be more interdisciplinary and focused on emerging areas, the ideal skills will include spatial thinking, scientific and computer literacy, mathematics and statistics, languages and world travel, and professional ethics. These skills are not always taught in university programs. Although spatial thinking is increasingly being taught in undergraduate programs, math and computer skills remain a gap in many natural and social science programs, and spatial perspectives remain a gap in most computer science and engineering programs.
Individuals with the knowledge and skills needed for a geospatial intelligence position at NGA are available, but NGA may not be looking for them in all the right places. Only about one-third of the universities and colleges where NGA currently focuses recruiting have strong programs in core or emerging areas. The academic institutions discussed in this report may provide a useful start for finding programs in core and emerging areas.
In summary, the analysis for Task 2 revealed both current and future gaps in knowledge and skills relative to NGA’s needs. Although the supply of experts is larger than NGA demand in all core and emerging areas, competition may be making GIS and remote sensing experts hard to find. Long before 2030, competition and a small number of graduates will likely result in shortages in cartography, photogrammetry, geodesy, and all emerging areas. In NGA’s future workforce, which is likely to be more interdisciplinary and focused on emerging areas, the ideal skill set will include spatial thinking, scientific and computer literacy, mathematics and statistics, languages and world culture, and professional
ethics. Although NGA is currently finding employees with skills in statistics, ethics, cultural analysis, and scientific methods, graduates with the ideal skill set will remain scarce until interdisciplinary and emerging areas develop. NGA could improve its chances of finding the necessary knowledge and skills by extending recruiting to the example university programs identified in this report.