10
How Might We Better Observe, Analyze, and Visualize a Changing World?

The observation, mapping, and representation of Earth’s surface have been an integral part of geographical research since Ancient times, and these activities remain central to the modern geographical sciences. Maps of Earth’s surface have provided an essential underpinning to a long series of research breakthroughs, such as the theory of continental drift and plate tectonics, our understanding of patterns of early human migration, inferences about the causes and transmission of disease, and global climatology. Today, the tasks of documenting, visualizing, and understanding the increasing rate of change that we observe on the planet’s surface are far more complex than traditional mapmaking, demanding an increasingly elaborate network of Earth-observing satellites, ground-based observers and sensors, servers, and broadband communications, and the tools to analyze, model, and visualize. Collectively, these systems exploit decades of research by geographical scientists, represent an infrastructure investment that accounts for tens of billions of dollars annually, and address needs in all walks of life, from the work of scientists striving to understand and perhaps to predict, to the broader interests of the media and the general public, to the needs of emergency managers, planners, and a host of other professionals.

In recent years, the rate of development of new tools and data sources has been spectacular. Today most computer users are familiar with sites such as Google Earth, Microsoft’s Virtual Earth, and other virtual globes that allow them to visualize the current physical appearance of Earth’s surface using imagery obtained from satellites, aircraft, or specially equipped vehicles. Many people make daily use of navigation systems, and of the host of Web-based services that now link diverse sources of data based on geographical location (Box 10.1). However, despite these developments, the geographical tools, knowledge, and understanding with which humanity faces the uncertainties of the 21st century are inadequate in many respects. Surprisingly, conventional mapmaking has been in decline worldwide for several decades (Estes and Mooneyhan, 1994), and many basic maps, including those covering the United States, are no longer updated. Although Web-based virtual globes display stunningly detailed images of how the geographical world looks, they provide little sense of how the world is changing, how it will look in the future, how certain or uncertain we may be about future states, and how the world works as an interconnected system. The general public makes use of these technologies to obtain driving directions and to look at recent events around the world, but gains very little understanding of the many problems that humanity faces, and the options available for dealing with them that have been discussed in previous chapters. The methods of analysis that search for pattern, anomaly, and correlation in geographical information reflect the needs of an earlier, slowly changing world and its somewhat leisurely pace of investigation, rather than the rapidly changing planet of the new century, and its need for quick, science-based response.

Geographical data and tools are now essential for the full range of research disciplines that study the Earth’s surface and near surface, from the social sciences



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 97
10 How Might We Better Observe, Analyze, and Visualize a Changing World? T he observation, mapping, and representation from satellites, aircraft, or specially equipped vehicles. of Earth’s surface have been an integral part Many people make daily use of navigation systems, of geographical research since Ancient times, and of the host of Web-based services that now link and these activities remain central to the modern diverse sources of data based on geographical location geographical sciences. Maps of Earth’s surface have (Box 10.1). However, despite these developments, the provided an essential underpinning to a long series of geographical tools, knowledge, and understanding research breakthroughs, such as the theory of conti- with which humanity faces the uncertainties of the nental drift and plate tectonics, our understanding of 21st century are inadequate in many respects. Surpris- patterns of early human migration, inferences about ingly, conventional mapmaking has been in decline the causes and transmission of disease, and global worldwide for several decades (Estes and Mooneyhan, climatology. Today, the tasks of documenting, visual- 1994), and many basic maps, including those covering izing, and understanding the increasing rate of change the United States, are no longer updated. Although that we observe on the planet’s surface are far more Web-based virtual globes display stunningly detailed complex than traditional mapmaking, demanding an images of how the geographical world looks, they increasingly elaborate network of Earth-observing provide little sense of how the world is changing, how satellites, ground-based observers and sensors, servers, it will look in the future, how certain or uncertain we and broadband communications, and the tools to ana- may be about future states, and how the world works as an interconnected system. The general public makes lyze, model, and visualize. Collectively, these systems use of these technologies to obtain driving directions exploit decades of research by geographical scientists, and to look at recent events around the world, but gains represent an infrastructure investment that accounts for very little understanding of the many problems that tens of billions of dollars annually, and address needs humanity faces, and the options available for dealing in all walks of life, from the work of scientists striving with them that have been discussed in previous chap- to understand and perhaps to predict, to the broader ters. The methods of analysis that search for pattern, interests of the media and the general public, to the anomaly, and correlation in geographical information needs of emergency managers, planners, and a host of reflect the needs of an earlier, slowly changing world other professionals. and its somewhat leisurely pace of investigation, rather In recent years, the rate of development of new than the rapidly changing planet of the new century, tools and data sources has been spectacular. Today most and its need for quick, science-based response. computer users are familiar with sites such as Google Geographical data and tools are now essential for Earth, Microsoft’s Virtual Earth, and other virtual the full range of research disciplines that study the globes that allow them to visualize the current physical Earth’s surface and near surface, from the social sciences appearance of Earth’s surface using imagery obtained 

OCR for page 97
 UNDERSTANDING THE CHANGING PLANET BOX 10.1 Linking Information Sources with Geographical Location Geographical location is now widely used on the Web as a means to link two or more existing sources of information to create a new service, providing information that would otherwise have been unavailable. These combined services are commonly known as mashups, from a term used in the recording industry to refer to music created by combining previously independent tracks. Together they form what is often termed the GeoWeb, a spider’s web of services held together through the use of geographical location, whether expressed as latitude/longitude, place name, street address, or any other convenient system of georeferencing. As an example, houses listed for sale on Craigslist (www.craigslist.com) are first analyzed by a mashup service to obtain their street addresses, which are then converted to latitude/longitude using a Web-based geocoding service (see Figure). The mashup service, www.housingmaps.com, combines these coordinates with a mapping service (Google Maps, maps.google.com) to provide detailed and useful maps of where houses are listed for sale, along with other information about the house—something that neither Craigslist nor Google Maps is capable of providing on its own. A map of houses currently listed for sale in the $150,000-300,000 price range by Craigslist for part of Los Angeles. This www.housingmaps.com mashup combines Craigslist data with Google Maps. SOURCE: www.housingmaps.com.

OCR for page 97
 OBSERVE, ANALYZE, AND VISUALIZE such as criminology or political science to the environ- ing” phenomena with the geographical knowledge mental sciences such as climatology or ecology. They and understanding that are embedded in today’s tools are an absolute necessity if scientific understanding is to and data acquisition systems, or with the processes of be translated into effective evidence-based and place- reasoning and inference that translate raw observations based policy, and if the average citizen is to become into scientific knowledge and ultimately into improved fully informed and engaged in discussions about poten- decisions. Even though it is possible to make a decent- tial uses and abuses of geographical technologies. looking and informative map by doing no more than Earth observing satellites provide powerful obser- accepting the default options of a software package, vations of the physical parameters that characterize the science of cartography is still an active field. Re- the Earth system, including sea-surface temperatures, mote sensing, a workhorse of modern geographical rainfall, and land cover. Geographical scientists have data acquisition, continues to advance, making use benefited from the massive investments that have been of new kinds of sensors that exploit different parts of made not only by U.S. government agencies, but also the electromagnetic spectrum, access a broader range by the private sector and by agencies in other countries. of acoustic frequencies in the case of marine remote They have discovered effective means of turning raw sensing, and achieve finer levels of spatial, temporal, or sensor measurements into characterizations of the land spectral resolution, and each new development opens surface, and have applied these in numerous ways, many opportunities for research into additional applications of them discussed in other chapters of this report (e.g., in the geographical sciences. Chapters 2, 3, and 4). However, many of the social None of these developments, however, has in any aspects of the Earth system cannot be observed or way reduced the importance of expertise in reasoning measured from above, and understanding of them must and inference, and it is clearly in this area that the role therefore rely on people’s willingness to divulge sensi- of the geographical scientist is most critical (Longley tive information about themselves in programs such as et al., 2005). No one would suggest, for example, that the U.S. Census. technology has in any way reduced the need for expert pedologists, both as creators of knowledge about soils and as interpreters of that knowledge to users, even role oF The geograPhical scieNces though that knowledge may be expressed through The activity of describing and drawing the world—the the simple medium of a map. Similarly the expertise literal root of the term geography—has changed of the cultural geographer in observing and interpreting dramatically over the centuries. An editorial in Nature the human landscape can never be encapsulated in soft- (Editorial, 2008) went so far as to argue that our ware. Geographical scientists understand the nature of museum collections are full of specimens collected at particular classes of phenomena through their training vaguely recorded locations (Guo et al., 2008), but there and knowledge of the literature. They are aware of the is now no longer any excuse for not recording the loca- importance of spatial concepts such as scale, location, tion of any scientific observation made of phenomena place, and interaction (Gersmehl, 2005); are familiar on or near Earth’s surface. For many researchers much with their underlying theories and the pitfalls associ- of the science of cartography, or the drawing of maps ated with each; and are capable of making effective use that accurately convey knowledge, has been encapsu- of the geographical data and tools at their disposal to lated in a series of default options in software. Access augment the sum of human knowledge about the geo- to the products of remote sensing, once a task requiring graphical landscape, while remaining cognizant of the a great deal of technical knowledge, is now reduced inevitable uncertainty associated with that knowledge. for many users to the manipulation of the simple user Past investments in research into geographical interface of a desktop or laptop computer, or even a data and tools have produced dramatic progress. The hand-held PDA (personal digital assistant). stunning zooms and pans of the virtual globes rely on However, although it has undoubtedly become fundamental research by geographical scientists into easier to work with geographical data and tools, we ways of capturing the curved surface of the planet in should not confuse the act of “describing and draw- digital form, a research area known as discrete global

OCR for page 97
00 UNDERSTANDING THE CHANGING PLANET FIGURE 10.1 Illustrations of icosahedron­based discrete glob­ al grids at several resolutions. SOURCE: Sahr et al. (2003). grids or geodesic grids (Figure 10.1), and on research in computer graphics into managing the level of detail in various parts of the visual field (level-of-detail manage- FIGURE 10.2 Soil map of the Yucatan Peninsula compiled ment). Research on geographical semantics is making from datasets from three different countries: Mexico (brown), it possible to integrate data across national boundaries Belize (red), and Guatemala (blue). For clarity, only the bound­ despite differences in mapping practice, terminology, aries between areas of homogeneous soil are shown. Note the different densities of boundaries, indicating differences between classification schemes, and data formats (Figure 10.2; national standards and practices (and also state standards in Goodchild et al., 1999). Much more is known now Mexico). SOURCE: Sifuentes (2005). about the sources of uncertainty in geographical data, about visualizing its importance, and about model- ing its impact on the results of analysis (Zhang and Goodchild, 2002). New techniques of geovisualization tunities created by advances in mainstream information allow geographical data to be displayed simultane- technology, by new sensors, and by new ways of think- ously from many different perspectives, leading to ing about the role of the geographical perspective as a new insights and the generation of new hypotheses cross-cutting theme in science. The following questions (Figure 10.3). Significant advances have been made illustrate lines of research that would be particularly recently by geographical scientists in areas such as visu- productive to advance this topic. alizing dynamic, multidimensional phenomena (Dykes et al., 2005; Slocum, 2009) and adapting mapmaking research suBQuesTioNs to the needs of the visually impaired (Golledge et al., 2005; Rice et al., 2005). All of these make use of digital how will new geographical knowledge be acquired technology, but require expertise not only in computer and shared? science but also in the disciplines that are intimately Many of the more recent developments in geo- concerned with phenomena on Earth’s surface. There graphical technology relate to data gathering. Satellite is no doubt that progress will be at least as dramatic in remote sensing ( Jensen, 2007), which dates from the the future, as geographical scientists explore the oppor-

OCR for page 97
0 OBSERVE, ANALYZE, AND VISUALIZE FIGURE 10.3 Screenshot from the software package GeoDa, developed for the exploration and analysis of geographical data and available free for download from the GeoDa Center at Arizona State University, geodacenter.asu.edu/. This example shows data on the rate of sudden infant death syndrome by North Carolina county from four different perspectives: map (upper left; darkest colors depict the highest rates), table (lower left), histogram (lower right); and in the upper right, a scatterplot of each county’s rate against the average of its neighbors, a plot that allows the easy identification of counties with anomalous rates that stand out from their neighbors. The rectangle, which can be maneuvered by the user, results in the highlighting of contained counties in every window. SOURCE: GeoDa 0.9 (Beta). spy satellites of the Cold War, has evolved into a multi- War II, has grown exponentially in importance, given national enterprise producing terabytes of data per day the need for detailed mapping and analysis of water- on various characteristics of Earth’s surface, at spatial column properties for climate change studies, of the resolutions that now extend to the submeter level. Pas- related need for ocean fisheries habitat restoration and sive sensors gather radiation from the surface in visible conservation, and for seafloor or subseafloor energy ex- and near-visible wavelengths, while active sensors “ping” traction and tsunami modeling. The geographical sci- the surface using radar, light, and microwave radiation ences have been central to the design of these systems; to create accurate maps of topography, of land cover to the techniques used to interpret imagery to identify and snow, and of the immediate subsurface. Satellite- crop types, land-cover types, benthic habitat types, and based remote sensing also provides needed inputs for biomass production; and to the research that emerges models of the atmosphere that lead to accurate fore- from the detailed analysis of the results ( Jensen, 2005, casts of temperature, rainfall, and extreme events; for 2007; Lillesand et al., 2008). They will continue to do monitoring of sea-surface temperature; and for a wide so as the science of remote sensing advances in the range of other parameters. Similarly, acoustic remote coming years. Finer spatial resolution, more frequent sensing, which dates from the naval vessels of World overpasses, finer spectral resolution to allow more accu-

OCR for page 97
0 UNDERSTANDING THE CHANGING PLANET rate identification of cover and habitat types, more reli- that can yield useful longitudinal data. All too often able tools for classification, faster processing, and more geographical datasets are no more than snapshots, effective systems for distributing data will all work to obtained at a few times over a short interval. Chang- enhance the benefits of remote sensing to science, to ing technology, lack of interest, and the short-sighted the development of public policy, and to the work of a nature of many programs present a major obstacle to myriad of agencies and communities. any concerted effort to build lengthy time series that Recently, there has been much interest in ground- can support analysis of change. We run a real danger based remote sensing, using networks of fixed, autono- that, in the not too distant future, much of what we now mous sensors capable of gathering useful information know about the planet through our current programs on properties of the environment. Sensor networks now of remote sensing will be lost because of a lack of both provide a steady flow of information on worldwide sea the resources and the organizations needed for long- levels, allowing scientists to test predictions of sea- term preservation. Research is needed to identify robust level rise; they provide a cheap and dense network of approaches to preservation, to ensure easy and reliable ground-based weather measurements; and they provide access by the researchers of the future. a steady flow of useful information on traffic densities on roads. What techniques will be needed to visualize earth However, there are obvious limits to what can futures? be observed, identified, and measured from remote In an address written for the opening of the sensors, whether in space, on the ground, or in the California Science Center in Los Angeles in early 1998, ocean. People use place names, distinctive landmarks, V ice President Al Gore expanded on an idea he had and addresses to describe locations, but maps of these first proposed in 1993 (Gore, 1993) for an immersive are created by humans on the ground. Data on demo- environment he termed Digital Earth, which would let graphics, socioeconomic status, and all of the wealth its users explore a visual representation of the planet. of information created by censuses, surveys, and other The idea of a digital replica of Earth, or mirror world, mechanisms for gathering social data are available only also appears in Neal Stephenson’s novel Snowcrash, from people, and only by expensive processes of official published in 1992. data gathering. It is important that researchers find Although the technology of 1998, with its limited ways of sustaining these processes despite their cost Internet connectivity and lack of advanced three- and what appears to be an increasing unwillingness of dimensional graphics, made this concept seem almost some members of the public to participate. impossibly futuristic, by 2005 the average personal An active and promising line of research by geo- computer had sufficient power in visual display, and graphical scientists focuses on the use of unconventional a sufficient number of households were connected by sources to elicit or infer detailed social information. For broadband communications, to make Gore’s basic vision example, Longley developed powerful techniques for de- a reality for millions of people. In that year Google riving spatially detailed social data using personal names acquired, rebranded, and substantially enhanced a (Longley et al., 2007; Gibin et al., 2008), while the field 4-year-old software product called Earthviewer, and of geodemographics (Harris et al., 2005) mines spatially launched the result as Google Earth, in many ways an disaggregated social statistics to sort neighborhoods into implementation of what Gore had described 7 years distinct social types. Recently, a range of projects has earlier. Several other virtual globes have followed, and focused on eliciting and compiling detailed geographical more than 300 million copies of the Google Earth data from the general public, a topic discussed in greater client software have been downloaded. Most users detail in Chapter 11. In all of these cases research is are content to zoom from global to local scales to see needed to understand the quality and usefulness of the their own houses and neighborhoods, but the publica- results and to protect the interests of the providers in tion of the Google Earth application programming preserving confidentiality. interface has allowed users to create literally hundreds Underlying all of these approaches to data acquisi- of thousands of applications. Google Earth has been tion is the problem of sustaining long-term programs

OCR for page 97
0 OBSERVE, ANALYZE, AND VISUALIZE described as “the democratization of GIS [geographic Today, detailed forecasts of global climate change information systems]” (Butler, 2006) because it has al- are readily available in the scientific community, but lowed millions of users to become familiar with some the high level of technical and domain expertise needed of GIS’s very basic operations. to access them limits their value. Policy makers and However, although virtual globes allow us to see individual citizens are far more likely to respond to a rough approximation of how the world looks (at the such forecasts when they can see their implications time the base imagery was acquired), Gore goes on to locally. The idea of bringing the global message home, describe how a Digital Earth could be used to explore of localizing the global, was very much part of Gore’s both the past and the future of the planet. Our abil- original vision for Digital Earth—and far beyond what ity to simulate the future relies on our understanding we can do today. of the processes, both social and environmental, that shape the planet, and to do so at apparently increasing how can geographical tools be adapted to a world rates. Google Earth now has the ability to display a that needs just-in-time answers? sequence of snapshots representing the development Because of the legacy of mapping relatively static of some phenomena through time, and many mash- phenomena, our methods of analysis are similarly ups have been created to show historical data, but the geared to so-called cross-sectional data, or data of spa- ability to visualize future scenarios by exploiting our tial distributions at one point in time. This limitation understanding of processes remains essentially futur- is exacerbated by the tendency for many social data- istic. Enhancing forward-looking visualizations will gathering exercises, such as the U.S. Census, to take require research in many disciplines (Craglia et al., place at fixed intervals. Remotely sensed images have 2008), including the geographical sciences, on topics also provided timed snapshots, although the effective such as the following: frequency of overpasses has been improving recently • Methods to use the spatial structure of virtual as more satellites are launched, allowing the recovery globes (discrete global grids) as the basis for a range of effort following the Wenchuan earthquake in China in simulation models; May 2008, for example, to make use of images collected • Methods of visualization that go beyond the from dozens of satellite sensors. Another reason for the current emphasis on rendering a realistic approxima- paucity of lengthy time series has been the difficulty tion of Earth’s surface appearance, and include prop- of maintaining the flow of public resources needed to erties that are abstract or nonvisual in nature, such as keep large-scale, expensive government data-gathering personal income, gross domestic product, or rainfall; programs in operation decade after decade. • Standards and mechanisms that allow a user As change accelerates and as sensor networks begin to search not only for data, but also for simulation to provide densely sampled data in both space and time, models, and to implement them in a virtual globe we will need to add rapidly to our collection of spatio- environment; temporal techniques of analysis. At this time, we know • Tools that allow models of a wide range of little about how to analyze and mine the increasing processes, from environmental to social, to be repre- supply of data resulting from the tracking of vehicles, sented using a common and reusable set of software people, and animals (Miller and Han, 2001). We know primitives; little about how to assess the significance of an apparent • Methods for downscaling predictions to the change on Earth’s surface detected by remote sens- local level, so they can be made meaningful in a local ing. We need a comprehensive battery of easy-to-use context; and models to simulate a range of social and environmental • Ways of enhancing understanding of the un- processes, and to investigate the footprints they leave certainty associated with simulations of process, and on Earth’s surface. of how that uncertainty can be displayed on a virtual Even more urgent is the need for methods that globe and communicated to the user. can continuously monitor the stream of data coming from our acquisition systems, searching constantly for

OCR for page 97
0 UNDERSTANDING THE CHANGING PLANET summarY anomalies and novel patterns, and initiating appropri- ate investigations. Such real-time analysis of medical Advances in technology in the contemporary area pres- diagnoses by primary care physicians, for example, ent new opportunities and challenges for the age-old could provide early warning of disease outbreaks and task of observing, mapping, and representing Earth’s health hazards. Data coming from individuals could be changing geographical character. Pursuit of the illus- used to provide early assessments of the damage from trative research questions presented in this chapter will disasters, and could speed the initiation of response. enable the geographical sciences to collect, analyze, and Spatial decision-support systems based on real-time share information in ways that are critical to the mul- streams of data could provide new levels of effectiveness tidisciplinary task of understanding and assessing the in the management of numerous social and environ- human and environmental processes that are shaping mental problems. the future of the planet.