Next-generation Digital Earth

Michael F. Goodchilda,1, Huadong Guob, Alessandro Annonic, Ling Biand, Kees de Biee, Frederick Campbellf, Max Cragliac, Manfred Ehlersg, John van Genderene, Davina Jacksonh, Anthony J. Lewisi, Martino Pesaresic, Gábor Remetey-Fülöppj, Richard Simpsonk, Andrew Skidmoref, Changlin Wangb, and Peter Woodgatel aDepartment of Geography, University of California, Santa Barbara, CA 93106; bCenter for Earth Observation and Digital Earth, Chinese Academy of Sciences, Beijing 100094, China; cJoint Research Centre of the European Commission, 21027 Ispra, Italy; dDepartment of Geography, University at Buffalo, State University of New York, Buffalo, NY 14261; eFaculty of Geo-Information Science and Earth Observation, University of Twente, 7500 AE, Enschede, The Netherlands; fFred Campbell Consulting, Ottawa, ON, Canada K2H 5G8; gInstitute for Geoinformatics and Remote Sensing, University of Osnabrück, 49076 Osnabrück, Germany; hD_City Network, Newtown 2042, Australia; iDepartment of Geography and Anthropology, Louisiana State University, Baton Rouge, LA 70803; jHungarian Association for Geo-Information, H-1122, Budapest, Hungary; kNextspace, Auckland 1542, New Zealand; and lCooperative Research Center for Spatial Information, Carlton South 3053, Australia

Edited by Kenneth Wachter, University of California, Berkeley, CA, and approved May 14, 2012 (received for review March 1, 2012)

A speech of then-Vice President Al Gore in 1998 created a vision for a Digital Earth, and played a role in stimulating the development of a first generation of virtual globes, typified by Google Earth, that achieved many but not all the elements of this vision. The technical achievements of Google Earth, and the functionality of this first generation of virtual globes, are reviewed against the Gore vision. Meanwhile, developments in technology continue, the era of “big data” has arrived, the general public is more and more engaged with technology through citizen science and crowd-sourcing, and advances have been made in our scientific understanding of the Earth system. However, although Google Earth stimulated progress in communicating the results of science, there continue to be substantial barriers in the public’s access to science. All these factors prompt a reexamination of the initial vision of Digital Earth, and a discussion of the major elements that should be part of a next generation.

scientific communication | visualization

igital replicas of complex enti- surface and near-surface of the Earth, and jections entirely by showing the Earth as ties and systems are now rou- could be used to illustrate and even ad- seen from space (technically a perspective D tine in many fields—in the dress a host of problems facing humanity, orthographic projection onto the 2D design and testing stages of from climate change and natural disasters plane of the screen), measured distance by aerospace engineering, in digital re- to warfare, hunger, and poverty (see, for using the length of the shortest path over constructions of ancient cities, or even in example, the 2009 Beijing Declaration the curved surface, and reduced scale to physiology, in which digital cadavers can on Digital Earth; 159.226.224.4/isde6en/ the simple metaphor of raising or lowering replace real ones in the training of medical hykx11.html). The Chinese Academy of the user’s viewpoint. A fly-by, the “magic students. The concept of a Digital Earth, Sciences responded by organizing the first carpet ride” of the Gore speech, the a digital replica of the entire planet, occurs International Symposium on Digital Earth capstone achievement for generations of in Al Gore’s 1992 book Earth in the in Beijing in 1999; the symposium became undergraduate GIS students, could be Balance (1), and was developed further a biennial event, and has been held in created by a child of age 10 in 10 minutes in a speech written for delivery by Gore Canada, the Czech Republic, Japan, the by using nothing more than a home at the opening of the California Sci- United States, China, and Australia. In computer and a downloaded Google ence Center in January 1998 (portal. 2006, the International Society for Digital Earth client. opengeospatial.org/files/?artifact_id= Earth (ISDE) was established, and bi- There will always be a need for flattening 6210). By the turn of the century, support ennial Digital Earth Summits have been the Earth, to see the entire surface at once, for advanced 3D graphics had become convened in New Zealand, Germany, and albeit distorted, or to present information ’ fi standard on personal computers, allowing Bulgaria. In 2008, ISDE sof cial publica- on paper. However, there is no doubt that real-time manipulation of complex ob- tion, the International Journal of Digital virtual globes had enormous advantages jects. Following the speech, the Clinton Earth, was inaugurated. over traditional maps as a means of com- Administration directed NASA to form In this concept of Digital Earth, sharing municating data, information, and knowl- fi fi a Digital Earth of ce, a number of pro- of scienti c information about the planet edge about the surface and near-surface of fi totype projects were begun, and, with the could extend well beyond the scienti c the planet. Here for the first time were 2005 launch of the Google Earth service, community, to people with very limited software environments that could overlay fi — a ne-resolution, digital replica of the technical skills and computing resources layers of information by using geographic planet, or at least of its surface, was within the vision of an information food chain location as a common key, show buildings reach of anyone with a broadband Internet extending all the way from science to and vegetation as complex 3D structures, connection. public policy seemed almost within reach. fi ’ pan and zoom at the click of a mouse, and The effect on the scienti c community Google s own Earth Engine project ex- visualize extracts from petabytes of rapidly was immediate and palpable (2). Here was emplifies this vision by making available a readily accessible technology that could Gore’s “vast quantities of geo-referenced be used to present scientific data and re- information.” Unlike geographic infor- fi Author contributions: M.F.G., H.G., A.A., L.B., K.d.B., F.C., sults (scienti c applications of Google mation systems (GISs), which have a rep- M.C., M.E., J.v.G., D.J., A.J.L., M.P., G.R.-F., R.S., A.S., C.W., Earth were recently reviewed in ref. 3) in utation for being difficult to learn, and and P.W. wrote the paper. easily digestible, visual form to collabo- force users to confront the intuitively dif- The authors declare no conflict of interest. rators and a general public that perceived ficult spatial concepts of scale and map This article is a PNAS Direct Submission. it as free, fast, and fun. Digital Earth had projections, Digital Earth implementa- 1To whom correspondence should be addressed. E-mail: relevance to any science dealing with the tions such as Google Earth avoided pro- [email protected].

11088–11094 | PNAS | July 10, 2012 | vol. 109 | no. 28 www.pnas.org/cgi/doi/10.1073/pnas.1202383109 Downloaded by guest on September 28, 2021 PERSPECTIVE

accumulating georeferenced information ence. Open-source packages such as dragged to contain the pole itself, or to in fractions of a second. NASA’s World Wind can be extended intersect with 180° longitude. The mathe- It is now 7 years since the launch of with new code, and Google’s Application matics of the spheroid are difficult, and Google Earth, and a similar period has Programming Interface allows Google perhaps shortcuts are being taken, for ex- elapsed since the release of the earliest of Earth functions to be embedded within the ample, by resorting to calculations on the what is now a long list of comparable virtual user’s own application. sphere or on a planar projection when the globes, including NASA’s open-source Virtual globes that are able to capture spheroid becomes less tractable. WorldWind (worldwind.arc.nasa.gov), and display elevation above or below the Problems of dealing with uncertainty Wuhan University’s GeoGlobe, the Chi- surface have obvious applications in may have broader implications. Positions nese Academy of Sciences Digital Earth oceanography, atmospheric science, and on the Earth are the result of measure- Prototype System, Microsoft’s Bing Maps geomorphology. Since its initial launch, ment, and thus have accuracies that are (www.bing.com/maps), Esri’s ArcGIS Google has also enhanced the temporal limited by the measuring device. It is Explorer (www.esri.com/software/arcgis/ dimensions of Google Earth, allowing the a long-established principle of science that explorer/), Unidata’s Integrated Data display of time series and the use of historic measurements should be reported with Viewer (www.unidata.ucar.edu/software/ base maps. Unidata’s Thematic Realtime a numerical precision that matches their idv/), and Digitnext’s VirtualGeo (virtual- Environmental Distributed Data Services accuracy, and the principle provides easy geo.diginext.fr/EN/). It is also more than project (www.unidata.ucar.edu/projects/ access to estimates of inherent uncertainty. a decade since the Gore speech and the THREDDS/) aims to facilitate sharing of However, latitudes and longitudes are vision that did much to motivate these data among Earth scientists using the routinely reported by Google Earth to as efforts, which, in some cases, have ex- Integrated Data Viewer platform, and many as six decimal places of degrees tended well beyond the Gore vision. This includes a full range of 3D and tempo- (corresponding, in the case of latitude, or seems an appropriate time, therefore, to ral support. longitude at the equator, to ∼10 cm) ir- examine what this first generation has Nevertheless, this first generation of respective of the spatial resolution of the achieved, and to envision what might be virtual globes has several limitations, each display. Distances, for example, between achievable and desirable in the future, with implications for potential uses. For Los Angeles and New York, can be re- another 6 or 7 years from now (an earlier a scientist, issues of accuracy, replicability, ported to hundredths of centimeters, an and now somewhat dated analysis is pro- and documentation are likely to be of absurd precision given the lack of an ac- vided in ref. 4). At the same time, these special concern, although they may be of cepted definition of the distance between virtual globes, especially Google Earth, less concern for the general public. Some of two extended objects. Clearly the design- are examined from a critical scientific the more technical limitations are dis- ers of this software lost track of an im- perspective to draw out some of the issues cussed in this section, and the following portant scientific principle, preferring, that lie at the intersection of commercial section asks what new developments might instead, to exploit the full numerical pre- software and the scientific enterprise. characterize a second, more powerful cision of the computing system. generation of Digital Earth. Readily accessible virtual globes are at- Achievements of Virtual Globes to tractive bases for rapid determination of Date Accuracy, Replicability, and Documentation. latitude and longitude. It is easy, for ex- A common reaction to the Gore speech in The Earth does not conform to any sim- ample, to find the location of a feature such 1998 centered on data volume: with ap- ple mathematical shape. The geoid or as a road intersection by identifying it on proximately 5 × 1014 sq m of surface, as isopotential surface, best understood as the the virtual globe’s already georegistered little as a single byte allocated to each surface formed by sea level and its imagi- image and capturing its coordinates. Any square meter results in half a petabyte of nary extension under the continents, must misregistration of imagery (by tying an data before compression. Zooming down be approximated by a mathematical shape image to the Earth’s surface inaccurately) to 1 m resolution and panning across the to define latitude and longitude and thus to is inherited by any locations captured by surface requires some clever algorithms measure location. Google Earth uses the using it, and when new imagery is inserted and data structures if it is to be achievable specific oblate spheroid known as WGS84, in the virtual globe, any previously regis- with a standard personal computer at the an international standard World Geodetic tered feature will now be offset by any end of a typical broadband connection, System, and, in principle, measures made positional difference in the new registra- with limited capacity for local caching on the surface of this implementation of tion. It is important to recognize that this of data. Virtual globes use a variety of Digital Earth should conform to this is a measurement problem, and that exact hierarchical tiling structures known as standard. Lines of latitude should grow measurement of location is impossible. discrete global grids (and reminiscent of further apart away from the equator, Each new registration of the base imagery, Buckminster Fuller’s geodesic domes) to whereas lines of longitude will, of course, in effect, creates a new datum—a new lo- enable rapid zoom (reviewed in ref. 5), converge on the poles. cal approximation to WGS84 that may or precompute tiles on the server to avoid Google Earth allows the user to add may not be better than previous approx- extensive local computation, and use several types of figures to the Earth’s sur- imations. A scientist expects extensive sophisticated level-of-detail management face, and in some cases to drag them freely documentation (i.e., metadata) on the to allow the field of view to be refreshed at over the surface. Dragging a circle, for registration process, its temporal history, video rates (6). example, produces several expected be- and its implications. One of the most attractive features of the haviors [these experiments were con- The actual amount of misregistration of virtual globes for the scientific user has ducted with Google Earth Pro-6.0.1.2032 base imagery varies over space and over been their capacity for extension and ad- (beta); build date, December 10, 2010]: its time, depending on the availability of aptation to individual needs. Google’s radius remains constant, but its circum- small, recognizable features with known Keyhole Markup Language allows the ference grows longer toward the poles as locations that can be used as reference user’s 2D and 3D data to be readily su- the local curvature becomes less. How- points, on the spatial resolution of the perimposed on the virtual globe in a vari- ever, some behaviors are unexpected, in- imagery itself, and on many other factors. ety of formats, opening the door to the cluding a stepwise change of circle area as Potere (7) found that the mean mis- vast number of “mashups” that have been it is dragged toward the poles. Behavior registration of Google Earth imagery was constructed recently in every area of sci- becomes almost chaotic when the circle is ∼40 m relative to Landsat GeoCover.

Goodchild et al. PNAS | July 10, 2012 | vol. 109 | no. 28 | 11089 Downloaded by guest on September 28, 2021 Over Santa Barbara, CA, misregistration nificant burden on the user to execute rations, and unavailable to others, or has been as much as 40 m at times relative a successful process of search, discovery, heavily restricted in its reuse. However, the to high-quality global positioning system extraction, and use that is very different case for open data access is becoming in- (GPS) measurements. What is important, from the user-friendly, visual paradigm of creasingly compelling, given the severity of course, is not whether misregistration the virtual globes. and urgency with which problems must be exists, as it must, but whether it is sub- Third, Google Earth’s emphasis on vi- addressed at a global scale. The thrust stantial enough to affect a given applica- sualization makes it difficult to communi- toward open data is gaining momentum in tion. Moreover, positional uncertainty cate information that is not inherently several countries that embrace it for rea- is only one of the many dimensions visual. A virtual globe that looks like the sons of transparency, administrative effi- of uncertainty that characterize geo- Earth itself is intuitively straightforward, ciency, and the economic potential of graphic data. easy to navigate, and easy to understand. reuse. They support open government Topography can be shown in three di- through legislation and practical measures, Achievements to Date Against the Gore mensions, rather than by using the con- such as the production of data in machine- Vision. Despite the progress they repre- tours and other coded techniques adopted readable formats and the creation of sent, today’s generation of virtual globes by cartographers; the effects of sea-level data portals. fall short of Gore’s vision of a Digital rise can be visualized by flooding areas The ubiquity of GPS has allowed anyone Earth in several key respects. [Goodchild with solid blue. However, other properties to measure location on the Earth’s surface (8) provides a more complete analysis of important to science, such as temperature with an accuracy of a few meters, and to the uses of Digital Earth mentioned in the fields, the spatial variation of soil types or tag other information, such as photo- Gore speech, and compares them to those vegetation properties, biodiversity, or the graphs, with geographic locations. Crowd- of GISs.] First, whereas Gore imagined distributions of social, economic, and de- sourcing is now a major source of geo- a Digital Earth that would allow the user mographic variables, are not inherently graphic information (12), exemplified by to explore backward in time by using his- visual, and not easily conveyed in imme- the enormously successful, worldwide torical data and forward in time by using diately meaningful ways. Uncertainty is OpenStreetMap project and Google’s computational models to simulate future even more problematic, as there are no MapMaker. New forms of citizen science scenarios, virtual globes remain largely obviously intuitive ways of conveying its are encouraging the general public to be- centered on portraying the Earth as it magnitude or implications—we are used come involved in the acquisition of data looks at the current time. Older imagery is to seeing a single, solid world rather than on phenology, weather, disasters, and available in Google Earth for many areas, a world of many conflicting possibilities. many other Earth-related phenomena. and historical maps have been made ac- Many of these trends come together in the cessible as mashups using Keyhole Prospects for the Next Generation term neogeography (13), which implies Markup Language. Simulation-model Several fundamental changes in society and a breaking down of the traditional dis- outputs and time-series of imagery have in the world of digital geographic in- tinction between expert and amateur in also been linked to virtual globes. How- formation have occurred since Gore’s the world of mapping and geographic in- ever, to date, it is not normally possible for speech was formulated in late 1997. formation. Thanks to GPS and a range of a member of the general public to recover Mention has already been made of the Web-based services, the average citizen is what his or her neighborhood looked like advances in bandwidth and 3D graphics able to be a consumer and a producer (a at some arbitrarily defined point in the that have enabled the standard consumer “prosumer”) of geographic information, past, or to use a virtual globe to visualize device to visualize and manipulate a Digi- and to make maps, such as those gener- what his or her world will be like, in terms tal Earth (Fig. 1). The supply of geo- ated by in-vehicle navigation systems, that of urban growth, climate change, or sea- graphic information from satellite-based are valid only at an instant in time, cen- level rise, at some point in the future and ground-based sensors has expanded tered on the user, representing the world based on the best scientific modeling. rapidly, encouraging belief in a new, as seen from above or from the ground, Second, whereas Gore imagined an en- fourth, or “big data,” paradigm of science and useful only for the purposes of the vironment in which it would be possible to (11) that emphasizes international collab- moment. Many of these services have store, retrieve, and share “vast quantities” oration, data-intensive analysis, huge proven to be effective even through the of Earth-related information, in reality, computing resources, and high-end visu- interface of a simple mobile phone (e.g., not all such data are equal in the world of alization. It also demands improved tech- Ushahidi, www.ushahidi.com). The citizen the virtual globes. Although the user can niques of search and discovery such as as prosumer offers an entirely new vision overlay data of various kinds on the base those that can be readily implemented in of the role of the citizen in relation to imagery, he or she has no control over the a virtual globe. science: as volunteer observer, as in- process by which Google acquires and A large part of this new information is telligent recipient of the results of science makes available such imagery, and the the property of governments and corpo- as applied to the citizen’s own Terms of Use impose tight restrictions over what can be done with the results. Rather than exploit virtual globes, the strategies developed for sharing of geo- graphic data among scientists tend instead to follow the model of a data library [or geolibrary (9)]: a searchable catalog of holdings, formally structured metadata, and mechanisms for facilitating the downloading of selected data sets (see, e.g., the Geospatial One-Stop, geo.data.gov; or the Global Change Master Directory, gcmd.nasa.gov). Geoportals (10) provide a single point of entry into a distributed set Fig. 1. Handheld consumer devices such as this tablet computer can augment a real scene with data of holdings, but nevertheless impose a sig- obtained from Digital Earth, such as these underground pipes. (Image courtesy of NextSpace.)

11090 | www.pnas.org/cgi/doi/10.1073/pnas.1202383109 Goodchild et al. Downloaded by guest on September 28, 2021 surroundings, and as an informed stake- made at a location can readily be linked to ample, or food supplies. Such predictions holder in the Earth’s future. other known facts about that location, or are commonly the outputs of simulation Crowd-sourcing produces information of about nearby locations, in a rich con- models, and represented in the extreme by variable quality that lacks the structured textualization. We can distinguish between projects that aim to develop a Digital sampling and rigorous methods of mea- horizontal context, or the context estab- Earth capable of predicting a vast array of surement that scientists emphasize. New lished by knowledge about nearby loca- properties at fine resolution in space and methods will be needed to synthesize such tions, and vertical context, or the context time, such as Europe’s FuturICT (www. crowd-sourced information, and to assure established by other things that are known futurict.eu) or Japan’s Earth Simulator its quality, if it is to be used with confidence about the same location. Both are cap- (www.jamstec.go.jp/esc/index.en.html). by scientists and more broadly in society. In tured in the famous principle of Tobler, They might be generated in one of two future, techniques of synthesis applied to often termed the First Law of Geography: distinct ways: by running a model in its disparate data of varying quality may be as “All things are related, but nearby things own software environment and then visu- important as techniques of analysis have are more related than distant things” (14). alizing the results by overlaying them as been in the past. A virtual globe provides easy access to a topping on a Digital Earth, or by running These new technologies of neo- context, especially horizontal context, al- a model within the Digital Earth software geography offer an unprecedented oppor- lowing inferences to be made from the environment itself, perhaps by using the tunity for science to extend its findings conditions surrounding a location. Verti- virtual globe’s discrete global grid as a set from the laboratory not only to the pages of cal context is more difficult to visualize, of finite elements. The latter might allow refereed journals, but also to the eyes and because a virtual globe allows only one the virtual-globe user to modify parame- minds of the general public. As easy-to-use layer (or what one might term on a virtual ters or boundary conditions, to evaluate tools capable of visualizing vast amounts globe a “topping”) to be transparently alternative scenarios. In essence, such ap- of information at scales from the global to overlaid on the imagery base. Mapping proaches would allow the user to in- the local, the virtual globes clearly have techniques such as cross-hatching might vestigate how the Earth works and how it a major role to play in this new vision of allow for two layers, but, more generally, it might evolve in the future, rather than to scientific communication. Several aspects would be useful to look for more powerful observe how it currently looks. In the spirit of a renewed Digital Earth vision derive ways of making the user aware of the of neogeography, these approaches might from this perspective. The ability to link the vertical context of geographic attributes. allow the user to augment reality with global to the local, through rapid zoom, is Moreover, insights can often be gained not visualizations from the user’s own per- clearly a key aspect of the virtual globes. from the horizontal and vertical context of spective. They would require substantial At the same time, it would be wise to attributes, but by searching for analogous investment in model management, given recognize that a simple linear model of conditions elsewhere on Earth; virtual the vast number of models that are al- communication between scientists and the globes seem especially well suited to this ready available, their lack of intero- public is naive in several respects. Scientists kind of investigation. perability, and the difficulty of choosing have much to learn about communication, More important, perhaps, would be the between them. especially in dealing with the uncertainty development of ways of capturing the It would be important to incorporate inherent in all measurements and pre- linkages that exist between locations and a few simple analytic functions in the dictions, as the recent “Climategate” and across scales. Maps are powerful ways of interests of supporting the evaluation of its aftermath demonstrated. The general visualizing the properties of locations, such scenarios. A subset of the functions gen- public can often understand concepts of as elevation or soil type, or what one might erally available in GIS would allow the user uncertainty and probability if they are ex- term “unary” geographic information. to obtain summary statistics for areas, plained well, and made an inherent part of However, they are much less powerful at for example, or to compute correlations. systems like Digital Earth. Research over portraying the linkages that exist between However, wholesale adoption of the the past two decades has given us a rich places, such as flows of migrants or the functionality of a GIS is likely to lead to the understanding of the modeling and visu- configurations of social networks—in- same kinds of inaccessibilities the average alization of uncertainty in geographic formation that one might term binary be- citizen already encounters with GIS soft- information, and of how to communicate cause it involves the properties of ware. Careful consideration of cognitive uncertainty to the user. It is vital that the locations taken two at a time. A map of issues will be needed to find an appropri- next generation of Digital Earth replace migration flows between each pair of the ate compromise between the scientist’s the exaggerated precision of the first gen- more than 3,100 US counties, for example, need for advanced functionality and the eration with techniques that convey un- would show a mass of lines that would general public’s need for simplicity and certainty clearly and unambiguously to the be impossible to comprehend without ease of use. user. More broadly, the social sciences sophisticated techniques of abstraction Although it is always risky to offer any have much to offer on the subject of and generalization (e.g., ref. 15), whereas kind of prediction of technological futures, communication, the cultural differences a map of county life expectancy or eth- several trends are already apparent that that characterize the various environments nicity is easily understood. Linkages are should be incorporated in a renewed vision in which Digital Earth is accessed, and the clearly important to our understanding of of Digital Earth. We already have the ways in which data are understood and environmental and social processes, and means to keep track of many kinds of interpreted. Vast differences exist around access to information about them should objects: many vehicles, and the mobile the world in access to computer hardware, be an important part of a renewed vision. phones of their occupants, are constantly software, and the Internet, as well as in the This is especially true when linkages tracked by using GPS and managed as technical skills and attitudes of users. The represent the impacts of changing con- “probes” to collect data on traffic speeds next generation of Digital Earth will need ditions in one location on conditions in and congestion; commercial shipments are a strong commitment to involving the another location, such as downstream or often tracked, as are pets, parolees, and social sciences in its development. downwind. many types of animals, in the interests of One of the most powerful benefits of The Gore speech describes a Digital research. In the future, we should envision georeferencing—in other words, the asso- Earth that would be capable of presenting a world in which it will be possible to know ciation of attributes with locations on the predictions about the Earth’s future—of the locations of everything at all times, Earth’s surface—is context. Observations climate change and sea-level rise, for ex- through a georeferenced “Internet of

Goodchild et al. PNAS | July 10, 2012 | vol. 109 | no. 28 | 11091 Downloaded by guest on September 28, 2021 Things.” We are also seeing a massive storing the sounds and stories associated the Earth’s surface by connecting to sensor expansion in the functionality of mobile with locations, and by responding to spo- networks and situation-aware systems (17). devices, and can envision a world in which ken place names. Audio is already strongly Although great progress has been made computing will be ubiquitous, and in which associated with small portable or wearable in overcoming differences of software and Digital Earth will be accessible anywhere, devices, which may in the future become data formats in sharing and communicating at any time. the primary modality for access to geographic information, issues of seman- Although GPS positioning generally Digital Earth. tics continue to present major challenges. works more effectively outdoors, and Differences of mapping practice, the ways geographic information is mostly 2D de- The Way Forward in which land is classified, and the mean- spite the increasing prevalence of 3D It is one thing to describe a vision for the ings attached to data are being addressed building representations in virtual globes, next generation of Digital Earth, but it by projects such as the European Com- we should anticipate a day when it will be may be quite another to imagine how it munity’s Infrastructure for Spatial Infor- possible to support georeferencing and might be achieved. The top-down option mation in the European Community navigation within detailed 3D structures for implementing Gore’s Digital Earth (inspire.jrc.ec.europa.eu). However, the anywhere on the planet to centimeter ac- through large-scale public investment Earth’s surface is clearly heterogeneous, curacies (see, e.g., the work of the D_City stuttered to a halt shortly after the 2000 and any community will inevitably devise Network, dcitynetwork.net and Fig. 2). US Presidential election, and in the cur- its own ways of describing its own sur- Despite these advances, the problem of rent funding climate, a top-down initiative roundings. Although remote sensing ach- archiving remains: we have, as yet, no for a next generation seems equally ieves a degree of uniformity, other sources satisfactory solution to the long-term unlikely. of geographic information tend to reflect preservation of the vast amounts of digital In reality, the first generation resulted the standards and practices of their source. data already created, along with the ex- from advances in Earth observation; from One has only to compare, for example, the ponential growth in data that seems likely a timely evolution of bandwidth and 3D maps of the political boundaries of the to continue for some time to come. Much visualization, the latter driven in part by the Himalayan region provided by Google of what we know of the history of the planet video game industry; from the modest Maps in the US with those provided by is recorded on paper, and some of it has stimulus provided by government funding; Google Maps in China and India (Fig. 3) already survived for centuries. However, it from the impetus provided by a US Vice to realize the impossibility of achieving is very hard at this point to see how century- President; by the commercial possibilities uniformity when mapping the human long preservation of digital media can be sensed by large corporations such as condition. achieved, together with the means to Google and Microsoft; and by the interests Any effort to develop a next-generation search and retrieve useful information. and enthusiasm of the open-source com- Digital Earth will require new governance Extensive research and development will munity. Underlying this was a sense among models. In addition to the ISDE (www. be needed if the records now being as- a loose-knit community that Digital Earth digitalearth-isde.org), many other organ- sembled for Digital Earth are to survive for was both possible and desirable. We believe izations, including the Global Spatial Data more than a few years. similar factors can drive a next generation. Infrastructure Association (www.gsdi.org), Finally, today’s world of Digital Earth is The next generation of Digital Earth will the International Council for Science and predominantly silent, using only the visual not be a single system but, rather, multiple its Committee on Data for Science and channel to communicate with the user. connected infrastructures based on open Technology (www.codata.org), the Group Sound is at least as rich a medium for access and participation across multiple on Earth Observations (and its geo- communication as vision, yet it plays al- technological platforms that will address portal, the GEO System of Systems, www. most no part in the communication of the needs of different audiences (16). A earthobservations.org), the UN Commit- geographic information. The next genera- more dynamic view has also been pro- tee of Experts on Global Geospatial In- tion of Digital Earth might make use of posed of Digital Earth as a digital nervous formation Management (ggim.un.org), audio in several ways: by allowing the user system of the globe, actively informing and many national agencies will address to make requests through speech, by about events happening on (or close to) various aspects of the future Digital Earth. These organizations can play a helpful role in endorsing the concept of a next-gener- ation Digital Earth, and helping to elabo- rate its vision. Along with the scientific community, they can also play a useful role in ensuring that the next generation meets the highest standards of scientific rigor, especially careful and detailed documen- tation of uncertainty. Perhaps there is also room for a Digital Earth code of ethics that could set standards for behavior in a complex, collaborative enterprise. However, in the realities of today’s economies, it seems most likely that in- novation will come from the private sector, together with volunteer initiatives cen- tered around open-source tools. The pri- vate sector is fundamentally competitive, but organizations such as the Open Geo- spatial Consortium (www.opengeospatial. org) have been able to achieve remark- Fig. 2. A rendering of Auckland, New Zealand, using 3D models of all of its buildings and other built able degrees of cooperation and in- structures. (Image courtesy of NextSpace.) teroperability. A similar collaboration

11092 | www.pnas.org/cgi/doi/10.1073/pnas.1202383109 Goodchild et al. Downloaded by guest on September 28, 2021 among the private sector, academia, non- governmental organizations, and govern- ment, focused on Digital Earth, might be similarly successful. Collaboration at this scale will also be necessary to tackle the growing issues of privacy and ethics that are associated with access to fine-resolution geographic in- formation (18). Technologies such as GPS and radio-frequency identification (RFID) raise the possibility of a future in which it will be possible to know and record where anything is, at all times. Fine-resolution imaging from space can identify objects on the surface less than 1 m across, and im- aging from the ground can be used to recognize faces and license plates, creating a world that Dobson and Fisher charac- terize as “geoslavery” (19). To date, efforts to protect privacy have been patchwork in nature: for example, some countries now require Google Street View imagery to blur faces and license plates, and mobile- phone service providers in the US are restricted in their use of locational in- formation. Against this one must place the evident willingness of people to publicize their location through services such as FourSquare or Google Latitude. To date, no principle of privacy protection has emerged that finds an acceptable, consis- tent position between the extremes of universal access and universal protection. At minimum, perhaps any individual should have a guarantee of control over his or her own locational privacy, able to turn it on or off at will. It is clear to us that the first generation is powerful but limited, that technology is advancing rapidly and making possible what might have been inconceivable a few years ago, and that humanity needs better ways of linking the scientific community and its discoveries about the ways the Earth works and is structured with the general public and their growing concern about the planet’s future. We would even go so far as to argue that access to scientifically grounded information about the planet’s future is a basic human right; that it should be available to all, independent of national policies and strategies, and in a form that is readily understood and absorbed. This is, after all, the only planet we have, and we all have a vested interest in its future. Digital Earth seems to us the most acces- sible and compelling way to organize, visualize, and use that investment.

ACKNOWLEDGMENTS. This paper is largely the outcome of a workshop on next-generation Digital Earth held in Beijing in March 2011 and hosted by the Center for Earth Observation and Digital Earth of the Chinese Academy of Sciences and the Sec- retariat of the International Society for Digital Earth (ISDE). The authors thank Luke Driskell for his Fig. 3. Three versions of the boundaries of the Himalayas, as provided by Google Maps in response to extensive notes of the discussion. The paper also queries from (A) the US (Copyright Google, Kingway, MapKing, Mapabc, SK M&C, TeleAtlas, ZENRIN), (B) benefited from comments on earlier drafts from China [Copyright GS(2011)6020 Google, Kingway, MapKing, Mapabc, TeleAtlas], and (C) India (Copyright members of the ISDE Executive Committee: Mario Google, LeadDog Consulting, Mapabc, TeleAtlas). Hernandez and David Rhind.

Goodchild et al. PNAS | July 10, 2012 | vol. 109 | no. 28 | 11093 Downloaded by guest on September 28, 2021 1. Gore A (1992) Earth in the Balance: Ecology and the 7. Potere D (2008) Horizontal positional accuracy of Goo- 14. Sui DZ (2004) Tobler’s First Law of Geography: A big Human Spirit (Houghton Mifflin, Boston). gle Earth’s high-resolution imagery archive. Sensors 8: idea for a small planet? Ann Assoc Am Geogr 94: 2. Butler D (2006) Virtual globes: The Web-wide world. 7973–7981. 269–277. – Nature 439:776 778. 8. Goodchild MF (2008) The use cases of digital earth. Int J 15. Tobler WR (1975) WINDS: A Computer Program for the 3. Yu L, Gong P (2012) Google Earth as a virtual globe Digital Earth 1:17–30. Analysis of Geographical Interactions. Cartographic tool for Earth science applications at the global scale: 9. National Research Council (1999) Distributed Geolibra- Laboratory Report No. 13 (Univ Michigan, Ann Arbor). Progress and perspectives. Int J Remote Sens 33: ries: Spatial Information Resources: Summary of a 16. Craglia M, et al. (2012) Digital Earth 2020: towards 3966–3986. Workshop (National Academy Press, Washington, DC). a vision for the next decade. Int J Digital Earth 5:4–21. 4. Craglia M, et al. (2008) Next-Generation Digital Earth. 10. Maguire DJ, Longley PA (2005) The emergence of geo- A position paper from the Vespucci Initiative for the 17. De Longueville B, Annoni A, Schade S, Ostlander N, portals and their role in spatial data infrastructures. ’ Advancement of Geographic Information Science. Int J Whitmore C (2010) Digital Earth s nervous system for – Spatial Data Infrastructures Res 3:146–167. Comput Environ Urban Syst 29:3 14. crisis events: real-time sensor Web enablement of vol- 5. Sahr K, White D, Kimerling AJ (2003) Geodesic discrete 11. Hey AJG, Tansley S, Tolle KM (2009) The Fourth Para- unteered geographic information. International Jour- digm: Data-Intensive Scientific Discovery (Microsoft Re- global grid systems. Cartography and Geographic In- nal of Digital Earth 3:242–259. formation Science 30:121–134. search, Redmond, WA). 18. Curry MR (1998) Digital Places: Living with Geographic 12. Goodchild MF (2007) Citizens as sensors: The world of 6. Bar-Zeev A (2007) How Google Earth [really] works, Information Technologies (Routledge, New York). – RealityPrime Advanced Technology Research, July 3, volunteered geography. GeoJournal 69:211 221. 19. Dobson JE, Fisher PF (2003) Geoslavery. IEEE Technol 2007. Available at www.realityprime.com/articles/how- 13. Turner A (2006) Introduction to Neogeography Soc Mag 47:47–52. google-earth-really-works. Accessed July 16, 2011. (O’Reilly, Sebastopol, CA).

11094 | www.pnas.org/cgi/doi/10.1073/pnas.1202383109 Goodchild et al. Downloaded by guest on September 28, 2021