We propose a three-year project that will determine the ontology of geographic phenomena, both forms and processes, in a scientific context. To this end, we will select several domains of geography and the environmental and social sciences, domains that include a significant proportion of the geographic concepts used in science, and establish, within a single over-arching framework, the basic ontology of each domain. We propose to conduct this as a single integrated project because of the clear complementarities among the domains, the need for a comprehensive product, and the methods that will be shared among the six domains. Such a project will contribute both to geography and to geographic information science in fundamental ways The ontology of the separate domains will be a major step toward a complete ontology of geographic phenomena, and will form a sound basis for interoperable software to support research using GIS. We will develop the ontology at both the philosophical and the information systems levels of specificity. We will publish the results of the project in the form of reports, web documents, journal articles, and a comprehensive book.
After discussing results of prior NSF support, we will discuss the role of ontology in information systems. We will outline the methods that will be used to capture and formalize the ontology, and conclude by describing the domains that will be studied and presenting a management plan and a plan of work.
Each of the investigators has had several NSF awards active within the past five years. Of these, the original NCGIA award and the NCGIA's Varenius project ("Advancing Geographic Information Science") are the prior awards most closely related to the current proposal.
Collaborative Agreement 8810917 ($9,800,148, 11/1/88 to 12/31/97) between
NSF and the University of California, Santa Barbara (with subcontracts to the
University at Buffalo and the University of Maine) provided funding for the
National Center for Geographic Information and Analysis through the end of
1996, with a no-cost extension through 1997. Its objectives were to conduct
research in geographic analysis utilizing geographic information systems (GIS);
to promote the use of GIS throughout the sciences; to increase the nation's
supply of experts in GIS; and to provide a national focus of research, with
links to related efforts in other countries. Goodchild was PI in the later years
ofon this award. Under this award NCGIA
researchers addressed 18 topics, known as Research Initiatives; published 54
books, 646 refereed journal articles, and 734 other articles, and gave 1,006
research presentations; developed extensive materials in support of instruction
at all educational levels, including the 1990 NCGIA Core Curriculum in GIS;
developed and distributed software and data sets; and organized successful
international conferences in rapidly advancing areas of research.
The NCGIA’s Varenius project to advance geographic
information science was initiated in 1997 under Cooperative Agreement 9600465
($2,462,621, 2/15/97 to 1/31/00, extended to 1/31/01), with subcontracts to
the University at Buffalo, the University of Maine, and the University of
Minnesota, and with Goodchild as PI. The Varenius effort was organized into
three Strategic Research Areas, each with a panel of internationally known
experts: cognitive models of geographic space; computational implementations of
geographic concepts; and geographies of the information society. An Advisory
Board oversaw the entire project. During the period of the award, a total of
nine Specialist Meetings were held, three on topics of the highest priority in
each of the three Strategic Research Areas. Each meeting included approximately
30 scholars, drawn from all of the disciplines relevant to the topic by an
international process of open selection and invitation. Following each meeting,
pPrograms
of seed grants and visiting fellowships promoted collaboration, and the impacts
of the each meeting were identified through a
questionnaire survey of participants. Each Varenius meeting
resulted in a report summarizing the state of knowledge in the area and
prioritizing a multiyear research agenda; and . It also resulted in
substantial redirection of the interests and collaborative links of many of the
participants, thus helping to advance Geographic Information Science.
As a branch of philosophy, ontology studies the constituents
of reality. An ontology of a domain describes in formal terms the constituents
of reality within that domain: "Ontology as a branch of philosophy is the
science of what is, of the kinds and structures of the objects, properties and
relations in every area of reality" (Smith, 1999). Ontology thus provides
the basis for exchange of information, and. Thus, ontology is
a fundamental pre-requisite to description and explanation, in science and
elsewhere.
In simple terms, ontology seeks the classification of entities. Typically, philosophical ontologists produce theories that are very much like scientific theories, but of a far more general nature. Information systems also require a classification of entities, in this case so that separately built systems can satisfy the condition of interoperability. Ontology can provide in systematic fashion a framework for cross-disciplinary collaboration between different domains of science
Recently, the term 'ontology' has been used by information scientists to refer to canonical descriptions of knowledge domains, or to associated classificatory theories. In this sense, an ontology is "a neutral and computationally tractable description or theory of a given domain which can be accepted and reused by all information gatherers in that domain" (Smith, 1999).
Over the past two decades, progress has been made in
developing a comprehensive theory of geographic information, and in recognizing
the role of such a theory in the representation and modeling of phenomena and
processes. Several calls were made for such a theory during the early days of
quantitative geography in the 1960s, and several early ideas were published in
that period (a frequently cited example is Berry's geographical matrix, Berry
1964). However, it is important to note that geographic information systems
(GIS) software developed largely independently of this framework, and
in the absence of a comprehensive and widely shared theory, and with more
attention to geographic form than to process. As a result, the field has
been marked by a number of flaws. First, initial GIS prototypes tended to
implement only limited data models. Subsequently, new applications had to be
forced into these limited frameworks, or ad hoc extensions were required to
accommodate them (Goodchild, 1999). Second, the data models adopted by early
GIS tended to be chosen for their convenience in representing the contents of
maps, rather than for their effectiveness in supporting knowledge acquisition
and modeling in the geographic sciences. Third, because each GIS -vendor's
products have been developed largely independently, there is no standardization
of terminology among what are to a large degree independent information communities
centered on the various available products. Progress has been made, notably by
the Open GIS Consortium, but interoperability is clearly difficult to achieve
retroactively, and scientific applications are not of high priority within the
GIS industry. Finally, it has been difficult for new applications to make use
of GIS technology when the ontologies that underlie them fail to fit the data
models embedded in standard products. The costs associated with lack of a
common ontology increase steadily with the growth of GIS applications.
We believe that the time is ripe for the development of a
comprehensive geographic ontology for the domains of scientific research that
include or extend over geographic space. Much is now known about ontology in
general, and about its role in language, thought, description, data system
interoperability, and research. However, serious research on ontology of
geographic phenomena has begun only recently (Smith and Mark 1998; Mark et.
al., 1999). Furthermore, this work in geographic ontology focuses on form, and
either addresses rather specific kinds of geographic phenomena such as
fields (Peuquet et al., 1999), or relates to naive or common-sense geography
(Egenhofer and Mark, 1995; Mark et al., 1997) or to general principles.
Geographic ontology of scientific research domains has not received explicit
attention.
One of the first tasks in this proposed project is to define the formal framework for the entire domain of geospatial science, based on these and other ontological works. The understanding of geographic objects, fields, quantitative and qualitative relationships, spatiotemporal variation, scale, and accuracy developed in geography and geographic information science over the last several decades will provide detail for this general framework. The major task in this project is to examine key geographic research, application, and scientific domains within this framework.
A complete geographic ontology would define geographic objects, fields, spatial relations, processes, and their categories. It would include not only the basic data models, concepts, and representations necessary for scientific computing about geographic phenomena, but also the ontological principles and structures needed to support administrative, legal, commercial, and personal computing for phenomena in geographic space.
Consider objects first: At any scale, objects and their
definitions are intimately intertwined with the nature of their boundaries. In
the "blocks" world of the table top, things are often bona fide
objects, existing unambiguously and having crisp boundaries. Examples are an
apple, a book, or a pet. Larger organisms and artifacts, such as people, trees,
vehicles, and buildings also commonly have distinct boundaries. For geographic
things, however, as Smith and Mark (1998) have noted, there arises an issue of
individuation—if topography is a continuous field of elevation, how do people
parse such elevation fields into objects such as mountains, valleys, hills, and
ridges? If a mountain is an object, then clearly it is likely that it has
indistinct or indeterminate boundaries. Any ontology of geographic objects will
have to deal with boundaries of this sort (Burrough and Frank, 1996; Burrough,
1996). Also, some geographic objects have bona fide or genuine boundaries—‑islands
would be a good example in most cases. In contrast, other geographic objects
have fiat boundaries, created by legislative acts or other decisions—‑here
the prototypes may be land parcels or nation states (see Smith, 1995). Objects
themselves may be considered to be bona fide or fiat, depending on the nature
of their boundaries (Smith, 1994). Another aspect of the ontology of geographic
objects involves putting them into categories. Because individuation and
classification are not always independent, geographic object categories are
more likely to show individual, cultural, or disciplinary differences than are
table-top objects (Mark et al., 1999).
Of course, not all geographic phenomena are conceptualized as objects. Fields may be defined most simply as functions that map from position in space onto some measurement scale, including nominal scale. An ontology for common-sense geographic phenomena might be able to ignore geographic fields, but fields are critical to many scientific applications with geospatial dimensions, especially for fluid geographic domains such as the atmosphere and oceans. A complete geographic ontology would also provide definitions of spatial relations, of dynamic aspects of geospatial phenomena (events, change, motion, etc.), and of more complex geographic processes.
Although an ontology is in principle independent of a particular language, it is necessary to choose a language to describe it. In order to share, exchange, and combine ontologies, the language must be formal. Natural language alone would be insufficient, because much interpretation would be left up to the user, thereby potentially missing significant aspects of an ontology. The phenomena behind geographic ontologies are complex in nature, and therefore sophisticated knowledge representation methods are needed to represent or abstract them appropriately. In this section we briefly discuss basic elements of an ontology using the example of cave research (not one of our proposed domains, but an area that is familiar to one of us). An ontology can be formalized through definition of classes, relations, functions, and axioms [e.g., Uschold and Grüninger, 1996, and see Ontolingua (Farquhar et al., 1995) and CycL (Lenat and Guha, 1990) for examples of the languages and tools that have been defined in computer science for capturing ontologies]. This basic fourfold structure defines an ontology and will underlie all of our research within this project.
Classification is an abstraction mechanism that maps those individuals that respond to the same operations onto a common class. For example, the class "cave" is an abstraction for all hollow places of natural origin in cliffs or underground. A class definition includes the name of the class, its operations, and the names of its attributes. For instance, the class named "cave" includes such operations as enumeration and has such attributes as length, depth, and host geologic formation. No values are included for the attributes, since specific values are associated with individuals, not classes. A speleologist might create a class "sediment section" when investigating previous stages of cave development by cutting sections through deposits.
Classes are often interrelated by a generalization processes, capturing different levels of detail about the same individuals. The is-a relation describes the dependency between a subclass and a superclass, where the subclass has more detail than the superclass. For instance, a cave is often broken down into its component structural parts (passages, rooms) based on form, and into parts based on various stages of speleogenetic process, while a sediment section might be broken down into its constituent layers, with associated attributes of deposit type and interpreted age. A database recording the existence of caves might contain no detail on each cave's structural form, but might point to other databases describing the structural form of specific caves in great detail. In a simple case, an ontology describes a hierarchy of concepts created by a generalization process. Domains with more complex structures introduce additional relations and axioms that relate concepts, thereby constraining the interpretations of an ontology.
Two types of relations are critical to enable the integration of ontologies: synonyms describe when different class names are used for the same concept, while polysemic terms use the same class names for different concepts. For example, "canyon" and "fissure" may be synonyms when used to describe cave passages, while "deposit" is a polysemic term that could refer either to clastic sediment or to a speleothem.
Geographic ontologies are different from many other ontologies in that topology and part-whole relations play a major role in the geographic domain. Topology is important because geographic objects can be connected or contiguous, scattered or separated, closed or open, and because connections can exist both in geometric form and in conceptualizations of process. For example, two cave passages may be connected by a narrow squeeze; and two caves may be connected in the sense that both are part of the same hydrologic system. A theory of part and whole, or mereology, is important because geographic objects are typically complex and have constituent parts. The integration of topological methods with the ontological theory of part and whole leads to mereotopology. Therefore, part-of relations and other similar spatial relations are needed to describe geographic ontologies.
Functions are mappings or transformations applied to collections or entire classes of objects, and can be of various types. Basic GIS functions, such as projection and datum change, reformatting, resampling, or interpolation are familiar to and used by almost all scientists in working with geospatial data. For example, a speleologist might wish to display all objects of class cave by superimposing them on a digital topographic map, and recognize that to do so a transformation of coordinate system is necessary, e.g., from latitude and longitude to UTM. Spatial interpolation might be used to transform elevation data at sample points to a continuous surface. Other generic functions exist in the statistical analyses frequently used in scientific research. Of more significance here are functions that map one state of a system to a subsequent state, by emulating the operation of a process, because in these instances the objects involved in the mapping are likely to be domain-specific, as are the processes. For example, a speleologist might wish to simulate the evolution of a cave system based on initial and boundary conditions, and simulation code representing hypothesized solution and deposition processes.
Our concept of geographic ontology thus includes notions both of form and of process. While we currently have extensive technology for capturing and sharing information on geographic form, and summarizing such information in metadata, we have almost no techniques for sharing and summarizing knowledge of process. Thus our project will lay the groundwork for a move away from the limited concept of information as data, dominated by concepts of geographic form, to incorporate information on process. Although we argued above that techniques for dealing with data in GIS arose largely in ad hoc fashion, the lack of prior attention to techniques for dealing with process representation creates a valuable opportunity to build technology on sound theory.
The final element of an ontology is a set of axioms. These are rules about the system that can be imposed or evaluated. For example, it is axiomatic that an area must have a closed boundary, and this requirement can be checked by software, and in some cases any problems can be resolved automatically. In a less trivial case, it is axiomatic that mass is conserved in a cave system, allowing inferences to be made about erosion rates from knowledge of system output.
Of course, many of the issues reviewed above in relation to ontology have long been addressed in the GIS literature. In the GIS and computer mapping literature, the conceptualization of geographic phenomena and the implementation of the resulting concepts were variously referred to as data models or data structures. Peuquet (1984) provided one of the earliest reviews of GIS data modeling; in later papers she provided a more general conceptual synthesis (Peuquet 1988) and addressed time in a GIS context (Peuquet 1994). Mark (1979) wrote about these issues in the context of modeling elevation data, and included the idea that the digital representations adopted for geographic data should match the conceptual structure of the phenomena being modeled. Goodchild (1992) addressed the relationships between fields, objects, vectors, and rasters, while Frank (1992) clarified the relationships between spatial concepts, data models, and data structures. More recent comprehensive reviews of GIS data models can be found in the texts by Worboys (1995) and Burrough and McDonnell (1998), in the monograph by Molenaar and De Hoop (1994), and in the review of GIS by Longley et al. (1999).
Data modeling can be characterized as the task of creating
digital representations, and is intrinsically linked with digital technology, and in the
geographic case with form. Ontology, on the other hand, focuses on the
actual nature of the phenomena themselves. It seeks to capture, in a neutral,
formal framework that is independent of technical implementations, the basic
features of the objects—;
properties,;
relations,;
and processes— in [MG1]a
range of different domains. The resulting formal descriptions can then be used
as a basis for computational representations and for implementations with a
high degree of interoperability, as well as for scientific understanding.
In his "Naive Physics" manifesto (Hayes, 1978, 1985a), Patrick Hayes established an approach that went beyond the toy projects typically addressed by the artificial intelligence community. He followed this paper with another, entitled "Naive Physics I: Ontology of Liquids" (Hayes, 1985b), in which he gave an excellent example of a case where geographic phenomena differ ontologically from manipulable ones, providing a detailed discussion of how the ontology of lakes is different from that of many other objects composed of liquids. Egenhofer and Mark (1995) proposed that geographic information science should follow Hayes' approach, but in part because "geographic entities are ontologically different from enlarged table-top objects" (Egenhofer and Mark, 1995, p. 8), they proposed a new research enterprise termed "Naive Geography",
An explicit interest in geographic ontology was identified as an important researchable topic during the October 1996 Specialist Meeting for NCGIA's Research Initiative 21, on "Formal Models of Common-Sense Geographic Worlds." The Initiative 21 report (Mark et al., 1997) contained a draft geographic ontology as an Appendix, which included about 150 terms, grouped in a hierarchy with 18 categories at the highest level: conduit, intersection, landmark, place, topological feature, change, change in a property, egocentric feature, partition of the world, geometric feature, geographic feature, properties of geographic feature, location, spatial relation, shape, metereologica, institution, construction. Although this was a useful start, this ontology was an ad hoc effort that focused on common-sense (naive) rather than scientific concepts.
In his paper "On Drawing Lines on a Map", Smith (1995) examined a subclass of 2-dimensional geospatial objects that are distinguished by the fact that they have boundaries which are created by administrative decisions, boundaries which may be little influenced by topographical features of the Earth's surface. Smith termed the latter fiat boundaries and the resultant objects fiat objects[MG2]. Smith and Mark (1998) developed these ideas further and elaborated some important ways in which geographic objects differ from objects on the table to:
Geographic objects are not merely located in space, as are the manipulable objects of table-top space. Rather, they are tied intrinsically to space, and this means that their spatial boundaries are in many cases the most salient features for categorization. The ontology presented here will accordingly be based on topology (the theory of boundary, contact and separation) and on mereology (the theory of extended wholes and parts). Geographic reality comprehends mesoscopic entities, many of which are best viewed as shadows cast onto the spatial plane by human reasoning and language. (Smith and Mark, 1998, p. 308).
The analogy to shadows links this work to the foundational spatial ontology studies of Casati and Varzi (1995, 1999). Smith and Zaibert (1998) reported on the ontology of real estate, and Smith and Mark (1999) and Mark et al. (1999) reported on work that used results of human subjects testing to gain insight into the ontology of geospatial categories in general.
It is important to note that all of these prior works focussed on geospatial ontology at the most general levels, or on naive, common sense, or folk taxonomies and categories for geographic things. None were specifically involved in those aspects of the ontology of geospatial phenomena that arise when we adopt the perspective of the environmental and social sciences.
In 1992, the U.S. National Institute of Standards and
Technology (NIST), designated the Spatial Data Transfer
Standard (SDTS) as Federal Information Processing Standard (FIPS) 173,
mandating its use in data exchange involving U.S. government agencies
(Morrison, 1992, p. 277). In 1994, a Presidential Order (Clinton, 1994)
required federal agencies to work with State, local, and tribal governments,
and with the private sector to build the National Spatial Data Infrastructure,
NSDI. Fegeas et al. (1992) provided an overview of SDTS, which is available on
line from the Federal Geographic Data Committee (FGDC) web site. SDTS was the
culmination of a decade-long process of research and negotiation under the
National Committee for Digital Cartographic Data Standards (NCDCDS, 1988). The
intended role of SDTS is summarized in the following statements from the
standard:
"This document contains a specification of the Spatial Data Transfer Standard (SDTS), that will serve as a national spatial data transfer mechanism for the United States. As such it is designed to transfer a wide variety of data structures that are used in the spatial sciences. These sciences include cartography, geography, geology, geographic information systems and many other neighboring sciences." (NIST, 1992)
Later, the introduction to SDTS mentioned scientific research again:
"It is designed to serve the spatial data transfer needs of the Federal agencies, especially the proposed National Geo-Data System, and the work of State and local governmental entities, the private sector, and research organizations." (NIST, 1992)
Given these statements, one might assume that SDTS defines, and NSDI will populate, a geospatial data infrastructure to support research in the social and environmental sciences. However, the two sections quoted above from SDTS are the only places in the standard in which scientific research applications and uses are mentioned explicitly. In practice, the building of NSDI has rightly emphasized administrative and logistical data for use in the public and private sectors. It also has a decidedly cartographic flavor emphasizing phenomena normally portrayed on government topographic maps and navigational charts in the 1980s and earlier. We know of no serious efforts to assess the utility and limits of the SDTS model for scientific use. We propose to conduct such an evaluation, and provide details below. The ontology that will be defined in this research project will provide a basis for improving the utility of SDTS data in scientific research.
Currently, the International Standards Organization through its Technical Committee 211 is engaged in development of an international equivalent of SDTS. The effort is substantially more recent than SDTS, and reflects contemporary notions of data modeling. We propose to track the work of TC211 carefully during this project, and as with SDTS, to assess the utility and limits of the models proposed by TC211 for scientific use.
Of recent NSF awards relating to ontology, four mention geographic, spatial, or geospatial in the titles or abstracts. These are reviewed here to show that our proposed research does not duplicate prior and current NSF-funded projects.
The project directed by Ben Kuipers (1995) formalized the Spatial Semantic Hierarchy (SSH), a model for representations of spatial knowledge. This project, which ended in 1998, did not explicitly address spatial knowledge at geographic scales, and was mainly intended to support simulated and physical robots. The other three spatial ontology projects funded by NSF all began in 1999 and will extend from one to three years. Findler and Malyankar (1999) were funded by NSF's "Digital Government" program to determine an ontology for coastal entities such as shorelines and tide tables, in partnership with the US Coast Guard and NOAA. Mark and Smith (1999) have recently begin a 3-year project to determine the ontology of geographic objects and associated cognitive categories; unlike the scientific context of the current proposal, the context of Mark and Smith's project is general common-sense or naive geography, and the project emphasizes human subjects testing in a variety of languages. The results of the project are intended to contribute to spatial data transfer and semantic interoperability of general-purpose geographic software and data. Egenhofer's (1999) ontology project involves collaboration with the Brazilian National Institute for Space Research (INPE), and focuses on semantic interoperability of spatial and geographic databases. Again, it is important to note that none of these projects focused on ontology for scientific applications in the environmental and social sciences, and thus they have little overlap with the present proposed project. They will be complementary to the work proposed here, without being duplicative.
We will develop a three-part structure for the project. David Mark, Max Egenhofer, and Michael Goodchild will constitute the Investigative Team, with primary responsibility for the research. They will be assisted by graduate research assistants at each site. The project also will have a Consultative Panel to advise the investigative team on general and technical aspects of ontology-building. This panel will be composed of about ten international experts in general ontology and its methodologies, or in the theory and ontology of geographic phenomena. Barry Smith (Philosophy & NCGIA, Buffalo) will be a key member of the consultative panel. The project will begin with a joint meeting of the investigative team, the consultative panel, and the Domain Leaders (see below).
A series of Specialist Meetings will play a key role in the project. Each Specialist Meeting will examine a domains of geospatial science and scholarship in detail. Each meetings be led by a Domain expert, who will be neither an investigator nor a member of the consultative panel. The meetings will bring together approximately 30 people, including both domain experts and ontologists, and each will include about three members of the consultative panel. Approximately one third of the participants in each workshop will be experts on representation (computational or graphic) from the Geographic Information Science community. This mix of individuals from different academic communities will provide critical mass in each key area, maximizing the potential for progress.
In consultation with the domain leaders, we will appoint a small Steering Committee (of 6 to 8 people) for each Specialist Meeting. The steering committee members will assist the Leader and the investigative team with development of a draft ontology for the domain before the workshop, and with selection of other workshop participants. Each Specialist Meeting will include invited experts, and others selected from respondents to an open call for participation. Selection of participants from the open call will be based on brief position papers and biographic sketches. Participants will come from the academic, government, and private sectors, and will include experts from other countries, and we will strive for disciplinary and personal diversity across the participants.
To contribute to the development of human resources within
the U.S. research community, we will reserve five places in each domain
Specialist Meeting for U.S. early career scientists and engineers from
non-NCGIA institutions. Early career researchers will include people in the
later stages of their Doctoral studies, those who have completed their Ph.D.s
within five years of the date of the meeting, and holders of NSF CAREER awards.
Following NSF practice in education and human resources, we will restrict the early
career places to U.S. citizens and permanent residents. Otherwise, we will
actively solicit participants from outside the United States. Thus in addition
to contributing to the ontology-building process, the Specialist Meetings will
contribute to the extension of social networks within science, across academic
generations, disciplines, and international boundaries.
Prior to each Specialist Meeting, under the direction of the investigative team, the project's graduate assistants will compile a draft ontology for the domain, based on definitions of key concepts from the literature and other conceptual models of the domain. The draft ontology will include definitions of objects, digital representations, and entities from SDTS that appear relevant to the domain. Each draft ontology will be reviewed by the domain leader, other domain experts, and ontology experts from the Consultative Panel before distribution to Specialist Meeting participants. The draft ontology for each domain will be distributed to Specialist Meeting participants well in advance of the meeting, and will form a key basis for discussion at each meeting. Specialist Meeting participants will be asked to respond to questions regarding their domain's ontology. These questions will explore
· the extent to which data models in the domain are constrained or driven by those of data acquisition systems, software tools, and data sharing mechanisms;
· the ontology of the domain's body of theory and widely accepted process models; and
· the naive and vernacular ontology associated with phenomena within the domain.
As far as possible, participants' responses will be shared and discussed before the meeting, mainly through the use of electronic media.
The meeting itself will begin with plenary presentations on the objectives of the project and the methods being used. The three themes identified above will be used to structure a series of plenary and breakout discussions, with keynote presentations setting the stage for discussion in each area. Results of breakout discussions will be presented in plenary session. The meeting will conclude with a plenary session to address remaining issues and to endorse the major findings of the meeting.
Each specialist meeting will have three main products:
· a report on the discussions at the workshop;
· an interactive ontology of the domain; and
· a book chapter.
The first two products will be linked web documents. The Investigative Team will take the lead in producing these documents, in consultation with the domain leader and steering committee members, and assisted by the project graduate assistants and other student rapporteurs who attended the meeting.
After all six domain specialist meetings have been held, the investigators will compile a report on extensions or adjustments to SDTS and TC211's proposals that are needed to support scientific research and data exchange in those domains, and transmit that report to FGDC.
As noted above, the consultative panel will meet at the outset of the project to review and comment on the proposed methods. That panel will meet again at the end of the first year, after some specialist meetings have been completed, to review progress and to help the investigators modify the growing ontology and establish goals for the final four meetings. Finally, it will meet early in the third year to review the preliminary products of the project, and to comment on the methodology that has evolved. The leaders of the specialist meetings will be invited to join the consultative panel meetings.
A comprehensive ontology of geographic phenomena is
essential to geographic information science. However, in order to make maximum
progress within the time-frame of the current proposed project, we plan to
focus attention on a small number of critical domains. Scientific domains for
investigation have been selected based on two complementary principles. One
basis for the choice is that collectively, the domains should exhaust to the
greatest degree possible the geographic concepts involved in research in the
social environmental sciences. The domains chose must include both fluids and
solids, objects and fields, static and dynamic phenomena, and human and
physical systems. This strategy should maximize the degree to which the
combined ontology of these domains will cover other domains not studied
explicitly within this project. The second basis for choosing domains is their
importance as subdisciplines of geography, or other fields of social or
environmental science in which research using improved GISgeographic
information systems would be most likely to advance the domain
science. Improved ability of GIS to model these fields will enhance the
potential contributions of scientific research using GIS.
Based on these two sets of criteria, we propose to use six
domains for developing the geographic ontology. One of the most fundamental
categorizations based on phenomena and application domains would divide earth
systems into atmosphere, oceans, and land: ontological studies of atmospheric
sciences and oceanography are of very high priority, and we will address them
in this project. Land-based systems can be subdivided into lithosphere and
biosphere: here, we, we propose to focus on hydrology
and on terrestrial ecology. Of course, human society requires special
attention, and here we propose to highlight two critical domains: urban systems
and transportation. For each domain, we have asked a leading researcher to act
as Leader of the domain and to play a substantial role in populating and
organizing the specialist meeting. Names and short biosketches for each of the
proposed leaders are included below, and letters documenting their commitment
to the project are provided in Proposal Section I, Special Information and
Supplementary Documentation.
We will study these earth sub-systems in three pairs: atmosphere/ocean, hydrology/ecology, and urban/transportation. After the first specialist meeting in each pair, we will develop as complete as possible a formal ontology of the topic, and then use that as input to the second workshop. The topics have been paired in this particular way because of strong ontological similarities, as well as strong functional couplings.
Atmosphere and ocean are both fluid domains in which fields and field-like phenomena dominate. Except for organisms and artifacts, all 'objects' in the atmosphere and ocean, such as fronts, currents, and storms, result from human conceptualizations, both folk and scientific. Oceans and atmosphere also are fundamentally linked by energy exchanges through radiation, wind, and evaporation. However, relevant scales for modeling the two systems are different, in part due to the differences in the densities and viscosities of the fluids involved. Also, both domains have poor compatibility with commercial GIS software, and both include data measured by moving instruments that sample densely along their paths and sparsely across them.
As a dynamic, three dimensional gaseous domain, the
atmosphere might be best conceptualized as a number of four-dimensional
spatiotemporal fields of physical properties such as temperature, pressure,
humidity, etc. However, continuous four-dimensional multivariate fields are
difficult to visualize, think about, or compute. Folk theories of the weather
involve phenomena such as clouds and storms, which are conceptualized as
objects. Scientific models from the early 1900s introduced additional object
varieties such as fronts and air masses. Now we have tropical storms and
supercells that move as objects, and precipitation intensity data that
correspond roughly to moving fields. More recently, both short-term (weather)
and long term (climate) variations in atmospheric behavior have typically been
modeled in mathematical and computational systems over finite element grids or
meshes; computational limitations have often required these grids to be of
rather coarse spatial and temporal resolution. For some atmospheric processes,
the situation isf
further complicated by the necessity to model data and processes isotropically
over the sphere; spherical atmospheric modeling has been implemented in
Spherekit (Raskin, Funk, and Willmott, 1997), but needs further attention in
the context of the ontological issues discussed here.
Among the most active regions in the atmosphere and oceans
are the interfaces and transition zones, such as the planetary boundary layer
interfaces (the earth-atmosphere and ocean-atmosphere interfaces) and the
oceans' mixed layers. These regions not only are of above-average importance,
but they tend to be among the most difficult to adequately represent. Formal
structures for characterizing fluxes of energy, mass and momentum across
interfaces and through boundary layers will be a critical part of the ontology.
Representing the states of near-interface control
volumes also will be important. In the primarily fluid domain of the free
atmosphere, it is relatively easy to simply and accuracy represent
a control volume
simply and accuracy; that is, its physical, biological and chemical
states, as well as fluxes to and from it; this is not true near interfaces and
in boundary-layer situations. Representing fluid-field boundary conditions is
another aspect of the numerical representation and modeling of 4-D fluid
dynamics.
Against this backdrop of varied mathematical and
computational conceptualizations of atmospheric phenomena and processes,
atmospheric scientists must integrate and understand information coming from an
ever- increasing
range of sensing systems, including ground-based sensors, weather balloons,
Doppler radar, and satellites. The vertical spatial dimension is rarely sampled
at the same intensity or resolution as the two horizontal dimensions.
Atmospheric sensing systems present their own unique ontologies. Coupling
between atmospheric, oceanographic, and terrestrial processes is of critical
importance, especially in efforts to understand global environmental change.
Thus, the ontology of the atmosphere will have to be integrated with the domain
ontologies coming out of the elements of this project addressing oceans, hydrology,
ecology, and human systems.
Cort J. Willmott of the University of Delaware has agreed to lead the specialist meeting for this domain. Cort Willmott received his Ph.D. in Geography from UCLA in 1977. He is currently a Professor of Geography and Marine Studies at the University of Delaware where he also serves as Chair of the Geography Department, Director of the University's Center for Climatic Research, Director of the Environmental Science B.S. Program within the College of Arts and Science, and Associate Director of Delaware's (NASA-supported) Space Grant Program. His research interests include large-scale climate variability and change, land-surface processes and their influences on climate, spatial interpolation over extensive geographic domains, and the statistical evaluation of model performance. Grants primarily from NASA and NSF have supported Professor Willmott's research since 1980. He is a member of the Association of American Geographers, the American Meteorological Society, and the American Geophysical Union
This domain area will include physical oceanography, the
geology and geophysics of the ocean basins and margins, coastal geomorphology,
coastal zone management, marine biology, navigation, and fisheries. Physical
oceanography focuses on the structure and movement of the ocean, and with the
way in which it transports various quantities, and interacts with the air,
land, and ice that bound it. The science of physical oceanography requires
models and data that are truly four-dimensional, with intense sampling in all
three spatial dimensions and time. Physical oceanography also employs
object-like concepts of thermoclines and currents. Ocean floor topography and
processes are similar in many respects to those observed in the more familiar
terrestrial environment, but sensing devices differ sharply, as do objects of
interest such as spreading zones and fractures. Coastal geomorphology is
distinguished by the sharp difference in spatial resolution along and across to
the coastline that is needed to map or study processes. For
fisheries, concepts such as catch potentials and quotas must be formally
related to models of fish demography and predator-prey relations, just as the
predator-prey and fish growth models must include physical properties of the
ocean. Atmosphere-ocean linkages are especially critical, and we will evaluate
the ontology of the atmosphere with regard to modeling such interactions both
in general and in important specific cases such as the El Niño- Southern
Oscillation (ENSO) system.
We propose that the specialist meeting be led by Bruce Gritton, formerly of the Monterey Bay Aquarium Research Institute, who has a longstanding interest in data modeling in these domains.
This topic will focus on surface hydrology, but also will include groundwater flow and fluvial geomorphology. Hydrologic phenomena vary greatly in scale, and also in the level of aggregation that underlies modeling. Surficial phenomena are essentially two-dimensional, but groundwater modeling requires three spatial dimensions. Some models deal with steady states or equilibria, but others require high temporal resolution. Networks are important as representations of surficial systems (Mark, 1987), and are also sometimes useful for subsurface systems. Surface runoff models include parameters such as infiltration rates and surface roughness. Models of hillslope erosion include several distinct processes, and are coupled to fluvial erosion and deposition in channels.
In contrast to some of the other application domains, the
main data infrastructure for hydrology, namely surface topographic or elevation
data, has been a focus of attention in computer mapping and GIS since the
earliest days. Elevation may be thought of as a field, a single-valued function
of position in geographic space; since about 300 years ago, such fields have
been represented graphically by objects called contours, a special case of isolines. Computing about surfaces from
contours turned out to be inefficient, and initially, regular grids, lattices,
and altitude matrices proved more effective (Boehm, 1967). Later, it became
popular to represent surfaces within GISs as sets of non-overlapping triangles
in Ttriangulated
Irregular Networks (TINs; see Mark, 1997, for a history of TINs). However,
despite the fact that TINs are widely used in current commercial GISs, most
work modeling surface hydrology and fluvial erosion hasve
based those models on regular grid structures. A formal ontology of elevation
surfaces and of fluvial processes will be related to the data models used to
represent topography in current GISs and is expected to reveal how fluvial
process modeling can be implemented in the various representations.
David R. Maidment, a hydrologist and engineer at the
University of Texas at Austin, has agreed to lead the specialist meeting.
Professor Maidment is the Ashley H. Priddy Centennial Professor of Engineering
and Director of the Center for Research in Water Resources at the University of
Texas at Austin, where he has been on the faculty since 1981. He received his
Bachelor's degree in Agricultural Engineering with First Class Honors from the
University of Canterbury, Christchurch, New Zealand, and his MS and Ph.D.
degrees in Civil Engineering from the University of Illinois at
Urbana-Champaign. Dr. Maidment is a specialist in surface water hydrology, and
in particular in the application of geographic information systems to
hydrology. For a decade, he has been cooperating in this field with ESRI, manufacturers
of Arc/Info and ArcView, and with the Hydrologic Engineering Center of the US
Army Corps of Engineers. He and his research team have current projects
applying GIS for flood plain mapping, water quality modeling, water resources
assessment, hydrologic simulation, surface water-groundwater interaction, and
global hydrology. Since 1991, Dr. Maidment has taught each year a semester- long
course on GIS in Water Resources. Dr. Maidment is co-author of the 1988 text Applied Hydrology, and Editor in Chief of
the Handbook of Hydrology. From 1992
to 1995 he was Editor of the Journal of
Hydrology, and he is currently an Associate Editor of that jJournal
and of the Journal of Hydrologic
Engineering.
This domain emphasizes terrestrial ecology,
both for plants and for animals.
Perhaps the only bona fide objects in the ecological domain are individual
organisms (plants and animals), which themselves are not good examples of
geographic objects. However, there are strong mapping and GIS traditions in
this area that see vegetation as consisting of patches (polygons) with
internally -uniform
vegetation separated by sharp discontinuities. Real- world
geographic-scale objects in this domain tend to be aggregate objects such as
forests or herds of animals, or fiat objects such as administrative forest
stands. At a more generalized level, vegetation may appear more field-like, as
in the case of ecotones. Field-like concepts certainly are important in
characterizing climatic variation, often used as independent variables in
studies of plant and animal ranges. A field-object dichotomy underlies
continuing debates about vegetation mapping (Goodchild, 1994) as well as many
other geographic characterizations, including the delimitation of species range
and territory. Ecological concepts, such as niche, home range, and biome, must
be formally defined within a common framework that links these concepts with
the others listed here and with other environmental characteristics. The
ecosystems domain also includes fiat objects, where things like wetlands have legislated definitions.
Terrestrial ecology also is being revolutionized by new sensing technologies,
including animal and bird telemetry and remote sensing, and the ontological
foundations for ecological computing must model these measurements in ways that
allow them to be related formally to the phenomena being measured.
Janet Franklin, a geographer at San Diego State University,
has agreed to lead the specialist meeting for this domain. She has been on the
faculty at SDSU since 1988 as an Assistant, Associate and now Full Professor.
She received the Bachelors degree oin
Environmental Biology (1979), the Master of Arts (1983), and the Ph.D. (1988)
in Geography, all from the University of California, at
Santa Barbara. Her research interests include biogeography, landscape ecology,
biophysical remote sensing, digital terrain analysis, geographic information
science and conservation biology. She has conducted her research in the
Mediterranean-climate chaparral and conifer forest ecosystems of California,
and in arid regions of North America and West Africa, as well as the tropical
forests of remote Oceania. She is the Co-Editor of the journal The Professional Geographer and
Associate Editor of the Journal of Vegetation Science. She has published
over 30 refereed book chapters and papers in journals such as Remote Sensing of Environment, Journal of
Vegetation Science, Ecology, IEEE Geoscience and Remote Sensing, and and
Conservation Biology. She has received research support from NASA, NSF,
USGS, the U.S. Forest
Service, the National
Geographic Society, and other agencies.
The ontology of the transportation domain will include the
objects or substances moved, modes of transport, and fixed infrastructure to support
transportation. Terrestrial transportation systems (roads, railways, pipelines)
are commonly viewed as networks, and ship and air transportation can be modeled
in a similar fashion. Thus the transportation domain will provide context for a
focus on the ontology of networks, in particular networks embedded in
geographic space. Networks are composed of nodes, links, and link weights.
Nodes may function as origins, destinations, and demand points in the
transportation context. Intermodal transfer points such as harbors and airports
also are nodes with special roles. The fixed infrastructure component of
transportation systems often has been included in geographic databases. Early
GIS data models built representations from the links and nodes of street networks
(e.g., the TIGER databases). The limitations of this approach are well
documented, however (cf. Goodchild, 1998), including the problems modeling turn
restrictions at intersections, as well as overpasses, particularly for
applications to the modeling of human behavior. Paths are and routes are
subsets of the transportation network, and the transportation ontology will
include concepts of optimal paths and of algorithms and heuristics to find such
paths. Transportation can also be conceptualized as a special case of moving
objects, and the ontology of moving transportation objects will be formally
compared to the ontology of moving animals in the ecological domain. In
building the ontology of this domain, we will also model personal mobility,
including such classic topics as journey-to-work and multi-destination trip
modeling. The concept of telecommuting also will be included in the ontology,
in order to support its integration into general transportation demand models.
Harvey J. Miller, an Associate Professor of Geography at the
University of Utah, has agreed to lead the specialist meeting for this domain.
Miller received the Ph.D. in geography from The Ohio State University. His
research and teaching interests involve the interface between geographic information
science and regional science, particularly with respect to using
geocomputational methods to improve theory and application in transportation,
telecommunication and locational analysis. Since 1989, he has published close
to thirty papers in peer-referred journals and conference proceedings on these
topics. Dr. Miller is a recipient of the Hewings Award for Outstanding Young
Scholar from the North American Regional Science Council, the ESRI Award for
Best Scientific Paper in GIS from the American Society for Photogrammetry and
Remote Sensing and the Springer-Verlag Award from the Western Regional Science
Association. He serves on the editorial boards of the Geographical Analysis, Journal of Regional Science, Transportation, URISA Journal and is
North American Editor of the International
Journal of Geographical Information Science. Dr. Miller is also Chair of
the Mathematical Modeling/Quantitative Methods Specialty Group of the
Association of American Geographers. Miller directs the geographic information
science program at the University of Utah.
This domain will focus on urban
systems, includesing
human ecology, criminology, urban sociology, and urban economics. The study of
urban systems requires information that is field-like, in the form of densities
of population or income, or such variables as mean housing value or age;
information on network-like entities and processes, including transportation
systems; and information on the behavior of discrete objects, including
individual people and vehicles. Couclelis (1991) and others have commented on
the difficulty of representing the classic geographer's
classic concept
of situation in current GIS data
models, with its emphasis on relative rather than absolute space, and there is
an extensive literature on the potential for non-Euclidean concepts in modeling
the spaces of urban systems. Kevin Lynch's classic Image of the City (Lynch 1960), which has influenced both urban
planning and spatial cognition research, will provide some conceptual
frameworks
for this case study; Lynch's model was implemented in an artificial
intelligence context by Kuipers (1978). Place
is a particularly key and contested topic, and its multiple definitions and
meanings will be investigated. Similarly, neighborhood
is a familiar but ill-defined concept that mediates between social fields and
individuals, providing contextual input to human decision-making.
Urban studies are often based on aggregate data collected
for regions whose boundaries have little or no relation to discontinuities in
the phenomena being measured or counted. The prototypical example is census
data in North America, and census tracts, block groups, and other areas are
ideal examples of fiat objects. Often, urban analysts wish to compare data from
different surveys, but the measurement regions may not be the same. This gives
rise to what is now known as the modifiable areal unit problem (MAUP; Openshaw,
1984); the ontology of aggregate data based on fiat regions, and of the MAUP,
will be a key component of the ontology to support urban studies. Human
mobility is essential to understanding the city, and the urban ontology will
include portions of the transportation ontology. Hägerstrand's (1970) Time Geography model provides basic
concepts for linking space and time as joint constraints on human behavior, and also
will provide key concepts for the ontology.
Susan Hanson, Professor of Geography at Clark University,
has agreed to lead the Specialist Meeting on this topic. Professor Hanson is an
urban geographer with interests in urban transportation, urban labor markets,
and gender issues. Before earning the Ph.D. at Northwestern University (1973),
she was a Peace Corps Volunteer in Kenya. She has been the editor of several
academic journals including tThe Annals
of the Association of American Geographers and Economic Geography and currently serves on the editorial boards of
five other journals. Her publications include Ten Geographic Ideas that Changed the World (Rutgers University
Press 1997), Gender, Work, and Space
(with Geraldine Pratt, ) Routledge,
1995), The Geography of Urban Transportation
(Guilford Press, first edition 1986; second edition 1995), and numerous journal
articles and book chapters. For the past 20 years her research has been
supported by grants from the National Science Foundation, which is funding her
current project on Geography, Gender, and Entrepreneurship. She is a past
president of the Association of American Geographers, a Fellow of the AAAS, a
former Guggenheim Fellow, and a recipient of the Honors Award of the
Association of American Geographers and of the Van Cleef Medal from the
American Geographic Society. Hanson has served on many national and
international committees in geography, transportation, and the social sciences.
This project will result in two very significant products:
A comprehensive compendium of geographic concepts. We propose to determine the ontology of each of a carefully selected set of fields of social and environmental science that operate in geographic space. We will select domains that are representative of broad classes of phenomena and approaches, and also domains with a range of complexity from simple to problematic. Our project will also test and evolve a methodology that can be replicated elsewhere.
A test of the general framework. Implicit in the previous product is the assumption that it will be possible to construct a general ontology sufficient for the instantiation of all sciences dealing with geographic information. If we can find counter-examples, these become requirements for the modification or extension of the ontology.
These two products will have several benefits. We will know
more about how the needs of different domains can be met by geographic
information technologies, and will be able to provide better guidance to the
GIS software industry as it evolves to meet the needs of a larger number of
application domains. The ontology itself will be a fundamental piece of
geographic science, by linking theories of geographic information to the
broader array of scientific disciplines, and by
studying the importance of information in underpinning the geographic
dimensions of science, and by addressing the formalization of process as
well as geometric form. Finally, we will have developed a methodology
that can be applied to additional geographic domains, and to specialized
subdomains, in order to characterize their geographic information within a
general and comprehensive theoretical framework. We expect several journal
articles and conference presentations to result, as well as the web documents
and book described above.
The project will be under the direction of Mark as PI. The
project team has more than a decade of experience with the management of
distributed, multi-site research at the National Ccenter
for Ggeographic
Information and Analysis (NCGIA). We propose an award to the University at
Buffalo, with subawards to the University of Maine and the University of
California, Santa Barbara. Funds at each of the three institutions will be
allocated to summer salary for the respective investigative team member and to
research assistance. Other costs, including support of the six specialist
meetings and the meetings of the consultative panel, will be funded through
Buffalo.
The investigative team and research assistants will consult frequently by email, conference calls, and face-to-face meetings. We have collaborated productively in this way in the past and expect this to continue.
This project will have considerable synergy with existing projects at the NCGIA sites.
University at Buffalo received one of the first
cohort of Integrative Graduate Education and Research Training (IGERT) awards
from NSF in FY98, for a multidiscsiplinary
dDoctoral
studies program in Geographic Information Science. David Mark is Director of
the program. A feature of the program is a required Philosophy course in
geographic ontology, taught by Barry Smith and Roberto Casati. Fellows in the
program are from 7 departments in engineering and the social and environmental
sciences. The IGERT Fellows are not research assistants, but they are required
to participate in faculty-led research. We expect Buffalo IGERT Fellows to
participate in this project as additional project researchers, and in some
cases as rapporteurs at Specialist Meetings. Also at Buffalo, as noted above,
David Mark and Barry Smith are investigating the ontology of common-sense
geographic phenomena.
The proposed project also will complement several aspects of the work of the new Center for Spatially Integrated Social Science (CSISS) at UC Santa Barbara. Ontology plays a vital role in allowing researchers to build coherent and well-grounded models to support science. The urban and transportation domains are both firmly within the social sciences, and the development of geographic ontologies in these domains will integrate well with the work of CSISS. More specifically, we will incorporate the findings in these domains into the learning resources being developed by CSISS; into the development of software tools for spatial analysis under CSISS; and into the workshops that will be offered in each year of the CSISS award. The extensive network being developed by CSISS will also help us to ensure that a broad population of social scientists are aware of the opportunity to participate in the geographic ontology project, and to benefit from its results.
It is important to note that the project proposed
here will complement ongoing research at the NCGIA, but will not be possible
based only on the existing projects, since compilation of the ontology requires
synthesis across these and other projects, as well as the Specialist Meetings
to ensure scientific validity and to maximize completeness and adequacy to
support real geospatial research in the chosen and related domains[MG3].
Year 1: Appoint members of consultative panel, and hold the first meeting. Approve appointment of steering committees for the first group of domains, and develop draft ontologies for them. Issue the calls for participation in the specialist meetings for the first two domains (Atmosphere; Hydrology); select and invite participants for those meetings.
Year 2: Conduct first two specialist meetings, and publish their reports. Hold the midcourse meeting of the consultative panel and identify any needed corrections and improvements to the project. Approve appointment of steering committees for the remaining domains. Issue the call for participation in the remaining Specialist Meetings; select and invite participants. Conduct the second two specialist meetings (Transportation; Oceans), and publish their reports.
Year 3: Conduct the final two specialist meetings (Ecology; Urban), and publish their reports. Prepare a book tentatively titled "Ontology of Geographic Phenomena: Interoperable Science for Geographic Phenomena" and contract with a suitable publisher. Hold the final meeting of the consultative panel to review the project, assess its impacts, and review the evolution of the book.
Berry, B. J. L., 1964. Approaches to regional
analysis: A synthesis. Annals of the
Association of American Geographers 54: 2-11.
Boehm, B. W., 1967. Tabular representation
of multivariate functions with application to topographic modeling. Proceedings, 22nd National Conference,
Association for Computing Machinery, pp. 403-415.
Burrough, P. A.., 1996. Natural oObjects
with iIndeterminate
bBoundaries..
In Burrough, P. A. and Frank, A. U., eds., Geographic
Objects with Indeterminate Boundaries. Bristol, PA: Taylor & Francis,
pp. 3-28.
Burrough, P. A., and Frank, A. U., eds., 1996. Geographic Objects with Indeterminate Boundaries. Bristol, PA: Taylor & Francis.
Burrough, P. A., and McDonnell, R. A., 1998. Principles of Geographic Information Systems. New York: Oxford University Press.
Casati, R., and Varzi, A. C., 1995. Holes and Other Superficialities. Cambridge, Mass.: M.I.T. Press
Casati, R., and Varzi, A. C., 1999. Parts and Places. Cambridge, Mass.: M.I.T. Press
Clinton, W. J., 1994. Coordinating
Geographic Data Acquisition and Access: The National Spatial Data
Infrastructure. Presidential Order. Washington,:
DC:,
The White House, Presidential Order.
Couclelis, H., 1991. Requirements for planning-relevant GIS: A spatial perspective. Papers in Regional Science 70(1): 9-19.
Egenhofer, M. J., 1999. NSF-CNPq Collaborative Research on Integrating Geospatial Information. NSF Award IIS 9970123.
Egenhofer, M. J., and Mark, D. M., 1995. Naive geogrGeography.
In Frank, A. U. and Kuhn, W., editors, Spatial
Information Theory: A Theoretical Basis for GIS.,
Lecture Notes
in Computer Sciences No. 988. Berlin: Springer-Verlag, Lecture Notes in
Computer Sciences No. 988, pp. 1-15.
Farquhar, A., Fikes, R., Pratt, W., and Rice, J.,
1995. Collaborative
Ontology Construction for Information Integration. Technical Report
KSL-95-10. Stanford,
California: Knowledge Systems Laboratory, Stanford University.
Fegeas, R. G.,
Cascio, J. L., and Lazar, R. A., 1992. An oOverview
of FIPS 173, the Spatial Data Transfer Standard. Cartography and Geographic Information Systems 19 (5):,
278-293.
Findler, N. V., and Malyankar, R. M., 1999. Digital Government: An Ontology for Geospatial Knowledge. NSF Award EIA 9876604.
Frank, A. U., 1992. Spatial concepts, geometric data models,
and geometric data structures. Computers
and Geosciences, 18:,
409-417.
Goodchild, M. F., 1992. Geographic data modeling. Computers and Geosciences 18(4): 401-408.
Goodchild, M. F., 1994. Integrating GIS and remote sensing for vegetation analysis and modeling: methodological issues. Journal of Vegetation Science 5: 615-626.
Goodchild, M. F., 1998. Geographic information systems and disaggregate transportation modeling. Geographical Systems 5(1-2): 19-44.
Goodchild, M. F., 1999. Measurement-based GIS. In W. Shi, M.F. Goodchild, and P.F. Fisher, editors, Proceedings, International Symposium on Spatial Data Quality, Hong Kong. Hong Kong: Hong Kong Polytechnic University, pp. 1-9.
Hägerstrand, T., 1970. What aAbout pPeople
in rRegional
sScience?
Papers, Regional Science Association
24: 1-21.
Hayes, P., 1978. The nNaive pPhysics
mManifesto.
Iin:
D. Michie, D., ed. (Ed.),
Expert Systems in the Microelectronic Age.
Edinburgh, Scotland: Edinburgh University Press, pp. 242-270.
Hayes, P., 1985a. The sSecond nNaive
pPhysics
mManifesto.
Iin:
J. Hobbs, J., and R. Moore, R., eds, (Eds.),
Formal Theories of the Commonsense World.
Norwood, NJ: Ablex, pp. 1-36.
Hayes, P., 1985b. Naive pPhysics
I: Ontology of lLiquids.
Iin:
J. Hobbs, J., and R.
Moore, R., eds., (Eds.), Formal Theories of the Commonsense World. Norwood, NJ: Ablex, pp.
71-108.
Kuipers, B. J., 1995. An
Ontological Hierarchy for Spatial Knowledge. NSF Award IIS 9504138.
Lenat, D. B., and
Guha, R. V., 1990. Building Large
Knowledge-Based Systems: Representation and Inference in the Cyc Project. California:
Addison-Wesley.
Longley, P. A., Goodchild, M. F., Maguire, D. J., and Rhind,
D. W., editors, 1999. Geographical
Information Systems: Principles, techniques, management and applications.
New York: Wiley.
Lynch, K., 1960. The Image of the City. Cambridge, Massachusetts: MIT Press.
Mark, D. M.,
1979. Phenomenon-based data-structuring and digital terrain modelling. Geo-processing, 1:,
27-36.
Mark, D. M., 1987. Network models in geomorphology. Chapter
4 in: M. G. Anderson,
M. G., editor,
Modelling Geomorphological Systems,
pp. 73-97.
Mark, D. M., 1997. The hHistory
of gGeographic
iInformation
sSystems:
Invention and rRe-iInvention
of Triangulated Irregular Networks (TINs).
Proceedings, GIS/LIS'97.
Mark, D. M., and Smith, B., 1999. Geographic Categories: An Ontological Investigation. NSF Award BCS 9975557.
Mark, D. M., Egenhofer, M. J., and Hornsby, K. 1997. Formal Models of
Commonsense Geographic Worlds: Report
on the Specialist Meeting of Research Initiative 21.,
Report 97-2. Santa
Barbara, CA: National Center for Geographic Information and Analysis,
Report 97-2.
Mark, D. M., Smith, B., and Tversky,
B., 1999. Ontology and gGeographic
oObjects:
An eEmpirical
sStudy
of cCognitive
cCategorization.
In Freksa, C., and Mark, D. M., editors,
Spatial Information Theory: A Theoretical Basis for GIS.,
Lecture Notes
in Computer Sciences. Berlin: Springer-Verlag, Lecture Notes in
Computer Sciences, pp. 283-298.
Molenaar, M., and De Hoop, S., 1994. Advanced Geographic Data Modelling. Publications on Geodesy, New Series, Number 40. Delft: Netherlands Geodetic Commission.
Morrison, J. L., 1992. Introduction. Cartography and Geographic Information Systems 19 (5):,
277.
National Committee for Digital Cartographic Data Standards,
1988. The pProposed
sStandard
for dDigital
cCartographic
dData.
The American Cartographer,
15(1)[1]:,
9-140.
National Institute of Standards and Technology (NIST), 1992. Spatial Data Transfer Standard, Part 1, Logical Specifications.
Openshaw, S., 1984. The
Modifiable Areal Unit Problem. Concepts and
Techniques in Modern Geography (CATMOG) No 38., Norwich: Geo
Books.: Norwich
Peuquet, D. J., 1984. A conceptual framework and comparison
of spatial data models. Cartographica,
21:,
66-113.
Peuquet, D. J., 1988. Representations of geographic space:
towards a conceptual synthesis. Annals of
the Association of American Geographers, 78,: 375-394.
Peuquet, D. J., 1994. It's about time: a conceptual
framework for the representation of temporal dynamics in geographic information
systems. Annals of the Association of
American Geographers, 84:,
441-461.
Peuquet, D. J., Smith, B., and Brogaard-Pederson, B., 1999. Ontology of Fields.
Varenius Project
Specialist Meeting Report.,
Santa Barbara, CA: National Center for Geographic Information and Analysis.
Raskin, R., Funk, C. and Willmott, C. 1997. Interpolation
over large distances using Spherekit. ACSM/ASPRS Annual Convention & Exposition, Technical PapersSeattle, Washington,
Technical Papers, Vol. 5
(Auto-Carto 13), Seattle, Washington, pp.
419-428.
Smith, B., 1994. Fiat oObjects.
InFrom:
N. Guarino, N., L.
VVieu, L., and S. Pribbenow, S., eds., (eds.), Parts and Wholes: Conceptual Part-Whole
Relations and Formal Mereology,
11th European Conference on Artificial Intelligence, Amsterdam, 8 August 1994.,
Amsterdam: European Coordinating Committee for Artificial Intelligence, 15-23.
Smith, B., 1995. On dDrawing
lLines
on a mMap.
Iin
Frank, A. U., and Kuhn, W., (eds.),, Spatial Information Theory. Proceedings of COSIT
'95.,
Berlin: Springer Verlag, pp. 475-484.
Smith, B., 1999. Ontology: Philosophical and Computational. Unpublished manuscript, http://wings.buffalo.edu/philosophy/faculty/smith/articles/ontologies.htm,
Smith, B., and
Mark, D. M., 1998. Ontology and gGeographic
kKinds.
In T. K. Poiker, T. K., and N. Chrisman, N., eds., (eds.), Proceedings. 8th International Symposium on Spatial Data Handling (SDH'98), Vancouver.:
International Geographical Union, 1998pp.,
308-320.
Smith, B., and Mark, D. M., 1999. Ontology with hHuman
sSubjects
tTesting. American Journal of Economics and
Sociology, v. 58(2):
245-272.
Smith, B., and Zaibert, L., 1998. The Metaphysics of Real Estate. http://wings.buffalo.edu/ philosophy/faculty/articles/lz.html.
Uschold, M., and Grüninger,
M., 1996. Ontologies: Principles, methods and applications. Knowledge Engineering Review, 11(2).
Worboys, M.
F., 1995. GIS: A Ccomputing
Pperspective.
London: Taylor and Francis.
Ferrari, G., 1996. Boundaries, Concepts, Language. Geographic Objects with Indeterminate Boundaries. Bristol, PA: Taylor & Francis Inc, pp.99-108.
Smith, B., forthcoming. Fiat and Bona Fide Boundaries. In Philosophy and Phenomenological Research.
Smith, B., 1994. The Formal Ontology of Space: An Essay in Mereotopology. in: L. Hahn (Ed.), The Philosophy of Roderick Chisholm. Chicago and LaSalle: Open Court.
Smith, B., 1995. Formal Ontology, Common Sense and Cognitive Science. International Journal of Human-Computer Studies, 43, 641-667.
Smith, B., 1995. More Things in Heaven and Earth. Philosophische Studien, 50, 187-201.
Smith, B., 1995. The Structures of the Common-Sense World. Acta Philosophica Fennica, Vol. 58, 290-317.
Smith, B., 1997. Boundaries: An Essay in Mereotopology. In Hahn, L., ed, The Philosophy of Roderick Chisholm (Library of Living Philosophers), LaSalle: Open Court, 534-561.
Smith, B., 1997. On Substances, Accidents and Universals: In Defence of a Constituent Ontology. From: Philosophical Papers, 105-127.
Talmy, L., 1995. Fictive motion in language and 'ception. In Bloom, P., Peterson, M., Nadel, L., and Garrett, M. (editors), Language and Space. Cambridge, Mass.: MIT Press.
Teller, P., 1991. Substance, Relations, and Arguments About the Nature of Space-Time. The Philosophical Review. Vol. C, No. 3, pp363-397.
Yuan, M. In press. Modeling geographic information to support spatiotemporal queries. In Life and Motion of Socio-Economic Units edited by A. U. Frank, J. Raper, and J. P. Cheylan. European Science Foundation (ESF) Series (London: Taylor and Francis).
Yuan, M. 1999. Representing geographic information to enhance GIS support for complex spatiotemporal queries. Transactions in GIS 3(2): 137-160.
Ontology of Geographic
Phenomena:
Geospatial
Foundations for Social and Environmental Science
David M. Mark, Max J. Egenhofer, and Michael F. Goodchild
Recently, the term ontology has been extended to describe the formal conceptualization of knowledge domains so that they can be modeled in computer systems. This includes what is commonly referred to in the geographic information systems (GIS) literature as the data model and related generalizations. The definition of an ontology, and how mental acts relate to the real world, provides a definition for information itself.
When a mental act, or a system of mental acts, is directed towards an object in the real world, information is produced. Information may be defined as the conceptual or communicable part of the content of mental acts. Communication of information from one person to another is a central aspect of human life. Together, concepts and sensory data can be referred to as the content of the mental act. Mental acts link content to objects in the world, and also connect sensory data to concepts. This way of conceptualizing the nature of information is a simplified version of the theory of intentionality developed by philosophers such as Brentano and Husserl
An information system is an external (non-mental) system designed to store content. Information systems afford indirect transmission of content between people, some of whom build or populate the information system and some of whom use it. In a sense, the information system acts as a medium for this indirect communication. However, in order for communication to happen, the conceptual systems of the originators and users of the information must be sufficiently similar. When information systems or programs are used for the purposes of information exchange, they must satisfy similar conditions; when these are satisfied, the systems are said to be interoperable.
[MG1]I'm not sure this is what you intended
[MG2]Note that these terms have already been used without definition - we could move the definition earlier, or drop this definition section on the grounds that the definitions are intuitive
[MG3]Quite a sentence! I decided I didn't understand what it was saying. Can you rephrase, or drop?