Recent efforts by scientists and managers to inventory, map, and model impacts of human activities on the environment have focused on land transformation and urbanization processes. To test the efficacy of any single model, algorithm or procedure which defines land transformation processes a standard database calibration reference resource is required. Therefore, a set of georeferenced, spatially structured and well documented data sets has been designed for the Baltimore-Washington Region as a test and evaluation resource for the community of environmental modelers and global change scientists.
Land transformation processes are being examined from a variety of perspectives and scales using a variety of indicator parameters and mensuration variables. Tools and techniques applied to land transformation assessments range from creation of simple population expansion maps to change detection calculations using remotely sensed satellite data. A variety of point and cell growth models have been applied to simulate the land transformation phenomenon. These activities have demonstrated the reality that urbanization and land transformation processes involve complex interacting variables.
A team of scientists are expanding the efforts of the USGS Human Impacts on Land Transformation (HILT) project to build an Internet accessible "collaboratory" containing quality controlled spatially referenced calibration and validation databases. The Baltimore-Washington Regional Testbed provides for the calibration, verification, and validation for multiple scalar, temporal, thematic, and spectral assessments or models. This design and documentation procedures for creating the Baltimore-Washington Regional "Collaboratory" are presented in relation to its use for environmental modeling applications.
Urbanization can be described as a massive unplanned global experiment affecting increasingly large acreages of the Earth's surface (Alig and Healy 1987). Peter Vitousek (1994) described the ongoing land use/land cover change, along with increasing concentrations of carbon dioxide in the atmosphere and alterations in the global nitrogen cycle, as well-documented factors of concern for global change community. This massive change in land surface characteristics is just beginning to be studied by Earth system scientists in terms of ecological processes, atmospheric implications and micro and macro-climatic impacts. Questions regarding ecosystem structure and function along the urban-rural gradient have been raised are appropriate for the global change agenda (McDonnell and Pickett 1990). Recent investigations along the urban-rural gradient have provided new insights into the apparent impacts urbanization has introduced to stable ecological systems (Pouyat and McDonnell 1991; et al., 1994; et al., 1995). Conceptually, Earth scientists may agree that urbanization is a key determinant in the litany of ecosystem transition processes of interest to the environmental modeling community. However, a significant gap exists in the capabilities to address design and implementation of integrated modeling structures along the urban-rural gradient. Challenges of integrating and developing models to understand and predict processes of urbanization affecting environmental conditions requires an interdisciplined approach towards collaboration of the talents, modeling resources, and spatial data at local, regional, and global scales (Asrar and Dozier 1994, Blood 1994, Pickett and Cadenasso 1995). A collaborative testbed for the Baltimore-Washington region has been designed to meet many of these integrated environmental modeling challenges. The testbed (or Collaboratory) is a comprehensive set of spatial databases put together in a manner which reconstructs and represents the real world.
In partnership with the US. Geological Survey (USGS), the University of Maryland Baltimore County is compiling historic maps, demographic data, environmental parameters, and satellite images to map human-induced land transformations for the Baltimore-Washington region. This effort includes collaboration with the Bureau of the Census, the University of California at Santa Barbara, the Goddard Space Flight Center, and numerous other federal, state, local, and private institutions. This work builds upon earlier research by the USGS that documented urban development phenomena for the San Francisco Bay region. That effort, working under the premise that historic overviews of urban sprawl can be provide insight into future scenarios, used a geographic information system to compile and visualize historic perspectives from 1850 to 1990 (Kirkland et al. 1994). A methodology was developed to combine the information from a variety of sources into an integrated, multi-scale, and multi-resolution dataset. This methodology was expanded for the Baltimore-Washington effort to promote a variety of environmental, social, and economic models related to urban-rural dynamics. Contemporary analysis focuses on the use of remotely sensed data, existing digital land use data, digital census information, a variety of Earth science infrastructure data, such as Digital Line Graphs, Digital Elevation Models, and other key ancillary demographic information. The resulting database of temporal urban demographic changes, which forms the framework of the Baltimore-Washington testbed, provides an ideal source of information to calibrate and verify models for urban geographers, environmental scientists, and global change scientists.
Design and construction of large spatial databases for environmental modeling has remained topical for both the GIS and environmental modeling community as evidenced in part by the interest and content of the three international conferences sponsored by the National Center for Geographic Information and Analysis. Significant progress is being made by the modeling community for employing effective GIS entity-relationship-attribute schemas thus offering promise of improved GIS integration with environmental models in general. Object and feature-based schema are described by many modelers as a successful path for improving performance of environmental models relative to articulation with spatial databases (Guptill and Fegeas 1988, Raper and Livingstone 1995). Use of the temporal domain for modeling, requisite for change or trend analysis, poses additional challenges for integrating GIS and environmental models. Peuquet (1994) offers an overview of temporal data structure theory that indicates various avenues of approach are available for time series analysis. Raper and Livingstone (1995) offer object-oriented structures as an approach to including temporal dynamics in environmental spatial models. While alternative designs are beginning to incorporate temporal datasets as input to environmental models, contemporary modeling with GIS remains primarily focused on the application of time slices defining geographic entities either in raster or vector data structures (Kemp 1993, Kemp and Kowalczyk 1994, Mitasova et al. 1995, Farmer and Rycoft 1991). The Baltimore-Washington regional testbed is designed along the more general time slice database structures with emphasis on improved metadata documentation. While more conservative in terms of database development, this approach offers better calibration opportunities to modelers with the standard arc-node entity definition and relational attribute definitions. This paper details the decisions and steps that led to the creation of the Baltimore-Washington regional testbed and the environmental science application of the spatial database resources.
As part of the U.S. Global Change Research Program, the USGS initiated urban mapping research activities to understand the urban transition from a historical and multi-scale perspective appropriate for modeling and predicting regional patterns of urbanization into the future (Kirkland et al., 1994). The USGS Human-Induced Land Transformations (HILT) project initially involved mapping the growth of urban development for the San Francisco Bay area using archival topographic maps and Landsat satellite images to delineate changes in the urban extent over time (Bell et al. 1995). Visualization of the urban data maps using time-series animation resulted in an effective videotape presentation to both scientists and the public alike (Acevedo and Bell 1994). The modeling component of this effort involved developing a cellular automation urban growth model using the San Francisco Bay database (Clarke et al. 1996). Clarke (et al. 1995) developed the model by adapting a wildfire behavior, environmental model and derived calibration techniques applicable to the land cover and other urbanization growth barriers using the temporal urban database. Continued refinement of the model which includes examining multi-scale extensions and interrelationships of various urban parameters is reported in a separate paper by Clarke (1996) at this conference.
Building upon the San Francisco Bay HILT activities was a collaboration
that included UMBC, the U.S. Census Bureau and others to extend the
temporal mapping into the Chesapeake Bay region. This collaboration
created a multi-phased research plan, Table 1, which entails the creation
of a multi-thematic, multi-temporal, multi-scale and multi-resolution
spatial database structure for the greater Baltimore-Washington region,
Figure 1.
This multi-year collaboration continues to support activities assembling an integrated and flexible temporal urban land characteristics database encompassing the period from 1792 to 1992. Phase I focused on testing the HILT methodologies for an area encompassing greater Baltimore metropolitan area. Database design and construction, metadata documentation, and basic visualization methods have been tested and implemented using the Phase 1 database (Acevedo et al. 1996, Masuoka et al. 1995). Phase II efforts are currently expanding the database development for the entire 2-degree by 2-degree region. Phase III and IV will focus on experimenting with selected mapping themes, analyzing spatial patterns and rates, and linking with various environmental models. Included in the Phase I and II database development are temporal mapping layers for primary transportation, hydrography, and population density. Derivation of these data layers comes from the archives of historic maps and records prior to the 1970's and digital data in the post 1970's era using Landsat imagery, Digital Line Graphs, Digital Elevation Models, and DIME and TIGER files.
The expansion of activities and contributing agencies related to the
Baltimore-Washington regional spatial database are lending support to the
testbed concept, Figure 2.
In as much as this testbed provides a distributed, Internet accessible resource for environmental scientists as well as interested local and regional planners, the term "collaboratory" has been applied. Other activities associated with the Baltimore-Washington Collaboratory include data visualization research, NSDI metadata compliance and testing, user community outreach, and applied science modeling. A multi-disciplinary team has expanded on methodology, definitions, and collection criteria used to define the various data layers, ensuring consistency in data definitions and data collection techniques among the different collaborators.
In the design of large spatial databases a variety of concerns and constraints must be taken into account to avoid outright failure or at a minimum reduce the inefficiencies of the system using the database structure (Marble 1988). Databases designers have adopted various approaches to addressing the creation of spatial databases, and databases in general, which almost universally begin with defining the user requirements or functional requirements of an individual, organization, or application process or model. From this setting the designer can construct the bounding parameters of systems, data, people, and financial resources into a conceptual template for the database to exist and perform its primary function. Systems designers have many models to follow but the similarity among these models is more distinctive than the minor, and more often, semantic differences (Pressman 1987, Teorey and Fry 1982, King 1984). In GIS applications, approaches to successful systems designs were adapted from the more ubiquitous non-spatial systems engineering protocols (Calkins 1982, Calkins and Marble 1987). Today, GIS consultants use essentially the same design to develop an integrated GIS installation and the requisite database for businesses or municipalities. Experience with "successful" installations has kept the system design process more narrowly defined than many outsiders might expect, to the point that "boilerplate" assessments comprise a significant portion of many contracted design studies. This does not demean the professionalism of GIS design consultants, who must have the expertise to know what is boilerplate and what is not, but illustrates that the process of system design and database design follows a structured series of fundamental steps from design to implementation.
For the Baltimore-Washington Collaboratory, the design of the spatial database follows a new paradigm outside the traditional GIS construction domain. This new paradigm is based on the realities of the collaborating data providers and users and therefore can be described from a couple of perspectives. From the data providers perspective the Collaboratory is following the general constructs of both the National Spatial Data Infrastructure (NSDI) and the NASA Mission to Planet Earth (MTPE) plans for handling Earth observations from space (National Research Council 1994, NSTC 1995). From the data user community another set of perspectives is evident which includes environmental modelers who will plan to use the Collaboratory assets as not only model input but as a means to calibrate and validate their models (Oreskes et al. 1994, FGDC 1995). Other users will likely view the Collaboratory as a source of input for regional planning purposes to analyze and predict rates of land use change and establish the causal factors related to the land use transitional processes. The most challenging user community represents those developing integrated regional models to couple environmental, human, and physical models (Blood 1994, NSTC 1995). This latter group of modelers will be instrumental in redefining both the identity of content and structure of the Collaboratory's assets. It therefore becomes incumbent on the Collaboratory designers and modeling community to creatively deal with issues of calibration, validation, uncertainty and error propagation, simplification or aggregation, resolution and scale as they impact the performance of integrated regional models or environmental models in general. These issues serve as primary assumptions in the Collaboratory design and therefore must in part be assessed on a case by case basis with individual modeling teams but with attention to the ramifications of any design constraints on the use of the spatial database for general modeling applications. With these caveats, a conservative approach has been applied to the creation and documentation of the spatial database.
Using contemporary GIS capabilities, the Baltimore-Washington Collaboratory assumes data assets to be either digital vector or raster with associated attribute files and metadata. Data sets initially represented in the Collaboratory, Figure 3 encompass:
It is assumed that most vector datasets will be converted to grid formats for input in cellular autonoma, finite element or finite difference modeling structures (Coucleilis 1985, Clarke et al. 1996). Other uses of the vector data sets would include referencing of geographic phenomena, via hydrology or transportation alignments or as vector overlays for improved comprehension of associated datasets. Error propagation attendant to vector-to-raster conversion remains unavoidable and will necessarily be the responsibility of the user (Lunetta et al. 1991).
Digital data resources of the Baltimore-Washington Collaboratory will be accessible via the Internet and in bulk media formats. Physical location of the Collaboratory's spatial database assets is distributed among the cooperating data providers with a few nodes accepting additional responsibilities to serve as Regional Data Centers (RDCs) under the guidance of NASA's Earth Observing System Data and Information Systems (EOSDIS) general protocols. A significant assumption for the Collaboratory is that data providers (e.g., NASA, USGS, NOAA) should remain stewards of data generated whenever possible to keep data redundancy to a minimum, maintain metadata documentation, and provide updates and upgrades to data as appropriate. The Collaboratory will ensure that data resources included in the data catalog have been compiled with cooperative protocols as defined later in the text under Documentation/Metadata. The RDC design acknowledges that many of the digital data resources will require various preprocessing steps to make the data suitable for some applications for local or regional users and environmental modelers. For example, AVHRR or Landsat data may require some data format handling before use on PC based GIS software packages by county planners and decision makers.
In functioning as a Regional Data Center, the Baltimore-Washington Collaboratory, using personnel from UMBC and NASA Goddard Space Flight Center are working with various environmental modelers. One example of an environmental application incorporates the testing of the Hydrologic Simulation Program in Fortran (HSPF) hydrologic model for performance along an urban-rural gradient using input parameters from the Collaboratory resources. The modeling evaluation, in cooperation with personnel of the USGS, Yale School of Forestry, and the Institute of Ecosystem Studies, entails examining how HSPF performs along this gradient at cascading spatial scales or grid resolutions. The RDC functions for the Collaboratory then serve in an iterative fashion to both supply data for modeling and to share the results back to the community for assessment and planning purposes. Another modeling application entails providing data for testing and calibrating an urban growth cellular autonomon model developed by Clarke (1996). The results of this model are planned to be available over the Internet, including operating code and data. This approach will provide local land use managers with virtually no-cost tools to examine population growth for their regions 50 years into the future, while also allowing them to analyze the past growth phenomena with the Collaboratorys historic land use and demographic data sets.
The importance of using the Baltimore-Washington Collaboratory to verify, calibrate, and validate researcher's models includes research on performance of remotely sensed data. By attending to careful geographic registration of the regional infrastructure, land use, demographics, topography, and other physical data sets as part of the Collaboratory shared digital resources, remote sensing scientists can utilize Collaboratory resources for ground truth calibration. This will become increasingly important as a host of new sensors being developed for the EOS program begins to produce data. In addition, the user community of local environmental and land use managers and planners and commercial entities will look to the RDC/ Collaboratory data sets to determine the applicability of EOS information for their local/regional applications.
A variety of data integration processes are involved in the creation of the Baltimore-Washington Collaboratory assets that must be explicitly defined for the user community. These definitions are included under the metadata design protocols directly from the federal metadata standards (FGDC 1994). Both federal and Maryland state agencies have been directed to comply with these standards. While the data provided to the Collaboratory varies in format and quality, the attention to the details of metadata documentation provide environmental modelers with the information required to determine goodness-of-fit for their modeling use. While the federal metadata standards have been viewed as unfunded mandates, compliance provides the modeling community with a rich resource of digital data that would otherwise be risky in terms of adding uncertainty to their modeling parameters (Berk 1994). This adherence to metadata documentation is not a trivial exercise and has required significant use of project personnel resources but serves as critical input into the NSDI national resources. Numerous technical problems have been discovered in the implementation of the FGDC metadata standards, solutions to these implementation problems should help with the design of future NASA RDCs as NASA will need to understand the requisite administrative overhead necessary to keep in FGDC compliance.
Initial results from the Baltimore-Washington Collaboratory have demonstrated the usefulness of an hierarchical approach to metadata documentation under a hybrid FGDC schema currently under testing. The hybrid approach is designed to streamline the inclusion of local and regional digital data resources from agencies that do not comply with federal or state standards. This will enable an increase in data resources available through the Collaboratory at fine resolution scales (1 meter to 10 meters) while still attending to the philosophy for goodness-of-fit labeling requirements, Figure 3. Performance testing of the hybrid schema is scheduled for summer of 1996.
The creation and utilization of a regional spatial data testbed is critically needed in the environmental community to calibrate, verify, and validate the various models. This requires that the design of a regional, digital spatial data resource be established in a general GIS structure to support the ready import of data into a variety of environmental models. Development of spatial models continues to mature for multi-scale, multi-temporal, and multi-thematic applications under a variety of schema for entity-attributes. Progress has been reported by Raper and Livingstone (1995) and others (Guptill 1988, Shi and Zhang 1995) in the use of object or feature base representation of spatial, temporal, and attribute modeling. A regional digital database must attend however to the needs of the many and therefore a conservative GIS structure was selected for the Baltimore-Washington Collaboratory. In addition, by aligning the Collaboratory protocols with national standards and trends, modelers using the digital data resources can benefit from working in the context of the national spatial data infrastructure where efforts to correct and improve the national protocols will result in more meaningful approaches for long-term use of their models.
Environmental modeling needs to be understood from a broader context than the disciplines of origin. Integrated assessment models and integrated regional models will require increased understanding of the semantics and parameter formats of different modeling schools. Environmental models using GIS data structures and resources will require extension into the domains of human ecology, urban environments, landscape ecology, sustainability, and ecological economics to meet the demands for improved decision making and management applications. It is envisioned that through the application of quality documented data resources, available from RDCs such as the Baltimore-Washington Collaboratory, development of integrated environmental modeling can be better accomplished in the future.
We wish to thank Janet Crawford and Susan Clark of the USGS in Reston, Virginia for their assistance in this project. The work was funded in part by NASA Research Grant NAGW-1743.
Acevedo, W. and C. Bell. (1994) Time series animation of historical urban growth for the San Francisco Bay region. Abstracts Association of American Geographers 90th Annual Meeting. San Francisco, CA, pp. 2.
Alig, R.J. and R.G. Healy. (1987) Urban and built-up land area changes in the United States: an empirical investigation of determinants. Land Economics. Vol. 63, pp. 215-226.
Asrar, G. and J. Dozier. (1994) EOS: Science Strategy for the Earth Observing System. AIP Press, Woodbury, NY.
Bell, C., W. Acevedo and J. T. Buchanan. (1995) Dynamic mapping of urban regions: growth of the San Francisco Sacramento region. Proceedings, Urban and Regional Information Systems Association. San Antonio, TX, pp. 723-734.
Berk, Richard A. (1994) Uncertainty in the construction of interpretation of mesoscale models of physical and biological processes. In P. Groffman and G. Likens (eds.). Integrated Regional Models: Interactions Between Humans and Their Environment, Chapman and Hall, New York, NY, pp. 50-64.
Blood, E. (1994) Prospects for the development of integrated regional models. In P.M. Groffman and G.E. Lines (eds.). Integrated Regional Models: Interactions Between Humans and Their Environment. Chapman and Hall, New York, NY, pp. 145-153.
Calkins, H. W. (1982) A pragmatic approach to geographic information system design. In Pequet and O'Callaghan (eds.). Proceedings U.S./Australia Workshop on Design and Implementation of Computer-Based Geographic Information Systems. Amherst, NY: IGU Commission on Geographical Data Sensing and Processing.
Calkins, H. W. and D. F. Marble. (1987) The transition to automated, production cartography: design of the master cartographic database. The American Cartographer. 14 (2).
Clarke, K.C., J. A. Brass, and P. Riggan. (1995) A cellular automaton model of wildfire propagation and extinction. Photogrammetric Engineering and Remote Sensing, Vol. 60, No. 11, pp. 1355-1367.
Clarke, K., S. Hoppen, and L. Gaydos. (1996 forthcoming) Methods and techniques for rigorous calibration of a cellular automaton model of urban growth. Third International Conference/Workshop on Integrating GIS and Environmental Modeling, Santa Fe, New Mexico, January 21-25, 1996.
Coucleilis, H. (1985) Cellular worlds: a framework for modeling micro-macro dynamics. Environment and Planning , Vol. 17, pp 585-596.
FGDC. (1994) Content standards for digital spatial metadata. Federal Geographic Data Committee, Washington, D.C.
FGDC. (1995) National geodata forum. Federal Geographic Data Committee. May 1995, Reston, VA.
Farmer, D.G. and M. J. Rycroft. (1991) Computer Modeling in the Environmental Sciences. Oxford: Clarendon Press.
Guptill, Stephen C. and Robin G. Fegeas. (1988) Feature based spatial data models--the choice for global databases in the 1990's? In H. Mounsey and R. Tomlinson (eds.). Building Databases for Global Science, Philadelphia, PA: Taylor & Francis, pp. 279-295.
Kemp, K. (1993) Environmental Modeling with GIS: A Strategy for Dealing with Spatial Continuity. Technical Report 93-3, National Center for Geographic Information and Analysis, Santa Barbara, USA.
Kemp, Z. and A. Kowalczyk. (1994) Incorporating the temporal dimension into a GIS. In M. Worboys (ed.). Innovations in GIS. London: Taylor & Francis, pp. 89-103.
King, D. (1984) Current Practices in Software Development: A Guide to Successful Systems. New York: Yourdon Press.
Kirkland, D., L. Gaydos, K. Clarke, L. DeCola, W. Acevedo and C. Bell. (1994) An analysis of human-induced land transformation in the San Francisco Bay/Sacramento area. World Resource Review, Vol. 6, No. 2, pp 206-217.
Lunetta, R. S., R. G. Congalton, L. K. Fenstermaker, J. R. Jensen, K. C. McGwire, and L. R. Tinney. (1991) Remote sensing and geographic information system data integration: error sources and research issues. Photogrammetric Engineering and Remote Sensing, 57(6), pp. 676-687.
Marble, D. F. (1988) Approaches to the efficient design of spatial databases at a global scale. In H. Mounsey and R. Tomlinson (eds.). Building Databases for Global Science, Philadelphia, PA: Taylor & Francis, pp. 49-65.
Masuoka, P., T. Foresman, S. Fifer, W. Acevedo, S. Clark, J. Crawford, and J. Buchanan. (1995) Visualization techniques for the analysis of Baltimore regional GIS data. Proceedings Volume 2, GIS/LIS '95. Nashville, TN, pp. 704-712.
McDonnell, M.J. and S.T.A. Pickett. (1990) Ecosystem structure and function along urban-rural gradients: an unexploited opportunity for ecology. Ecology, Vol. 71, pp. 1232-1237.
Mitasova, H., L. Mitas, and W. M. Brown, D. P. Gerdes, I. Kosinovsky, and T. Baker. (1995) Modelling spatially and temporally distributed phenomena: new methods and tools for GRASS GIS. International Journal of Geographical Information Systems, 9(4), pp. 433-446.
National Research Council. (1994) Science Priorities for the Human Dimensions of Global Change. National Academy Press, Washington, D.C.
National Science and Technology Council (NSTC). (1995) Our changing planet: the FY 1995 U.S. global change research program. A Report by the Subcommittee on Global Change Research, Committee on Environment and Natural Resources Research of the National Science Technology Council.
Oreskes, N., K. Shrader-Frechette, and K. Belitz. (1994) Verification, validation, and confirmation of numerical models in the earth sciences. Science, Vol. 263, pp. 641-646.
Pickett, S.T.A. and M.J.Cadenasso. (1995) Landscape ecology: spatial heterogeneity in ecosystems. Science, Vol. 269, pp 331-334.
Peuquet, Donna 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, Vol. 84, No. 3, pp. 441-461.
Pouyat, R.V. and M.J. McDonnell. (1991) Heavy metal accumulations in forest soils along an urban-rural gradient in southeastern New York, USA. Water, Air, and Soil Pollution. Vol. 57-58, pp. 797-807.
Pouyat, R.V., M.J. McDonnell, and S.T.A. Pickett. (1995) Soil characteristics of oak stands along an urban-rural land-use gradient. J. Environmental Quality, Vol. 24, pp. 516-526.
Pouyat, R.V., R.W. Parmelee, and M.M. Carreiro. (1994) Environmental effects of forest soil-invertebrate and fungal densities in oak stands along an urban-rural land use gradient. Pedobiologia., Vol. 38, pp. 385-399.
Pressman, R. S. (1987) Software Engineering: A Practitioner's Approach. (Second Edition). New York: McGraw-Hill Book Company.
Raper, J. and D. Livingstone. (1995) Development of a geomorphological spatial model using object-oriented design. International Journal of Geographical Information Systems, 9(4), pp. 359-383.
Shi, W. and M. Zhang. (1995) Object-oriented approach for spatial, temporal, and attribute data modeling. Proceedings Volume 2, GIS/LIS '95, Nashville, TN, pp. 903-912.
Teorey, T. J., D. Yang, and J. P. Fry. (1986) A logical design methodology for relational databases using the extended entity-relationship model. Computing Surveys, 18(2), pp. 197-222.
Vitousek, P. (1994) Beyond global warming: ecology and global change. Ecology, Vol. 75, No. 7, pp. 1861-1876.