An Open Geographic Modeling System
Abstract
Developing the complex computer models that are necessary for effectively managing human affairs through the next century requires new infrastructure supporting high performance collaborative modeling.
In this paper we describe an ongoing program to develop new modeling technology which supports 1) modular, hierarchical model construction and archiving/linking of simulation modules, 2) graphical, icon-based model construction, 3) transparent distributed computing, and 4) integrating multiple space-time representations. It is hoped that this type of infrastructure will open the simulation arena to a much wider set of participants, and facilitate the application of computer modeling to the study of complex multi-scale processes in support of science, education, and policy making on many levels.
We believe that effectively managing human affairs through the next century will require extremely complex and reliable computer models. Programs such as NASA's Mission to Planet Earth (MTPE) augmented by the development of the Open Geodata Interoperability Specification (OGIS) by the Open GIS Consortium have paved the way by providing an open framework for interoperable geodata access and processing. However, the crucial next step, the development of an open framework to support collaboprative dynamic modeling, has been all but ignored. This paper describes an ongoing program to develop this crucial infrastructure for open geographic modeling.
Generally speaking, a dynamic geographic simulation is a computer-based model that simulates the spatio-temporal dynamics of a landscape at scales ranging from a watershed to the globe. It is structured as a spatial grid/network of site-models, each representing the dynamics of a landscape component. A site-model incorporates multiple mathematical, logical, and stochastic process models created and defined by a multidisciplinary research team. These process models calculate the state of the system at a time t+dt as a function of the state at time t. The initial conditions are seeded by the use of a "picture" or "snapshot" of the system at some real or artificial start time. Seeding may use various forms of input data including raster images, vector data, point information, and object status and location. Recent advances in computer processing speed and efficiency have made it possible to incorporate spatial components of landscapes -i.e. population distribution, specific landscape conditions- into dynamic, geographic models.
Spatially explicit (geographic) modeling is essential for realistically addressing many resource management and environmental impact scenarios related to industry, agriculture and the natural environment. Across the nation, land management offices responsible for the management of natural resources, endangered species, water quality, aesthetics, and economic productivity of the land, are turning toward dynamic, geographic modeling. The development of these simulations has in general been limited by the ability of any single team of researchers to deal with the conceptual complexity of formulating, building, calibrating, and debugging complex models. "The complexities of science are outstripping the capability of any single institution to embrace the full range of pertinent expertise in any area" (Towards a National Collaboratory, NSF Workshop 3/17 '89). Collaborative, interpersonal computing must become a focus for efforts to increase the possibilities for and productivity of interdisciplinary teams working together on geographic modeling projects. This community must be tightly coupled within a worldwide "collaboratory" based on new electronic information sharing technologies, bringing together experts in computational and life sciences with policy makers and stake holders. The initial groundwork for this collaboratory has been laid with the development of the Open GIS interoperability standard (OGIS). Over the past six years our team has been developing strategies and software designs to take the next step toward a worldwide Open Geographic Modelling System (OGMS).
There exists a rich set of research problems associated with the implementation of computer based collaborative technologies for high performance geographic modeling. Five important areas of ongoing research and development are integrated support for 1) modular, collaborative model development, 2) transparent access to high performance computing resources, 3) graphical display & manipulation of model structure and dynamics, and 4) multiple integrated spatio-temporal representations.
The development of team approaches to computer-based resource management reflects the increasing reliance on experts from multiple disciplines with the variety of specialties necessary to address all aspects of the complex management scenarios. The development of realistic models in this field can require collaboration between species specialists, hydrologists, chemists, land managers, economists, ecologists, and others.
Supporting collaborative modeling requires structuring the model as a set of distinct modules with well-defined interfaces (Gauthier and Ponto, 1970; Goodall, 1974; Acock and Reynolds, 1990; Silvert, 1993). Modular design facilitates collaborative model construction, since teams of specialists can work independently on different modules with minimal risk of interference. Modules can be archived in distributed libraries and serve as a set of templates to speed future development. The inheritance property of object-oriented languages allows the properties of object-modules to be utilized and modified without editing the archived object. A modeling environment that supports modularity could provide a universal modeling language to promote worldwide collaborative model development.
High performance computing gains its strength from transparent availability of computational resources. Developments to date have been focused on new hardware and software architectures designed to efficiently map an application to the computing environment, and methods of distributing processes to appropriate platforms. Much less effort has gone into efficiently mapping the teams of researchers who use these high performance systems into the high performance environment. For this reason high performance systems have seen little use in the field of environmental modeling and resource management, even though this class of models is a near ideal application for distributed processing (i.e. a typical model consists of a large number of cells, each containing a computationally intensive unit model that can be executed semi-independently; each computational node can be assigned a different subset of cells, and most inter--node communication is nearest-neighbor only). Our development program is decicated to addressing this issue by transparantly integrating high performance distributed computing resources into the model development environment.
A second step toward reducing model complexity involves the utilization of graphical, icon-based module interfaces, wherein the structure of the module is represented diagramatically, so that new users can recognize the major interactions at a glance. Scientists with little or no programming experience can begin building and running models almost immediately. Inherent constraints make it much easier to generate bug-free models. Built-in tools for display and analysis facilitate understanding, debugging, and calibration of the module dynamics.
One major advantage of this graphical approach to modeling is that the process of modeling can become a consensus building tool. The graphical representation of the model can serve as a blackboard for group brainstorming, allowing students, policy makers, scientists, and stakeholders to all be involved in the modeling process. New ideas can be tested and scenarios investigated using the model within the context of group discussion as the model grows through a collaborative process of exploration. When applied in this manner the process of creating a model may be more valuable than the finished product.
Building realistic spatially explicit geographic models requires the integration of multiple spatial data structures in a single model. For example, variables such as elevation and vegetation cover may be require a grid representation, while entities such as roads, rivers, and canals may favor a vector representation. An "area" representation may be most appropriate for lumped-parameter models that may be embedded in a spatial grid, such as a spatially-aggregated lake model that covers multiple grid cells in a landscape. Other objects may be represented as mobile points, such as entities that can wander around in the landscape. These and other spatial data structures should be implemented in the modeling environment, and the details of linking, transferring data between, and decomposing (over multiple processors) spatial representations should be transparent to the modelers.
Building realistic geographic models requires the integration of multiple dynamic modes in a single model. For example, many processes are best represented using differential equations, others are best represented using event-based simulation, and others, such as input-output economic models, use a "black-box" or look-up table implementation. Some processes, such as storm events, are best handled with a hybrid approach. Since continuous (differential-equation based) simulation can be emulated in a discrete event framework (but not vice-versa), the underlying temporal dynamics of the simulation environment should be event based, but structured to efficiently emulate continuous systems.
In an attempt to address the conceptual and computational complexity barriers to geographic model development, our team has been developing an integrated environment for high performance spatial modeling, called the Spatial Modeling Environment (SME). This environment, which transparently links icon-based modeling environments with advanced computing resources, allows users to develop models in a user-friendly, graphical environment, requiring very little knowledge of computers or computer programming. Automatic code generators construct spatial simulations and enable distributed processing over a network of parallel and serial computers, allowing transparent access to state-of-the-art computing facilities. The modeling environment imposes the constraints of modularity and hierarchy in program design, and supports archiving of reusable modules in our Modular Modeling Language (MML). An associated library of "module wrappers" will facilitate the incorporation of legacy simulation models into the environment. This paradigm encourages the development of libraries of modules representing model components that are globally available to model builders, enabling users to build on the work of others instead of starting from scratch each time a new model is initiated. A menu-driven run-time interface will provide the user with a single familiar environment in which to interact with simulations running on any one of a number of parallel or serial computers. This interface will allow users with widely varying goals and background knowledge (from scientists and students to policy makers) to easily control the simulation parameters and generate graphical output (at their Web browser) in a manner appropriate to their level of expertise. The adoption of this paradigm should greatly facilitate the application of computer modeling to the study of spatial systems in support of research, education, and policy making.
The SME design has arisen from the need to support collaborative model development among a large, distributed network of scientists involved in creating a global-scale ecological/economic model. It is intended that it's design be progressively more inclusive of the full range of relevant geographic modeling activities. The following sections give a brief description of the current design of the SME. A more detailed description can be found in the web page (SME). The three-part Modelbase-View-Driver architecture of the SME is displayed in Fig. 1 and described below.
The View component of the SME is an graphical, icon-based simulation environment used to construct, run, calibrate, and test biological/ecological/economic modules in a desktop environment. This component is represented by an off-the-shelf application such as STELLA (HPS) or Extend (Imagine That Inc., San Jose, CA).
The ModuleConstructor application converts the View ecosystem component modules into Module objects defined in our text-based Modular Modeling Language (MML). The MML objects can then be archived in the ModelBase to be accessed by other researchers, and/or used immediately to construct a working spatial simulation. Many MML objects can be combined hierarchically in the MML. This MML hierarchy can then be converted by the Code Generator into a C++ object hierarchy within the spatial modeling environment (SME), where it can drive a spatial simulation.
Figure 1: The ModelBase-View-Driver Architecture.
The PointGrid library (PGL) is a set of C++ distributed objects designed to support computation on irregular, distributed networks and grids. It contains the core set of objects on which the SME Driver is constructed.
The PGL builds spatial representations from sets of Point objects (see below) with links. It transparently handles: 1) creation and decomposition (over processors) of Point Sets, 2) mapping of data over and between Point Sets, 3) Iteration over Point Sets and Point Sub-Sets, 4) data access and update at each Point, and 5) swapping of variable-sized PointSet boundary (ghost) regions. Some of the important PGL classes are:
The driver is a distributed object-oriented simulation environment which incorporates the set of code modules that actually perform the spatial simulation on the targeted platform. It is implemented as a set of distributed C++ objects linked by message passing, layered on top of the PointGrid library.
The current management applications of this framework include the Everglades Landscape Model (ELM), the University of Illinois GMSLab's Threatened and Endangered Species Models (TESM), and the Patuxent Landscape Model (PLM). The SME is also being used for educational purposes at the University of Illinois to support ecosystem modeling classes. These examples demonstrate the use of the SME by interdisciplinary research teams coordinating similar large-scale, ecological modeling efforts. Additional resource management and education programs will be initiated as the system moves out of the development stage.
The Patuxent Landscape Model (PLM) is a regional landscape simulation model that can address the effects of different management and climate scenarios on the ecosystems in the Patuxent Watershed. The PLM is being developed as part of the Ecological Ecosystem Models for Evaluating the Interactive Dynamics of the Patuxent River Watershed and Estuary Project funded by the Chesapeake Bay Research & Monitoring Division, Maryland Department of Natural Resources. The PLM contains about 6,000 spatial cells each containing a dynamic simulation model (based on the GEM model (Fitz et al., 1995)) containing approximately 20 state variables partitioned into 14 modules. It uses two frames, a 2D grid frame (for modules such as Consumers, Nitrogen, Hydrology, Macrophytes, Detritus, etc.) and a tree-network frame for the River modules. The model is being calibrated with data from 1973 and 1985, and run for a scenario analysis period from 1985 to 2020 with selectively variable time steps from hourly to daily depending on forcing function dynamics. Application of this model in the Patuxent watershed is expected to allow extensive analysis of past and future management options, and will form the basis for future application to other areas in the Chesapeake Bay watershed.
The Everglades Landscape Model (ELM) is designed to be one of the principal tools in a systematic analysis of the varying options in managing the distribution of water and nutrients in the Everglades. This system has myriad indirect interactions, constraints and feedbacks that result in complex ecosystem structure (biotic and abiotic components and their flow pathways) and function (the modes of interaction and their rates). For this reason, it is critical to develop a systems viewpoint towards understanding the dynamics inherent in that ecosystem structure and function. Part of this process is the development of a dynamic spatial simulation model. Using this spatially explict process model, changing spatial patterns and processes in the Everglades landscape can be analyzed within the context of altered management strategies. Only by incorporating spatial articulation can an ecological model realistically address large scale management issues within the vast, heterogeneous system of the Everglades.
A unique example of an organization working to understand and manage complex landscape systems is the US military. Having sequestered large tracts of land to function as military training centers, the military is now responsible, as mandated by guidelines in the Endangered Species Act, for managing and preserving the vast collection of ecosystems living on those landscapes. In support of this effort a group of scientists and students at the University of Illinois have developed a number of dynamic spatial ecosystem models designed to help manage and protect a threatened species. The species modeled includes the desert tortoise (Gopherus agassizii), living at Fort Irwin, a U.S. Army training center in the central Mojave Desert of California. The Desert Tortoise model is a collection of four separate sub-models, all constructed within the SME modeling environment. Each model was built by separate groups of the research team members. This division of labor and cross-disciplinary approach allowed individuals to apply their expertise to the most appropriate aspects of the project. The final model is an interdependent web of four distinct aspects of Fort Irwin's landscape: climate, including hydrology and temperature; vegetation; tortoise population growth; and tortoise movement.
We believe that effectively managing human affairs through the next century will require extremely complex and reliable computer models. Widespread utilization of modeling environments supporting graphical, hierarchical/modular design linked to advanced computing resources may be essential in facilitating reliable, economical model construction. General adoption of this paradigm will support the development of libraries of modules representing reusable model components that are globally available to model builders, as well as making advanced computing resources available to users with little computer expertise.
Acock, B. and Reynolds, J.F., 1990. Model Structure and Data Base Development. In: R.K. Dixon, R.S. Meldahl, G.A. Ruark and W.G. Warren (Editors), Process Modeling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, OR.DMSTS, 1995.
Fitz, H.C., DeBellevue, E., Costanza, R., Boumans, R., Maxwell, T. and Wainger, L., 1995. Development of a General Ecosystem Model (GEM) for a range of scales and ecosystems. Ecological Modelling, 88: 263-297.
Gauthier, R.L. and Ponto, S.D., 1970. Designing Systems Programs. Prentice-Hall, Englewood Cliffs, NJ.Goodall, D.W., 1974.
Maxwell, T. and Costanza, R., 1994. Spatial Ecosystem Modeling in a Distributed Computational Environment. In: J. van den Bergh and J. van der Straaten (Editors), Toward Sustainable Development: Concepts, Methods, and Policy. Island Press, Washington, D.C.
Maxwell, T. and Costanza, R., 1995. Distributed Modular Spatial Ecosystem Modelling. International Journal of Computer Simulation: Special Issue on Advanced Simulation Methodologies, 5(3):247-262.
Maxwell, T. and Costanza, R., 1997. An Open Geographic Modeling Environment. Simulation, 68(3):175-185.
Silvert, W., 1993. Object-Oriented Ecosystem Modeling. Ecological Modeling, 68:91-118.