The Modular Modeling System (MMS) - A Modeling Framework for Multidisciplinary Research and Operational Applications

Table of Contents

To address the problems of model selection, application, and analysis, a set of modular modeling tools, termed the Modular Modeling System (MMS) is being developed by the NRP Precipitation-Runoff Modeling Project. The approach being applied in developing MMS is to enable a user to selectively couple the most appropriate process algorithms from applicable models to create an "optimal" model for the desired application. Where existing algorithms are not appropriate, new algorithms can be developed and easily added to the system. This modular approach to model development and application provides a flexible method for identifying the most appropriate modeling approaches given a specific set of user needs and constraints.

MMS OVERVIEW
The conceptual framework for MMS has three major components: pre-process, model, and post-process . A system supervisor, in the form of an X-window graphical user interface (GUI), is proposed to provide user access to all the components and features of MMS. The present framework has been developed for UNIX-based workstations and uses X-windows and Motif for the GUI. The GUI provides an interactive environment for users to access model-component features, apply selected options, and graphically display simulation and analysis results. The current GUI is being expanded and enhanced into the full system supervisor, incorporating the linkages needed to access features in all the system components.

Pre-process Component
The pre-process component includes the tools used to input, analyze, and prepare spatial and time-series data for use in model applications. A goal in the development of the pre-process component is to take advantage of the wide variety of existing data- preparation and analysis tools and to provide the ability to add new tools as they become available. Spatial data analysis is accomplished using geographic information system (GIS) tools that users have installed on their computer system.

The data base(s) used to store the spatial and time series data provide(s) the interface between the pre-process and model components. Time-series data from existing data bases as well as from field instrumentation currently are prepared for use in selected model applications by generating and combining these data into a single flat ascii file. Procedures are being developed to enable the interface of a variety of commercial and user-definded data bases directly with a model. The goal is to enable the use of a variety of data bases, dependent on user preference or prescribed needs.

Model Component
The model component is the core of the system and includes the tools to selectively link process modules from the module library to build a model and to interact with this model to perform a variety of simulation and analysis tasks. The module library contains a variety of compatible modules for simulating water, energy, and biogeochemical processes. Several modules for a given process may be present, each representing an alternative conceptualization or approach to simulating that process. A module can be written in either the FORTRAN or C programming language.

Modules are located in both read-only directories, where tested, documented, and approved code reside, and in user-defined work directories where new modules are being developed. The user selects and links modules from these directories to create a specific model using an interactive, graphical, model-builder tool (MBUILD). Modules are linked by coupling the outputs of user-selected modules to the required inputs of other user-selected modules. Tools are provided to display a module's input requirements and to list all modules available that will satisfy each of these inputs. When the inputs for all modules are satisfied, a model is complete. Once a model has been built, it may be saved for future use without repeating the MBUILD step.

When a model is executed, the user is interfaced with the model through a series of pull-down menus in the GUI which provide the links to a variety of system features. These include the ability to (1) select and edit parameter files and data files, (2) select a number of model execution options such as a basic run, an optimization run, or a sensitivity analysis run, and (3) select a variety of statistical and graphical analyses of simulation output. During a basic run, up to four graphical display windows can be opened to display any of the variables that have been declared in the model modules. As many as 10 variables can be displayed in each window and plotted results can be output in HPGL or PostScript formats either to a digital file or to a printer.

Post-process Component
The post-process component provides a number of tools to display and analyze model results, and to pass results to management models or other types of software. One tool included in this component is the copyrighted freeware package XMGR. XMGR can be used to visualize selected model variables in user-defined combinations and to analyze these variables using a number of statistical procedures. Model output can also be directed to user-specific analysis programs using an ascii flat-file format.

Some post-processing capabilities interact directly with the model component. The parameter-optimization and sensitivity-analysis tools are provided to optimize selected model parameters and evaluate the extent to which uncertainty in model parameters affects uncertainty in simulation results. A modified version of the National Weather Service's Extended Streamflow Prediction Program (ESP) (Day, 1985) provides forecasting capabilities using historic or synthesized meteorological data.

GIS INTERFACE
A geographic information system (GIS) interface is being developed to provide tools for the analysis and manipulation of spatial data in the pre-process, model, and post- process components of MMS. Pre- and post-processing interfaces are being developed as generic interfaces to support a variety of existing GIS packages. To date, GIS tools have been developed and tested using Arc/Info and the Geographical Resources Analysis Support System (GRASS) (U.S. Army, 1991). Functions developed include the ability to (1) delineate and characterize watershed subbasin areas for distributed-parameter modeling applications, (2) estimate selected model parameters for these subbasins using digital elevation model (DEM) data and digital data bases that include information on soils, vegetation, geology, and other pertinent physical features, (3) generate an MMS input parameter file from these estimates, and (4) display and analyze the spatial and temporal variation of selected model results.

Within the model component, the GIS interface enables the visualization of the spatial and temporal variation of simulated state variables during a model run. The capability to store selected images from this animation for user-defined time periods is currently being developed.

COLLABORATIVE DEVELOPMENT
MMS began as a cooperative research effort between the U.S. Geological Survey (USGS) and the University of Colorado's Center for Advanced Decision Support for Water and Environmental Systems (CADSWES). As MMS took shape, interest in the MMS concepts was expressed by many national and international agencies and organizations. Agreements established with several of these groups have provided new ideas for system enhancement and the contribution of resources, in terms of money and/or people, to add these enhancements to the system. In addition, these groups continue to contribute their modeling expertise to the system by converting their models to MMS modules and by providing test sites for system evaluation and development.

The U.S. Bureau of Reclamation (BOR) has aided in the development of some of the GIS capabilitiies and the ESP tools. They have coupled MMS to a river basin management model and are currently using the system as a real-time management tool on the Gunnison River basin in Colorado. River basin management tools coupled with power generation management tools is also the focus of a joint project between the Tennessee Valley Authority (TVA) and the Electrical Power Research Institute (EPRI). MMS was selected to provide the hydrologic and ecological modeling support to this project and EPRI is funding five post doctoral positions in the NRP to support the research and develop of components for a noncalibrated hydrologic model.

The TERRA Laboratory is a joint Agricultural Research Service (ARS)-Forest Service (FS)-USGS consortium that was formed to facilitate the development of decision support systems for terrestrial ecosystem problems. TERRA has provided several systems programmers to work on MMS components and has provided funding to a number of research projects to incorporate their models in MMS. One such project is funding the placement of the ARS Root Zone Water Quality Model (RZWQM) in the modular library and a joint USGS-ARS research project to convert RZWQM from a point- to a watershed-scale model. RZWQM will be used to study the movement of nutrients and pesticides under different agricultural practices at several of the Mid- Continent Initiative sites being studied under the USGS Toxics program.

The FS is also using MMS to integrate a number of watershed and forest-growth modeling capabilities to enable the assessment of the effects of different land- mangement decisions on hydrologic and ecosystem response. NASA is using MMS to investigate the development of coupled atmospheric and hydrologic models and the coupling of remotely sensed data with hydrologic models.

Internationally, the German Research Association, the equivalent of the U.S. National Science Foundation, is supporting the development and application of MMS on several projects in Germany. These include a 15-University consortium project titled the Regionalization of Hydrolgy in Germany, and projects studying the effects of climate and land-use change on watershed response at the University of Bonn and the Potsdam Institute for Climate Impact Analysis.

Initial USGS contributions to the module library are from the U.S. Geological Survey's Precipitation Runoff Modeling System (PRMS) (Leavesley et al., 1983) and TOPMODEL (Beven and Kirkby, 1979; Wolock, 1993). New modules for channel transport of solutes and sediment also have been developed and are currently being included. Additional modules resulting from continued collaborative efforts within the WRD and among other government agencies and univerities will be added as they are developed, tested, and documented.

SUMMARY
MMS is an integrated system of computer software that has been developed to provide the research and operational framework needed to support the development, testing, and evaluation of physical-process algorithms and to facilitate the integration of user- selected sets of algorithms into an operational model. MMS provides a common framework in which to focus multidisciplinary research and operational efforts. Scientists in a variety of disciplines can develop and test model components to investigate questions in their own areas of expertise as well as work cooperatively on multidisciplinary problems without each scientist having to develop the complete system model. Results that demonstrate improved simulation performance can be used to modify or enhance current operational models for application within the same framework.

Continued advances in physical and biological sciences, GIS technology, computer technology, and data resources will expand the need for a dynamic set of tools to incorporate these advances in a wide range of interdisciplinary research and operational applications. MMS is being developed as a flexible framework in which to integrate these activities.

REFERENCES
Beven, K.J., and Kirkby, M.J., 1979, A physically based variable contributing area model of basin hydrology: Hydrological Sciences Bulletin, v. 24, p. 43-69.
Day, G.N., 1985, Extended streamflow forecasting using NWSRFS: Journal of Water Resources Planning and Management, American Society of Civil Engineers, 111(2), p. 157-170.
Leavesley, G.H., Lichty, R.W., Troutman, B.M., and Saindon, L.G., 1983, Precipitation- runoff modeling system: User's manual: U.S. Geological Survey Water Resources Investigations Report 83-4238, 207 p.
U.S. Army Corps of Engineers, 1991, Grass version 4.0 user's reference manual, USACERL, Champagne, IL, 513 p.
Wolock, D.M., 1993, Simulating the variable-source-area concept of streamflow generation with the watershed model TOPMODEL: U.S. Geological Survey Water Resources Investigations Report 93-4124, 33 p.

U.S. Department of the Interior | U.S. Geological Survey
URL: http://wwwbrr.cr.usgs.gov/projects/SW_MoWS/software/oui_and_mms_s/mmsAbstract.shtml
Page Contact Information: <>
Page Last Modified: October 20, 2006