|The MUTATE project is funded by the Educational Multimedia Task Force of the European Union.|
MUTATE environmental modeling tools developed and implemented by ESS are based on a generic client-server architecture that combines a powerful model server for high-performance computations required for interactive modeling with the flexibility of a Java based user interface.
The models are run on a (conceptual) model server, that is linked to the main MUTATE http server either on a local area network (LAN) or on the same (multi-tasking) UNIX machine; a bandwidth of 100 Mb/s (at least 10 Mb/s) provided by fast Ethernet is recommended.
The client software (Java applets triggered within a standard web browser) is platform independent and can be run on any client hardware such as PCs or workstations, including light-weight low-end PCs.
The client software performs two major groups of interactive interface functions:
For the integration of the MUTATE Model Server into the courses of the MUTATE Bundle, several communication and integration mechanisms are foreseen:
individual calls to the model server from XML/HTML pages;
a complete set of scenario specifications is sent to
the Model Server cgi (Post request);
the model server returns its results in graphical format (GIF, PNG); this may consist of the model results (in different graphical renderings) with or without background maps, or possibly specific graphs derived from the primary model results;
output files deposited by the model server in a standard directory; the primary output of the simulation models are cell-grid files with scalar or vector data;
media player a separate interface that provides access to all the model functions; here the caller can specify a parameter mask that defines a (sub)set of parameters for user editing. The media player provides its own interface for the display of model results, and has an (optional) data export function (see above).
Check the detailed description of communication protocol and file formtas for the client-server communication.
From a didactic perspective, the models are designed for scenario analysis or experimentation, where each scenario or experiment in turn should be designed to illustrate one or several basic concepts within the framework of a course module.
The scenario here is defined as a complete set of initial and boundary conditions for a given model, plus the results of running the model with these inputs.
For a given exercise or illustration, most of these parameters will be fixed at predefined values, but a few will be open (within well defined ranges) to be changed by the student interactively to explore the concept of the lesson.
The underlying experiments a student can perform will therefor be of one of two types:
A typical example could be the relationship between (atmospheric) dispersion parameters and the spatial distribution of pollutant concentration from one or several emission sources.
A typical example could be a (spatial) environmental quality standard as the target, and emission reduction options as the decision variables.
In the latter case, the student's experimental approach could also be checked against an optimization routine.
In the example case of the Air Quality simulation system AirWare, a standard simulation scenario for major industrial point sources and for the simplest possible (steady-state) models, consists of the following elements:
The model scenario involves the selection and definition of the following items:
choice of model (ISC-2, ISC-3 (with or without COMPLEX domain corrections; please note that the domain corrections require that a DEM is available for the model domain)
choice of simulation period (please note that this is related to the choice of models, since not all models are available in both a short-term (episode or event) and long-term (seasonal, yearly) version.
choice of spatial resolution (grid size).
The emission scenario involves the selection of one of the pre-defined emission inventories, and, on this basis, the definition of a set of emission sources. This includes:
The content of a weather scenario depends on the choice of model (see above) and the temporal resolution and duration of the simulation.
For a short-term (episode) model run, typically for a one-hour averaging period:
For a long-term seasonal or annual model run, frequency distributions for the above variables are required. In AirWare, they are generated automatically given the selection of a meteorological station, and the definition of the start- and end-dates of the desired simulation period.