On-demand post-processing in modeling systems

Abstract

We present the first results of a project aimed at developing and implementing an on-demand post-processing service based on a nonstop modeling system with the purpose of enhancing and optimizing the usability of simulation results. The on-demand post-processing implementation is part of a general project we called AUTOMET. The project foresees the development of an autonomous system able to run nonstop simulations of pollution dispersion over a certain area by covering all the involved aspects from data collection to internet publication. The on-demand approach has given encouraging results. It allows the user to access the widest range of data processed by a modeling chain from meteorological to diffusive parameters. Whoever accesses the service has the chance to visualize any datum or data set without any knowledge about its implementation or programming language. The interface adopted is user friendly and gives the opportunity, even to non-meteorologists, to query and interpret graphics. Users who may be interested in this information-rendering approach are environmental policy managers, technicians for quick observations, and general public. At present the AUTOMET project is limited to meteorological modeling. The model employed for this purpose is RAMS (Regional Atmospheric Modeling System) but a MM5 implementation for modeling comparisons is planned. The RAMS model has been implemented on a parallel Linux platform following an open-source orientation of the whole project farther then Linux configuration capabilities. A GrADS is employed due to its diffusion and its batch programming features. Other technologies considered by this project are PHP and Javascript.