MOSE' (Spatio-Temporal MOdelling of Environmental Evolutionary Processes by means of GeoSErvices) system, a Grid based problem solving environment (PSE) for the developing of geoscience applications. MOSÈ is a PSE able to support the activities that concern the modeling and simulation and mining of spatio-temporal phenomena for analyzing and managing the identification and the mitigation of natural disasters like floods, wildfires, landslides etc. The activities managed by MOSÈ are characterized by the need to handle large amounts of spatio-temporal data and to support the interoperability among simulation models, distributed GIS, visualization systems, parameter estimation services, discovery of spatio-temporal patterns in pre-existing data, etc. In this domain, the data conversion and the access, search, discovery and organization processes are complex problems because data are geo-referenced, stored in distributed GIS and can be used along three dimensions: temporal, spatial and referred to the physical properties.

MOSE' uses a Grid service based computing portal architecture to coordinate the access to the resources. Workflow technology is used to compose the services. The main components of MOSÈ are simulation services, geographic information (GI) services, knowledge discovery services (KDS), visualization services, geographic data and repositories. MOSÈ enables the creation, execution and monitoring of geo-workflows in grid environments through high level, graphical Web interfaces. Components of the workflows can be sequential, parallel and P2P applications. Each component is wrapped as Web/Grid Service for exploiting the potentialities of this architecture. Each Web service is semantically annotated and, consequently, domain specific ontologies support the user in building complex workflows, even without a deep knowledge of the domain itself.

MOSE' provides web based access to the spatial data by a browser and allows data to be observed and manipulated in a 2D/3D space by selecting regions in thematic maps. Natural phenomena can be modeled by cellular automata (CA) models and simulated by a parallel Grid service based on the CAMELotGrid environment. MOSÈ provides KDS based on the WEKA (Waikato Environment for Knowledge Analysis) data mining library and novel distributed data mining algorithms for spatial data analysis. Distributed data intensive mining algorithms are necessary to discover spatial patterns from large geospatial datasets. Novel algorithms must be developed to accomplish this task efficiently. We present an example of innovative KDS based on a bio-inspired P2P agent based algorithm for clustering distributed intensive geospatial data. The algorithm was implemented using the JXTA platform and then wrapped as a Web Service and integrated in the MOSÈ environment.

A first prototype of MOSE', available at the URL http://mose.icar.cnr.it/, was successfully applied for the analysis of landslide hazard areas in the Campania Region near the Sarno area. In this scenario, the main actor is a disaster manager who wants to get an overview of the Sarno area with the indication of the regions which are currently slid down and those which are susceptible to sliding down (landslide hazard areas) within a fixed time. For each scenario, the disaster manager generates a geo-workflow that orchestrates the web services necessary to obtain the outcome, and submitted it to the MOSÈ workflow enactment engine, which takes care of its execution.

References

G. Folino, A. Forestiero, G. Papuzzo, G. Spezzano. MOSE’: A grid portal for solving geoscience problems using distributed knowledge discovery services.
Future Generation Computer Systems, Elsever, Vol. 26, N. 1, pp. 87 – 96 (2010).