ALVIOLI M., MARCHESINI I., ROSSI M., SANTANGELO M., CARDINALI M., REICHENBACH P., ARDIZZONE F., FIORUCCI F., BALDUCCI V., MONDINI A.C., GUZZETTI F., 2013, Parallel processing in WPS services for geological and geomorphological mapping, International conference (AIG) on Geomorphology 2013, Parigi, 2013,
URL: http://www.cnr.it/prodotto/i/273234

Exploiting to the maximum possible extent the information contained in spatial data, along with specialized and optimized computing intensive analysis processes, is mandatory in order to cope with the challenges posed by the huge amount of existing and upcoming data. Dealing with spatial data and related computing processes can be a hard task for many subjects, preventing proper dissemination of information and modeling advances. One possibility for sharing data content, specialized model implementation and distributed computing power is represented by Web Processing Services, an Open Geospatial Consortium standard specifically designed to facilitate the inter-operability of geospatial data by client-server interfaces. The client side of the interface presents the user with a choice of services exposed by the server, and the possibility of specifying parameter values and uploading own maps. The main advantages of such an approach are the possibility of publishing specialized and optimized processes, corresponding to sound and well- established models; the complexity of the implemented processes is hidden by the server side, where a number of chained procedures can be summarized in a single service, optionally using different programming languages and, most importantly, exploiting distributed and parallel computing. We have implemented tools for (i) statistical models of the distribution of landslide areas (ii) estimation of the attitude of bedding planes from the corresponding bedding traces (iii) production of maps showing the geometrical relationship between bedding planes and terrain slopes (iv) automatic delineation of hydrological slope units. These few prototype services for landslide modeling, representative of a vast class of processes from the algorithmic implementation point of view and the parallelization schemes we have used, represent a starting point for a systematic publication of models through similar web services.

Data from https://intranet.cnr.it/people/