Presentation Details |
|||||
Name: | (PP23) i_SSS - integrated Support System for Sustainability | ||||
Time: | Tuesday, June 20, 2017 03:15 pm - 03:45 pm |
||||
Room: | Booth #L-212 | ||||
Breaks: | 03:15 pm - 03:45 pm Coffee Break | ||||
Presenter: | Jannek Squar, University of Hamburg | ||||
Abstract: | The target of our project is the development of a realtime digital decision system to optimise farming with regard to an increase of crop yield and minimal damage to the environment. Basis for this decision system is SAGA, which is an open-source geographic information system. During our development we add new SAGA tools, which analyse local circumstances provided through a combination of on-site measurements and remote sensing and allow to assess potential risks. Risks like runoff, erosion and mass transport may be inferred from ICON weather-forecasts, which we get from Deutscher Wetterdienst.
To handle the huge amount of data from weather forecasts, remote sensing and measurement stations but also from SAGA tools, the use of HPC is essential to ensure that the decision system may deliver realtime results in the end. Therefore we build an infrastructure for a HPC cluster which allows to download and preprocess weather-forecasts automatically and in parallel. The preprocessing offers the opportunity to reduce the amount of data (currently up to 90%) and prepare the data before it is loaded into SAGA. SAGA can be executed on a cluster but makes currently only use of OpenMP. Besides from optimising the new tools we want to expand its potential by introducing MPI parallelisation - this would result in a significant gain in performance. Authors: Christoph Beck, Universität Hamburg Michael Bock, Universität Hamburg Jürgen Böhner, Universität Hamburg Olaf Conrad, Universität Hamburg Tobias Kawohl, Universität Hamburg Michael Kuhn, Universität Hamburg Lars Landschreiber, Universität Hamburg Hermann Lenhart, Universität Hamburg Thomas Ludwig, German Climate Computing Center Jannek Squar, Universität Hamburg Sandra Wendland, Universität Hamburg |
||||
Download | PP23_Squar.pdf (1442 KB) |
||||