Data Management System GEORIOS – Documentation and evaluation of natural hazards

Natural hazards and their effects on the population and economy are for the Austrian Alpine region of increasing relevance. The events of the last years have shown that high meaning is to be attached to a comprehensive scientific understanding of mass movement processes. The quest for security gives rise to develop strategies and measures to counter the threats and to protect the people and the infrastructure. To determine where protective measures are necessary, we produce landslide inventory and risk assessment maps for many areas in Austria.
   Since the foundation of the Geological Survey of Austria (GBA) in 1849 have been received a lot of data or knowledge about geogen natural hazard into the archives of the department of engineering geology.
   To manage this data diversity and to make this entire pool of data available for everyone is necessary to develop a row of policies and strategies. Three main steps are followed in order to create this management system: (i) the development of spatial database, (ii) the development of an integrated procedure for design of susceptibility maps, (iii) and development of a tool set for the visualization and web-enabled data query. The final application based on the concept of Landslide Information Systems, will be used as an additional tool for risk and emergency assessment as well as for planning and decision making purposes.
   Landslides unfortunately, do not display a clear relationship between magnitude and frequency as do for example floods. Landslide studies are challenging to scientists, due to the difficulty to represent landslide hazards in quantitative terms over large areas.
   To be able to clarify which method to which conditions (scale of area, quality of data, area heterogeneity) and for which questions /objectives is suitable in the different measure, a classification of the areas to be modelled is necessary. However, this also means that the modelling results must be judged concerning her statement quality for different objectives (at least semi quantitatively). Otherwise it would be unclear furthermore for what, the produced maps (e.g., hazard potential maps, susceptibility maps) by means of different methods and data quality, are to be used generally.
   Experience has shown that in this regard the following criteria should be used:
   - Relative size of area
   - Data quality (the quality of process data, the quality of the parameter maps)
   Nevertheless, all these criteria cannot be quantified. Therefore, different areas with regional variety and different data quality are very important for model calculations. Only different models and methods, can be tested concerning her usefulness for the production of maps as bases for spatial planning, and estimated, under which conditions which method is for which question more suitable
   In addition, the GBA is also keen to apply methods and to develop strategies, through which an evaluation of existing data towards large scale maps (for example hazard potential maps).
   For one project area were used in addition to the simple heuristic method also a neural networks method to produce a susceptibility map. From the visual comparisons of the results can actually see any big differences. A comparison with the results of other methods shows only small differences exist, however, all validation results using artificial neural networks are slightly better than those using heuristic method.
   However due to the random and selective available process data, there is a risk, that the generated susceptibility maps lead to good validation results. This means that the causes of the well-validated results are not clear and further developments are necessary.