SOCIETY | Aug 23, 2018

Data and digital against avalanches and rockfalls

Here is how the RED project algorithms help monitor hydrogeological risk

There is no doubt about the fragility of the Italian territory, and with current climate changes the situation is worsening. This is confirmed by the most recent data: in 2017, 91% of Italian municipalities were at risk and more than 7 million people lived in areas of high vulnerability; the area potentially subject to landslides grew by 2.9% compared with 2015 while the area exposed to risk of flooding increased by 4%. The Risk Evaluation Dashboard (RED) research unit was created precisely to provide an innovative tool for supporting local bodies in charge of protecting the territory.

How did the project aimed at limiting hydrogeological risk come into being?

The RED project was started in July 2015 in response to a call for proposals for the creation and development of research units organized by the Valle d’Aosta Autonomous Region and financed through the European Regional Development Fund. It is made up of Engineering as lead partner, two important research institutes, the Polytechnic University of Turin and the Montagna Sicura Foundation, the GeneGIS GI company which is specialized in environmental surveys, and GMH helicopter services.

The main aim of the research unit is to design and implement a dashboard that enables efficient management of various data related to hydrogeological risk. The dashboard has been developed to address two particular natural hazards: snow avalanches and rockfalls. For avalanches, an algorithm has been created which, on the basis of collected or predicted nivo-meteorological data, is able to trace the most similar days that occurred in the past and call up the associated avalanche risk situation. For rockfalls, on the other hand, some algorithms have been implemented for assessing the susceptibility of and risk associated with road infrastructures: these infrastructures are in fact those that present the major problems related to aspects of rockfalls.


The data are there, but …

In order to provide a sufficiently robust and reliable tool, a genetic algorithm for evaluating the weights to be attributed to the numerous variables that come into play for the forecasting of avalanches has been studied and implemented. This algorithm requires a large amount of data which, fortunately, the Valle d’Aosta Region possesses. In fact, avalanche and nivo-meteorological parameters have been collected in the region with a certain continuity for over forty years. However, the databases created at the time and still in operation were designed for mainly consultative use, leaving ample space to qualitative information collected by various operators. This information, although essential, cannot be managed correctly by current IT tools: most of the project time was therefore dedicated to the “preparation” of the mass of data contained in the various databases.

A similar argument can be made for the databases relating to rock collapse phenomena. For this reason, the dashboard has been designed to allow for the systematic and standardized collection of geological, geotechnical and structural information. In addition, data types that have not yet been identified (such as those relating to the state of deterioration of defense structures) have been identified, and specific tools have been developed to guide operators during the information-gathering phase in the field.

What is the project’s biggest challenge?

Beyond the technical and theoretical aspects, the great and compelling challenge was to work within a heterogeneous and strongly multidisciplinary group. The goal was common to all, but being able to break free from complex theoretical concepts, the technical aspects of computerization and the concrete needs of operators was not easy.  To communicate effectively with all the different project players it was necessary to abandon some preconceptions, strengthen the capacity of abstraction and reshape a mindset that had necessarily formed after years of study in the engineering field. formed after years of study in the engineering field.

Finding the points of contact and creating effective channels of communication took time because each of us basically speaks (and reasons) through a sectoral language. Understanding one another is easy among computer scientists, engineers or technicians, but being able to translate theoretical-engineering complexity into blocks of computer code, all while respecting the real and practical needs of the technicians, has been as difficult as it has been exciting.

What developments for the future?

The need to put yourself out there for creating an innovative and functional product has created fertile ground for effective development of the ideas contained in the project. Although it sanctions the end of the RED project, realization of the dashboard prototype is actually the starting point for the development of other ideas and other projects.

Use of the product and acquisition of new territorial data through the standardized collection procedures implemented in the dashboard will make it possible to compare and above all validate the quality of the algorithms implemented. The data entered will in turn change the algorithms themselves, making them progressively more reliable. The infrastructure created for storage and display of collected data can also be applied, with appropriate modifications, to other hydrogeological phenomena such as landslides, floods and forest fires. Finally, with the increase of data collected inside the databases, it will be possible to search for connections and correlations among different types of data, in order to open the way to new theoretical research, transforming the collected data into useful information managed by the bodies charged with defense of the territory.

André Chaussod