Torrent hazard control in the european alps. practical tools and methodologies for hazard assessment and risk mitigation

Modelling works were performed and were related to several initiation processes in the identified source areas in the Arbonne catchment were carried out. The software FLAC (Itasca) was used for this purpose. Several potential initiation zones were selected and modelling works were carried out, considering the following scenarios: - Modelling of modifications of geotechnical characteristics of materials (shear strength); - Modelling of the influence of the rising of a water table; - Modelling of the lowering of an infiltration front; - Modelling of erosive processes at the toe of a slope. When conditions for an initiation process are present, the displacement vectors, as well as the development of plasticity indicators account for the mechanism of progressive deformation and localisation of deformations along a particular zone or surface. Scenarios of debris flow initiation processes can then be identified by reference to climatic triggering events.

A monitoring and warning system has been installed On the Wartschenbach. Within the setup of this warning system, from the collected experiences, guidelines for the setup of warning systems were developed. From the recorded data, threshold values were derived for debris flow warning by precipitation. Because of different recording technology it was not possible to find out a special threshold for the other sensors (geophones).

The analysis of the most important impacts related to man activity and resulting in some alteration of the environment responsible for torrential activity modifications, was carried out and can be summarized as follows: A first category of general impacts is related to modifications of the water balance in the catchment or sub-catchment : run-off, infiltration, evapo-transpiration. A second category of general impacts is related to modifications of slope stability conditions. Different elementary impacts can be listed: - Impacts resulting from engineering works by forestry agencies; - Impacts resulting from works related to land farming; - Impacts resulting from road constructions; - Impacts resulting from the rerouting and concentrating of drainage water; - Impacts resulting from the building of winter sport resorts, with the associated (excavations and fill placements which modify the local slope stability conditions, impermeable surfaces: roads, roofs, car parks, tennis courts, usually leaking water mains or sewer lines, deforestation and compaction of superficial soils on the upper slopes for ski practice). The example of La Ravoire de Bourg-Saint-Maurice and the construction of the Arcs winter sport resort was analysed by reference to these impacts. Conclusions can be drawn bu end-users about land management and prevention measures.

Firstly, a standard form devoted to debris flow risk analysis was a reference document. It was designed by ARMINES.M.GI/ENSMP.CGI in a previous research project. The predisposition factors, related to a whole catchment, concern topographical factors, geological and hydrogeological conditions, slope erodibility and slope stability conditions (with the role of vegetation), geotechnical factors (grain size distribution curves of materials in the potential material source areas, looseness of materials, etc), available volume of material liable to be mobilized near a torrent channel or in the torrent bed. During this research project, this methodology was improved for the mapping of the predisposition factors and resulted in a synthetic debris flow initiation hazard map. Field investigations were performed in the Arbonne catchment (Savoie, France). This map allows the identification of material sources areas and then various potential scenarios of debris flow initiation can be analysed, with the associated volumes. These scenarios have to be characterized with reference to a typology of debris flow initiation and mobilization processes. This typology, previously established, refers to 12 main processes. The classification proposed is based on geomorphologic surveys performed in many Alpine debris flow source areas and underlines the importance of local conditions related to geology, geomorphology, hydrology and hydrogeology. These analyses have to be performed at the scale of a whole catchment and may demonstrate that, in some cases, large parts of the catchment contribute to the initiation and mobilization of a given debris flow event during a heavy rainfall, for example, and that, in other cases, only a specific initiation area is concerned. As a consequence of these general results, scenarios of debris flow events can be analysed in a qualitative manner, with the objective of determining orders of magnitude of the following volumes, considering field conditions: - Volumes available in the source areas, - Volumes which are liable to be mobilized into debris flow materials, - Volumes which are liable to reach the lower parts of the catchment.

Measurements of debris flows characteristics and collection of field data related to them are very important for many theoretical and practical purposes Several are the debris flows parameters that have been measured by researchers and technicians. Among the most commonly monitored quantities there are mean flow velocity (U), surface velocity (us), peak flow height (hp), flow height as a function of time, h(t)), peak discharge (Qp), mean discharge as a function of time Q(t)), triggering rainfalls (Tr), ground vibrations (Gv), total volume (V), specific weight or density (f(×) or f(â)) and impact force (If). The monitoring systems are equipped with gauges and sensors for measuring parameters. Video cameras record images of debris flows. The systems are completely automated, remotely controlled and consists of on-site and off-site supervision station.

This result is a better comprehension of the process of body/snout interaction in debris-flows with a qualitative overview of conditions of occurrence or of non-occurrence, how granulars are entrained by the fluid and a quantitative estimation of the effect of the snout on the flow characteristics, mainly depth and velocity. Another result is a better comprehension of the overflow and subsequently build up of levees, with a qualitative overview of the conditions of occurrence and quantitative estimation of the effect on the flow both inside the channel and outside in the initiation zone of spreading. The expected benefit is a better representation of natural flows of debris and subsequently some better hazard assessment.

Hydro-meteorological parameters were analysed in order to identify possible relationship between them and debris flow events in order to contribute ro the elaboration of models for debris flow magnitude-frequency prediction. Different threshold values of rainfall intensity to predict the occurrence of catastrophic debris flows and to identify return periods were tested in the Arbonne catchment. Several types of triggering conditions were identified, as sufficient conditions: - A two-days heavy rainfall (higher than 60mm/2 days), - A two-days storm heavier than the mean storm value + 4* standard deviation storm value (summer period), - A threshold where the 20 days rainfall amount is associated to a following intense 2 days rainfall event, - A criterion where a temperature sudden change (snowmelt period) is associated to a rainfall event. As a matter of fact, it appears that due to the various processes of initiation and mobilization processes, of course the triggering conditions are very different. In a given catchment, where various source areas and the corresponding initiation and mobilization processes exist, several different types of threshold values can operate, being related to the intensity or/and duration of a rainfall event, to the antecedent status of soil before a rainfall event or to snowmelt conditions associated to a particular rainfall event. Up to now the question of models for debris flow magnitude-frequency prediction is improved but not solved.

Reconstructive methods use 'silent witnesses' left by debris-flow events in the past in order to obtain an estimate of hazardous conditions for torrential catchments for which otherwise insufficient preliminary information exist. Methods for geomorphological and sedimentological mapping as well as for dating of deposits are made available for end-users. An important part of the output of this work is the preparation of guidelines for end-users.

Validation of numerical models of spreading of muddy debris-flows by comparison with field data and progress in the definition of domains of validity, model capabilities and deficiencies: The result of this action is a numerical model of spreading of muddy debris-flows that has now been widely tested by comparison with observations of field events. This model has shown good capacities to represent natural events and its domain of validity and level of accuracy are now well defined. Associated to its interface with GIS Arcview, the model can now be used by a large number of field engineers. This will be possible due to the organisation of some common use with field engineers and the organisation of courses. The expected benefit is some improvement of hazard assessment by the use of a powerful tool rather than simply by expertise. The model clearly should be of some help to decision makers.

The GIS-based model developed for hazard assessment at catchment scale is indicated as Probability Model for Torrent Debris Flows (PROMOTOR-df) - it is a user-friendly version based on the application of graphical user interfaces. The model provides first-order probability predictions for different debris-flow magnitudes from an individual torrent, based on end-user specified scenarios for precipitation and using a wide range of catchment properties. PROMOTOR-df is based on the PCRaster Environmental Modelling Language, developed at the Faculty of Geographical Sciences of Utrecht University. As the name suggests, PCRaster is a raster-based Geographical Information System. The architecture of the system permits the integration of environmental modelling functions with classical GIS functions. The meta-language structure of PCRaster allows the direct use of all kinds of existing equations. In this way, users are able to build models in a flexible way. Using this principle, PROMOTOR-df contains groups of process equations for slope hydrology, debris-flow initiation, debris-flow and/or water-flow propagation, and deposition. For end-users, a specially designed interface allows an easy operation of the model both with regard to input and to visualizing the different types of output. The software is freely accessible on registration with the PCRaster managing team. All necessary information, software and instruction to apply the PROMOTOR-df software is brought together in comprehensive guidelines.

Mapping of debris flows in the Municipality of Cortina d'Ampezzo, initiated during the previous EU Project Debris Flow Risk, has been completed at scale 1:10000, through aerial photo analyses and field surveys; morphometric measurements of debris flow catchments have been carried out. A total number of 325 debris flows have been mapped. The geomorphological hazard map has been upgraded and now it includes 135 channelised and 190 hill-slope debris flows. In the data base, each record is identified by a number and it finds its corresponding in the debris flows map. For each debris flow the following parameters have been observed, measured and estimated: - Type (channelised, hill-slope); - Aspect; - Rock-basin area, perimeter, length, width; - Shape factor; maximum, minimum and mean rock-basin elevation; - Relief ratio; - Flow channel length; - Lower flow-channel elevation; - Mean channel slope; - Mean initiation-area slope; - Mean deposition-area slope; -Concentration time (by Giandotti method). This analysis has been extended to the whole Veneto Region where other 120 debris flows have been identified and classified in the same way. The analyses of morphometric data aimed to identify possible relationships between the different parameters. The preliminary results are shown in the following figures as a total of the investigated area and also considering five different sectors in which the same area may be subdivided.

The use of reconstructive methods for past torrent process activity by non-specialist end-users is possible only to a limited degree. All relevant theoretical and practical backgrounds were brought together in a study, the results of which are accessible. Detailed guidelines and instructions enable end-users to apply these methods within safe limits. Apart from technical information, the guidelines contain methodological backgrounds and warnings allowing end-users to verify that an intended use of these methods is scientifically justified and feasible for application by non-specialists. Further, the guidelines also indicate where specialist help is necessary. Interest in former torrent processes may concern the activity at the torrent outlet on the corresponding fan (mainly debris flows) but also processes within the catchment, especially on parts of the slopes where landslides may be intermittently active. The guidelines pay attention to these different processes and to the different sites of occurrence of process activity. In the former case reconstruction contributes more directly to hazard appreciation, in the latter case it may give a better idea of the contribution of certain parts of a torrent to a recognized hazard. The list of topics treated in the guidelines includes: - Description of situations for which a reconstruction is recommended / feasible - Analysis of historical data as a starting point (availability, interpretation, accuracy) - Applying earth-science methods: -- Geomorphological and sedimentological mapping of traces ('silent witnesses'); instruction for mapping and surveying; use of classifications and legends; -- Reconstruction of geometrical properties of individual events; assessment of transport and deposition processes (important for determination of hazard type); -- Dating of identified events; instructions for lichenometry and dendrochronology (the latter strictly limited to certain cases without expert help); - Interpretation of results in terms of temporal activity pattern, magnitude/frequency relationship, and from the latter: probability estimates for events of given magnitude.

The result consists in some interface between a numerical model of debris-flow spreading and the very classical GIS environment Arcview. In that sense, it makes possible the use of the very sophisticated hydraulic tool to a large number of engineers and technicians in charge of the hazard assessment. The benefit is a user-friendly access to tools, which up to now could be used only by researchers.

Firstly, a standard form devoted to debris flow risk analysis was a reference document. It was designed by ARMINES.M.GI/ENSMP.CGI in a previous research project. The identification of predisposition factors was improved using GIS tools. The methodology was improved for the mapping of the predisposition factors and resulted in a synthetic debris flow initiation hazard map, with an application in the Arbonne catchment (Savoie, France). This map allows the identification of material sources areas and then various potential scenarios of debris flow initiation can be analysed, with the associated volumes. These scenarios have to be characterized with reference to a typology of debris flow initiation and mobilization processes. This typology, previously established, refers to 12 main processes. The classification proposed is based on geomorphologic surveys performed in many Alpine debris flow source areas and underlines the importance of local conditions related to geology, geomorphology, hydrology and hydrogeology. As a consequence of these general results, scenarios of debris flow events can be analysed in a qualitative manner, with the objective of determining orders of magnitude of the following volumes, considering field conditions: volumes available in the source areas, volumes which are liable to be mobilized into debris flow materials, volumes which are liable to reach the lower parts of the catchment.

Collection of field data regarding debris flows is important for several scientific and practical purposes. However, debris flows generally occur with a frequency high enough to create serious hazards to human settlements, but too low for a convenient monitoring activity capable of providing enough data in a sufficiently short time. Moreover it is also difficult to locate mountain torrents where monitoring equipment may be easily and safely installed. In 1984 a small creek was located in the Eastern Italian Alps (the Moscardo Torrent) that presented both an easy accessibility and a very high debris flows occurrence. This catchment was thus chosen for the installation of a debris flow monitoring system for the collection of field-data (hydrographs, seismographs, rainfalls) useful for the entire scientific community and for the understanding of the phenomenon. Within the THARMIT project the monitoring system has been renewed and has collected three years of field data.

Within the scope of the Tharmit project, guidelines were prepared for GIS-based modelling of torrent hazard prediction to facilitate successful application by end-users. Especially for GIS-based modelling, Internet availability is essential. Both the GIS-based model for hazard assessment at catchment scale indicated as Probability Model for Torrent Debris Flows (PROMOTOR-df) and the software developed are described under separate Results. The end-user operates PROMOTOR-df from an interface showing several sheets for manipulating input data en displaying the different types of output. Several map and table options are available. For all items, default values are given, but end-user specified values could be introduced in all cases to match catchment-specific conditions. Comprehensive interactive guidelines for end-users of the Probability Model for Torrent Debris Flows (PROMOTOR-df) are developed, which can be reached and operated from a special PROMOTOR-df website. This site also allows direct access to all needed software, PCRaster included. The guidelines include introductory chapters, describing theoretical issues and delimiting the application fields of the model. These chapters are followed by an extensive demonstration, giving examples of how to use the model, its interface, how to display and interpret results. A further chapter leads the end-user through the running of the model using own data. Finally, several appendices supply background information and background data.

Progress in the comprehension of debris-flow / structure interaction, mainly development of a new 3D numerical model able compute transitory flows close to the structure and validation in the laboratory. The main result is a new numerical model based upon the smoothed particles hydrodynamics technique and which has been developed for the purpose of studying debris-flow / structures interactions. This 3D unsteady model makes it possible to compute flow close to structures and give a precise description of its characteristics in one hand and of the pressures applied by the flow to the structure in the other hand. This model has been partly validated by comparison to laboratory experiments. The expected benefit is to be able after further research, to precisely compute the pressures applied to structures while for now, only very rough approaches exist. The expected benefit is new methods and rules for the design of check dams for example.

In the context of the examinations was found, that reasons for insufficient function fulfilment results from one for the structures unsuitable function, in the woody debris difficulties, in the geological qualities of the debris as well as in an inadequate discharge. The result of the model experiments has been a strong coherence between the Froude-number and the maximum pressure peak occurring at the barrier. The pressure depends very much on the material properties, a significant difference was found between fast and liquid and slow and stony debris flows.

One of the main aspects in slope instability and debris-flow research is the capability to store, manipulate and analyse spatial and temporal data. In this connection and with respect to the application of new data collection technologies, computer tools should be considered necessary to guarantee effective linkages between data and models; but, despite encouraging progresses in computer technologies applications, the great potential of these tools is largely untested. Moreover, an increasing demand exists for a more and more easy access to natural phenomena data, as they can be used both to investigate their temporal pattern (time series analysis) and to consider their spatial pattern in terms of regional hazard assessment. The primary task is then to use temporal and spatial inventories of debris flow events and related information for the elaboration of hazard and risk models and for the analysis of events time series in relation to triggering factors. The original aim of the database is to collect data mainly concerning: - Type of natural hazard; - Topographic, geological and geomorphological maps of catchments and deposition areas; - History and chronology of events; - Predisposition and triggering factors description; - Physical, mechanical and hydrogeological characteristics of soils and rocks; - Event magnitude and frequency; - Characteristics and data of the existing observation stations; - Available results of tests on samples for the estimate of physical, mechanical and rheological parameters.

Method for geomorphometric analysis of mountain catchments (The Horton machine): evaluation of the zones prone to landslides and debris flow; evaluation of the liquid and solid discharge under assigned design conditions, and debris flow run-out estimation (in prototype). The developed software and the guidelines can serve for mapping idrogeological hazards starting from DEMs of the catchments and some (usually available) meteorological data. The tools are intended to be used by engineers, geologists and forest scientists (but require however some training). The maps, in turn, can be used as a base for sustainable urban and regional planning.

Method for GIS-based torrential hazard prediction. Using PCRaster Environmental Modelling Language, combining map-based geo information, time series, and process equations to describe processes of initiation, propagation and deposition involved in debris-flow activity (model PROMOTOR-df). The GIS-based model developed for hazard assessment at catchment scale is indicated as Probability Model for Torrent Debris Flows (PROMOTOR-df). The model provides first-order probability predictions for different debris-flow magnitudes from an individual torrent, based on end-user specified scenarios for precipitation and using a wide range of catchment properties. The scenarios simulate torrent behaviour during the full extent of the event and for all parts of the torrent involved, for which reason the modelling is indicated as 'dynamic'. The central concepts of the model combines georeferenced information from the torrent to be analysed (DTM, catchment properties), empirical equations describing torrent slope and channel processes, and detailed rainfall data (intensity and duration of precipitation). The 'dynamic' modelling applied in PROMOTOR-df is based on the PCRaster Environmental Modelling Language, developed at the Faculty of Geographical Sciences of Utrecht University. The model consists of two related parts, one for torrent catchment behaviour, and one for outflow of debris flow and/or water flow on the fan below the catchment outlet. At present, the catchment part is completed, while the fan part will be completed in the near future. The catchment part is, however, relevant as such because it predicts the probability of hazardous debris flows leaving the catchment for a given scenario, specified by the end-user. Potential end-users are managers of infrastructure, administration officers at local or regional level, etc. Comprehensive guidelines and instructions have been developed for end-users.

New approach of field assessment of the magnitude/frequency relations on the basis of the analysis of environmental factors and triggering mechanisms: The result is a new geomorphological approach dedicated to the assessment of magnitude/frequency relation for debris-flows and of characterization of the main features (essentially) of catastrophic events of debris-flows. The estimation of the magnitude of debris-flows is of paramount importance because it is the input to any other possible computation of the characteristics of the potential flow. The expected benefit is a better estimation of these input data leading in fine to better hazard assessment.

A barrier typology, based on the division between Solid Body Barriers and Open Barriers was developed. The Open Barriers are further described by a 2-part nomenclature in which the first part describes the opening and the second part describes the elements to cover the openings. By the use of this two-folded classification, torrential barriers can be exactly defined, which makes the classification useable for databases.

Deliverables

Publications

Share this page

Download