Search

Analog methods (AMs) allow predicting local meteorological variables of interest (predictand), such as the daily precipitation, based on synoptic variables (predictors). They rely on the hypothesis that similar atmospheric conditions are likely to result in similar local effects. The statistical relationship is first defined (e.g. which predictors, and how many subsampling steps) and calibrated (e.g. which spatial domain, and how many analogues) before being applied to the target period, may it be for operational forecasting or for climate impact studies. A benefit of AMs is that they are lightweight and can provide valuable results for a negligible cost. AtmoSwing is an open source software that implements different AM variants in a very flexible way, so that they can be easily configured by means of XML files. It is written in C++, is object-oriented and multi-platform. AtmoSwing provides four tools: the Optimizer to establish the relationship between the predictand and predictors, the Downscaler to apply the method for climate impact studies, the Forecaster to perform operational forecasts, and the Viewer to display the results. The Optimizer provides a semi-automatic sequential approach, as well as Monte-Carlo analyses, and a global optimization technique by means of Genetic Algorithms. It calibrates the statistical relationship that can be later applied in a forecasting or climatic context. The Downscaler takes as input the outputs of climate models, either GCMs or RCMs in order to provide a downscaled time series of the predictand of interest at a local scale. The Forecaster automatically downloads and reads operational NWP outputs to provide operational forecasting of the predictand of interest. The processing of a forecast is extremely lightweight in terms of computing resources; it can indeed run on almost any computer. The Viewer displays the forecasts in an interactive GIS environment. It contains several layers of syntheses and details in order to provide a quick overview of the potential critical situations in the coming days, as well as the possibility for the user to go into the details of the forecasted predictand distribution.

Hydrological Risk: Landslide

Analogue methods (AMs) rely on the hypothesis that similar situations, in terms of atmospheric circulation, are likely to result in similar local or regional weather conditions. These methods consist of sampling a certain number of past situations, based on different synoptic-scale meteorological variables (predictors), in order to construct a probabilistic prediction for a local weather variable of interest (predictand). They are often used for daily precipitation prediction, either in the context of real-time forecasting, reconstruction of past weather conditions, or future climate impact studies. The relationship between predictors and predictands is defined by several parameters (predictor variable, spatial and temporal windows used for the comparison, analogy criteria, and number of analogues), which are often calibrated by means of a semi-automatic sequential procedure that has strong limitations. AMs may include several subsampling levels (e.g. first sorting a set of analogs in terms of circulation, then restricting to those with similar moisture status). The parameter space of the AMs can be very complex, with substantial co-dependencies between the parameters. Thus, global optimization techniques are likely to be necessary for calibrating most AM variants, as they can optimize all parameters of all analogy levels simultaneously. Genetic algorithms (GAs) were found to be successful in finding optimal values of AM parameters. They allow taking into account parameters inter-dependencies, and selecting objectively some parameters that were manually selected beforehand (such as the pressure levels and the temporal windows of the predictor variables), and thus obviate the need of assessing a high number of combinations. The performance scores of the optimized methods increased compared to reference methods, and this even to a greater extent for days with high precipitation totals. The resulting parameters were found to be relevant and spatially coherent. Moreover, they were obtained automatically and objectively, which reduces efforts invested in exploration attempts when adapting the method to a new region or for a new predictand. In addition, the approach allowed for new degrees of freedom, such as a weighting between the pressure levels, and non overlapping spatial windows. Genetic algorithms were then used further in order to automatically select predictor variables and analogy criteria. This resulted in interesting outputs, providing new predictor-criterion combinations. However, some limitations of the approach were encountered, and the need of the expert input is likely to remain necessary. Nevertheless, letting GAs exploring a dataset for the best predictor for a predictand of interest is certainly a useful tool, particularly when applied for a new predictand or a new region under different climatic characteristics.

International audience

The Séchilienne landslide is located on the right bank of the Romanche River, South East of Grenoble (Isère, France). The active zone of the gravitational instability involves several millions of cubic meters. The geology consists in fractured hard rocks (micaschists) with double permeability and strong spatial heterogeneities. The deformation of the unstable slope is monitored by on-site extensometric gauges, inclinometers, GNSS and remotely by a terrestrial radar and a total station.Hydro-chemio-mechanical processes controlling the reactivation of the landslide are influenced by the evolution of the landslide deformation in space and time, and the water circulation in the highly heterogeneous fractured media.A hydrogeochemical monitoring of the unsaturated zone in the fractured hard rock has been carried out since 2010. This monitoring is supported by the French Landslide observatory (OMIV) and consists in continuous measurements of physico-chemical parameters on two groundwater outlets (T◦C, EC, flow rate) and weekly samplings of the waters for quality monitoring. Water chemistry is a good proxy to locate in time and spacethe origin of the infiltrated water. This tool is used to understand the complex relationships between chemical weathering, hydromechanical changes and weakening/deformation of the unstable material. This monitoring indicates a correlation between water chemistry, rainwater infiltration and rock mass deformation highlighting the impacts of rock-water interactions on the landslide dynamics. But a distributed information over area is still needed because the heterogeneities of the slope and the few sampling points currently prevent a detailed understanding of the global mechanisms involved.To better understand and constrain the hydrogeological and hydro-chemio-mechanical behavior of the slope, a multi-method monitoring of a flood wave infiltration has been carried out in early 2016 in order to distinguish possible signals related to significant displacements. Displacements were monitored by a GB-InSAR and a terrestrial laser scanner in order to obtain a global image of the deformation at high frequency (less than 1 hour). Repeated time-lapse geoelectrical profiles along four sections have been acquired each two hours on relevant plots which are suspected to be the main water flow paths from the surface to the depth. Water quality changes were monitored at high frequency in order to provide information on the water residence time. This first dataset gives insight into the moving volumes of rock and fluids. Imagery geophysics identifies a signal of fluid circulation in a fracture with a fast transit. The chemical signal identifies the heterogeneous functioning of the drainage system (drain/low permeable structure) with a fast transit.

International audience; The knowledge of the rockfall frequency in a cliff is needed for quantitative rockfall hazard assessment (Hantz et al., 2016). It can be estimated from historical data bases (e.g. Hantz et al., 2003) or from diachronic comparison of digital cliff topographic models (e.g. Dewez et al. 2013, Guerin et al. 2014). An empirical approach is proposed to be used when historical data bases are not significant (not enough events) and diachronic digital models not available.

Economic estimation of the impact of rural earthen road construction in Phewa watershed. (DOCX 20Â kb)

The Analogue Method (AM) aims at forecasting a local meteorological variable of interest (the predictand), often the daily precipitation total, on the basis of a statistical relationship with synoptic predictor variables. A certain number of similar situations are sampled in order to establish the empirical conditional distribution which is considered as the prediction for a given date. The method is used in operational medium-range forecasting in several hydropower companies or flood forecasting services, as well as in climate impact studies. The statistical relationship is usually established by means of a semi-automatic sequential procedure that has strong limitations: it is made of successive steps and thus cannot handle parameters dependencies, and it cannot automatically optimize certain parameters, such as the selection of the pressure levels and the temporal windows on which the predictors are compared. A global optimization technique based on Genetic Algorithms was introduced in order to surpass these limitations and to provide a fully automatic and objective determination of the AM parameters. The parameters that were previously assessed manually, such as the selection of the pressure levels and the temporal windows, on which the predictors are compared, are now automatically determined. The next question is: Are Genetic Algorithms able to select the meteorological variable, in a reanalysis dataset, that is the best predictor for the considered predictand, along with the analogy criteria itself? Even though we may not find better predictors for precipitation prediction that the ones often used in Europe, due to numerous other studies which consisted in systematic assessments, the ability of an automatic selection offers new perspectives in order to adapt the AM for new predictands or new regions under different meteorological influences.

International audience

International audience