Séminaire WP3 Incertitudes
le 22 mai 2018
Dans le cadre du WP3, Clémentine Prieur et Anne-Catherine Favre ont organisé un séminaire portant sur les incertitudes.
Résumé et téléchargement des présentations effectuées :
Guillaume Evin (IRSTEA)
Titre : Traitements des incertitudes avec un ensemble incomplet de scénarios climatiques.
Résumé : Les études qui s'intéressent aux impacts du changement climatique doivent composer avec un nombre croissant de trajectoires représentant des futurs possibles. Cette multiplicité résulte d'un ensemble de scénarios d'évolution des gaz à effet de serre et d'un nombre croissant de modèles climatiques globaux et régionaux et de modèles d'impact (hydrologie, biodiversité, etc.) comportant tous leur lot d'hypothèses. A cette incertitude se conjugue la variabilité naturelle du climat.De nombreuses méthodes sont désormais disponibles pour analyser ces différentes incertitudes à partir de scénarios multi-modèles et multi-membres. L'approche la plus populaire suppose que plusieurs membres sont disponibles pour chaque chaine de modèles. Cependant, il n' y a souvent pas de répétition des scénarios pour chaque combinaison de modèles (RCP/GCM/RCM) et des combinaisons manquantes. Nous présentons ici une approche statistique permettant de partitionner les différentes sources d'incertitude dans un tel contexte, avec des approches bayésiennes.
Résumé : The accelerating environmental changes clearly demand for biodiversity scenarios that provide trends and alternative states in biodiversity with respect to future emissions. Yet, a thorough analysis and communication of the associated uncertainties is still largely missing2,3. Here, we provide the first comprehensive uncertainty analysis of global biodiversity scenarios. We modelled the global distribution of ~11,500 amphibian, bird and mammal species and project their potential climatic suitability into the time horizon 2050 and 2070, while varying the input data used. By this, we explore the uncertainties originating from selecting species distribution models (SDMs), dispersal strategies, global circulation models (GCMs), and representative concentration pathways (RCPs). Using a full-factorial analysis of uncertainties, we demonstrate the overwhelming influence of SDMs and RCPs on future biodiversity projections, followed by dispersal strategies and GCMs. The relative importance of each component varies in space but also with the selected sensitivity metrics used. This suggests that the use of multiple SDMs, RCPs, dispersal assumptions and GCMs is a necessity in any biodiversity scenario assessment. We recommend that, to be of any use for biodiversity management and international programs (e.g. IPCC, IPBES), biodiversity scenario studies carry out a full analysis of uncertainties across space and sensitivity metrics, and report ensembles of results rather than average values.
Résumé : Pour le scientifique comme pour le gestionnaire, la prise en compte des incertitudes est devenue une nécessité bien comprise autant qu'un impératif affiché. Pourtant, en pratique, il s'agit bien souvent d'un vœux pieu faute d'outils efficaces appropriés par l'ingénieur et le décideur. L’objet de cet exposé est de montré sur des exemples tirés de l’ingénierie paravalanche qu’il est pourtant à la fois possible et souhaitable d’intégrer l’incertitude dans les décisions relatives à l’aménagement du territoire. Pour ce faire, le formalisme bayésien décisionnel est utilisé et appliqué au concept de risque combinant modèle stochastique d’aléa et quantification des dommages pour le bâti et ses habitants. On montre ainsi que la prise en compte du caractère limité de l’information disponible conduit à des choix plus prudents.
Résumé : Many physical phenomena are modelled numerically in order to better understand and/or to predict their behaviour. However, some complex and small scale phenomena can not be fully represented in the models. The introduction of ad-hoc correcting terms, can represent these unresolved processes, but they need to be properly estimated.
A good example of this type of problem is the estimation of bottom friction parameters of the ocean floor. This is important because it affects the general circulation. This is particularly the case in coastal areas, especially for its influence on wave breaking. Because of its strong spatial disparity, it is impossible to estimate the bottom friction by direct observation, so it requires to do so indirectly by observing its effects on surface movement. This task is further complicated by the presence of uncertainty in certain other characteristics linking the bottom and the surface (eg boundary conditions). The techniques currently used to adjust these settings are very basic and do not take into account these uncertainties, thereby increasing the error in this estimate.
Classical methods of parameter estimation usually imply the minimisation of an objective function, that measures the error between some observations and the results obtained by a numerical model. In the presence of uncertainties, the minimisation is not straightforward, as the output of the model depends on those uncontrolled inputs and on the control parameter as well. That is why we will aim at minimising the objective function, to get an estimation of the control parameter that is robust to the uncertainties. In this work, a toy model of a coastal is modelled and implemented. The control parameter is the bottom friction, upon which classical methods of estimation are applied in a simulation-reestimation experiment. The model is then modified to include uncertainties on the boundary conditions in order to apply robust control methods. A study on the meaning of different concepts of robustness is therefore carried on. Typically, one then seeks an optimal parameter set that would minimise the variance or the mean of the original objective function.
Wilfried Thuiller (LECA)
Titre : Uncertainty in ensembles of global scenarios of biodiversityRésumé : The accelerating environmental changes clearly demand for biodiversity scenarios that provide trends and alternative states in biodiversity with respect to future emissions. Yet, a thorough analysis and communication of the associated uncertainties is still largely missing2,3. Here, we provide the first comprehensive uncertainty analysis of global biodiversity scenarios. We modelled the global distribution of ~11,500 amphibian, bird and mammal species and project their potential climatic suitability into the time horizon 2050 and 2070, while varying the input data used. By this, we explore the uncertainties originating from selecting species distribution models (SDMs), dispersal strategies, global circulation models (GCMs), and representative concentration pathways (RCPs). Using a full-factorial analysis of uncertainties, we demonstrate the overwhelming influence of SDMs and RCPs on future biodiversity projections, followed by dispersal strategies and GCMs. The relative importance of each component varies in space but also with the selected sensitivity metrics used. This suggests that the use of multiple SDMs, RCPs, dispersal assumptions and GCMs is a necessity in any biodiversity scenario assessment. We recommend that, to be of any use for biodiversity management and international programs (e.g. IPCC, IPBES), biodiversity scenario studies carry out a full analysis of uncertainties across space and sensitivity metrics, and report ensembles of results rather than average values.
Michaela Bevione (INRIA, PACTE)
Titre : Enjeux socio-écologiques, métabolisme territorial, création de richesse : application à la Vallée de la MaurienneNicolas Eckert
Titre : Pourquoi et comment prendre en compte l’incertitude dans les décisions d’aménagement ? Des exemples en ingénierie paravalanche.Résumé : Pour le scientifique comme pour le gestionnaire, la prise en compte des incertitudes est devenue une nécessité bien comprise autant qu'un impératif affiché. Pourtant, en pratique, il s'agit bien souvent d'un vœux pieu faute d'outils efficaces appropriés par l'ingénieur et le décideur. L’objet de cet exposé est de montré sur des exemples tirés de l’ingénierie paravalanche qu’il est pourtant à la fois possible et souhaitable d’intégrer l’incertitude dans les décisions relatives à l’aménagement du territoire. Pour ce faire, le formalisme bayésien décisionnel est utilisé et appliqué au concept de risque combinant modèle stochastique d’aléa et quantification des dommages pour le bâti et ses habitants. On montre ainsi que la prise en compte du caractère limité de l’information disponible conduit à des choix plus prudents.
Elise Arnaud (LJK)
Titre : Parameter control in presence of uncertainties: robust estimation of bottom frictionRésumé : Many physical phenomena are modelled numerically in order to better understand and/or to predict their behaviour. However, some complex and small scale phenomena can not be fully represented in the models. The introduction of ad-hoc correcting terms, can represent these unresolved processes, but they need to be properly estimated.
A good example of this type of problem is the estimation of bottom friction parameters of the ocean floor. This is important because it affects the general circulation. This is particularly the case in coastal areas, especially for its influence on wave breaking. Because of its strong spatial disparity, it is impossible to estimate the bottom friction by direct observation, so it requires to do so indirectly by observing its effects on surface movement. This task is further complicated by the presence of uncertainty in certain other characteristics linking the bottom and the surface (eg boundary conditions). The techniques currently used to adjust these settings are very basic and do not take into account these uncertainties, thereby increasing the error in this estimate.
Classical methods of parameter estimation usually imply the minimisation of an objective function, that measures the error between some observations and the results obtained by a numerical model. In the presence of uncertainties, the minimisation is not straightforward, as the output of the model depends on those uncontrolled inputs and on the control parameter as well. That is why we will aim at minimising the objective function, to get an estimation of the control parameter that is robust to the uncertainties. In this work, a toy model of a coastal is modelled and implemented. The control parameter is the bottom friction, upon which classical methods of estimation are applied in a simulation-reestimation experiment. The model is then modified to include uncertainties on the boundary conditions in order to apply robust control methods. A study on the meaning of different concepts of robustness is therefore carried on. Typically, one then seeks an optimal parameter set that would minimise the variance or the mean of the original objective function.
Publié le 23 mai 2019