Today, a major challenge for climate science is to overcome what is called the "usability gap" be... more Today, a major challenge for climate science is to overcome what is called the "usability gap" between the projections derived from climate models and the needs of the end-users. Regional Climate Models (RCMs) are expected to provide usable information concerning a variety of impacts and for a wide range of end-users. It is often assumed that the development of more accurate, more complex RCMs with higher spatial resolution should bring process understanding and better local projections, thus overcoming the usability gap. In this paper, I rather assume that the credibility of climate information should be pursued together with two other criteria of usability, which are salience and legitimacy. Based on the Swiss climate change scenarios, I study the attempts at meeting the needs of end-users and outline the trade-off modellers and users have to face with respect to the cascade of uncertainty. A conclusion of this paper is that the trade-off between salience and credibility sets the conditions under which RCMs can be deemed adequate for the purposes of addressing the needs of end-users and gearing the communication of the projections toward direct use and action. Regional Climate Models • Climate change impact • Climate services • Climate uncertainties • Non-epistemic values • Usable or actionable information * Julie Jebeile
Parameterization and parameter tuning are central aspects of climate modeling, and there is wides... more Parameterization and parameter tuning are central aspects of climate modeling, and there is widespread consensus that these procedures involve certain subjective elements. Even if the use of these subjective elements is not necessarily epistemically problematic, there is an intuitive appeal for replacing them with more objective (automated) methods, such as machine learning. Relying on several case studies, we argue that, while machine learning techniques may help to improve climate model parameterization in several ways, they still require expert judgment that involves subjective elements not so different from the ones arising in standard parameterization and tuning. The use of machine learning in parameterizations is an art as well as a science and requires careful supervision.
International audienceNotice biblio Julie Jebeile réalise une thèse de doctorat sur la compréhens... more International audienceNotice biblio Julie Jebeile réalise une thèse de doctorat sur la compréhension scientifique dans les sciences empiriques. Elle s'interroge en particulier sur le rôle explicatif des modèles qui sont utilisés à la fois comme représentations de systèmes naturels ou sociaux et comme outils inférentiels. Dans ses travaux, elle cherche notamment à évaluer l'éventuelle dégradation ou amélioration de la compréhension induite par le tournant computationnel des sciences, c'est-à-dire par l'utilisation de l'ordinateur dans l'activité de modélisation. Résumé Dans cet article, je défends l'idée que la philosophie de la connaissance actuelle, telle qu'elle s'applique aux modèles classiques, se révèle être inopérante dans le cas particulier des simulations numériques. A cette fin, je recense dans un premier temps les procédures d'une justification dite « traditionnelle » des modèles classiques. Dans un second temps, je montre tour à tou...
Épistémologie Des Modèles et Des Simulations Numériques
Understanding phenomena often requires using mathematical models of the target systems. In partic... more Understanding phenomena often requires using mathematical models of the target systems. In particular, this requires obtaining, through them, reliable answers to whyquestions. In this context, we achieve understanding once the models are acceptable and intelligible; this is the central assumption in this thesis. This double requirement is thus studied first in the analysis of analytical models, and then in the analysis of simulation models. This study first allowed us to highlight the positive role of idealizations in understanding through analytical models. Next, it allowed for an identification of the consequences of the computational turn. There is in fact a gap between a computational model and its results, partly because of the epistemic opacity of computer simulations. This gap seems to doubly hinder our understanding of simulated phenomena. On the one hand, some epistemological difficulties arise which are specific to the justification and the use of simulation models. These difficulties contravene their acceptability. On the other hand, since simulation is not open to direct inspection, it seems difficult for a user to make the relation between the model content and its results. Nevertheless, visual representations seem to play a fundamental function in allowing us to overcome the opacity issue, and thus to provide us with explanatory elements to our why-questions
Expliquer et comprendre dans les sciences empiriques : les modèles scientifiques et le tournant computationnel
Comprendre les phenomenes consiste souvent a interroger les modeles mathematiques des systemes co... more Comprendre les phenomenes consiste souvent a interroger les modeles mathematiques des systemes consideres. En particulier. il s'agit d'obtenir par leur intermediaire des reponses fiables aux questions de type « pourquoi'? ». Nous y reussissons des lors que les modeles sont acceptables et intelligibles: c'est l'idee directrice de la these. Ce double requisit est ainsi etudie; d'abord dans l'analyse des modeles analytiques puis dans celle des modeles de simulation. Cela a permis dans un premier temps de mettre en lumiere le role positif des idealisations dans la comprehension par les modeles analytiques. Puis, dans un second temps, il a ete possible d'identifier les consequences du tournant computationnel. Il demeure en effet un fosse entre le modele computationnel et ses resultats, il cause, notamment de l'opacite epistemologique des simulations numeriques. Or ce fosse semble doublement entraver notre comprehension des phenomenes simules. En effet,...
Boston Studies in the Philosophy and History of Science, 2017
In empirical modeling, mathematics has an important utility in transforming descriptive represent... more In empirical modeling, mathematics has an important utility in transforming descriptive representations of target system(s) into calculation devices, thus creating useful scientific models. The transformation may be considered as the action of tools. In this paper, I assume that model idealizations could be such tools. I then examine whether these idealizations have characteristic properties of tools, i.e., whether they are being adapted to the objects to which they are applied, and whether they are to some extent generic.
In philosophical studies regarding mathematical models of dynamical systems, instability due to s... more In philosophical studies regarding mathematical models of dynamical systems, instability due to sensitive dependence on initial conditions, on the one side, and instability due to sensitive dependence on model structure, on the other, have by now been extensively discussed. Yet there is a third kind of instability, which by contrast has thus far been rather overlooked, that is also a challenge for model predictions about dynamical systems. This is the numerical instability due to the employment of numerical methods involving a discretization process, where discretization is required to solve the differential equations of dynamical systems on a computer. We argue that the criteria for numerical stability, as usually provided by numerical analysis textbooks, are insufficient, and, after mentioning the promising development of backward analysis, we discuss to what extent, in practice, numerical instability can be controlled or avoided.
Studies in History and Philosophy of Science Part A, 2020
Projections of future climate change cannot rely on a single model. It has become common to rely ... more Projections of future climate change cannot rely on a single model. It has become common to rely on multiple simulations generated by Multi-Model Ensembles (MMEs), especially to quantify the uncertainty about what would constitute an adequate model structure. But, as Parker points out ( ), one of the remaining philosophically interesting questions is: "How can ensemble studies be designed so that they probe uncertainty in desired ways?" This paper offers two interpretations of what General Circulation Models (GCMs) are and how MMEs made of GCMs should be designed. In the first interpretation, models are combinations of modules and parameterisations; an MME is obtained by "plugging and playing" with interchangeable modules and parameterisations. In the second interpretation, models are aggregations of expert judgements that result from a history of epistemic decisions made by scientists about the choice of representations; an MME is a sampling of expert judgements from modelling teams. We argue that, while the two interpretations involve distinct domains from philosophy of science and social epistemology, they both could be used in a complementary manner in order to explore ways of designing better MMEs.
International Studies in the Philosophy of Science, 2017
It is often said that computer simulations generate new knowledge about the empirical world in th... more It is often said that computer simulations generate new knowledge about the empirical world in the same way experiments do. My aim is to make sense of such a claim. I first show that the similarities between computer simulations and experiments do not allow them to generate new knowledge but invite the simulationist to interact with simulations in an experimental manner. I contend that, nevertheless, computer simulations and experiments yield new knowledge under the same epistemic circumstances, independently of any features they may share.
Computer simulations are often expected to provide explanations about target phenomena. However t... more Computer simulations are often expected to provide explanations about target phenomena. However there is a gap between the simulation outputs and the underlying model, which prevents users finding the relevant explanatory components within the model. I contend that visual representations which adequately display the simulation outputs can nevertheless be used to get explanations. In order to do so, I elaborate on the way graphs and pictures can help one to explain the behavior of a flow past a cylinder. I then specify the reasons that make more generally visual representations particularly suitable for explanatory tasks in a computer-assisted context.
Uploads
Papers by Julie Jebeile