A Program for Robustness Analysis: overcoming uncertainty in decision-making
Abstract
In an effort to produce actionable science, the hydrological community is increasingly aiming to make the uncertainty present in modeling practices and models transparent to decision-makers. However modelers are faced with model uncertainty- that is uncertainty in choosing between competing available models which could be used to represent a given hydrological application or system. Traditionally in the sciences, replication of laboratory or field experiments have been viewed as the methodology for ensuring that models are reliable internally and externally, and thus having some form of confidence that they map onto the world. However, in the case of large-scale simulations, I argue that the replicating a model's results only provides evidence for the internal validity, but not the external validity, of the model. Robustness analysis is required to provide evidence for the external validity of a model, which requires comparison to the results of other models regarding similar climatological and/or hydrological scenarios. I lay out a program for hydrological modelers and decision-makers to work together to use inferences from robustness to ensure that climate model inferences map onto the particular local systems appropriately, giving confidence in model-world-fit.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFMGC13F1256F
- Keywords:
-
- 3355 Regional modeling;
- ATMOSPHERIC PROCESSESDE: 1807 Climate impacts;
- HYDROLOGYDE: 1847 Modeling;
- HYDROLOGYDE: 1952 Modeling;
- INFORMATICS