Explainability, uncertainty & physics
Ensemble‐Based Experimental Design for Targeting Data Acquisition to Inform Climate Models
In one sentence. An ensemble-based algorithm targets data acquisition to regions and times that most reduce parametric uncertainty in a GCM convection scheme, within calibrate-emulate-sample.
Abstract
Abstract Data required to calibrate uncertain general circulation model (GCM) parameterizations are often only available in limited regions or time periods, for example, observational data from field campaigns, or data generated in local high‐resolution simulations. This raises the question of where and when to acquire additional data to be maximally informative about parameterizations in a GCM. Here we construct a new ensemble‐based parallel algorithm to automatically target data acquisition to regions and times that maximize the uncertainty reduction, or information gain, about GCM parameters. The algorithm uses a Bayesian framework that exploits a quantified distribution of GCM parameters as a measure of uncertainty. This distribution is informed by time‐averaged climate statistics restricted to local regions and times. The algorithm is embedded in the recently developed calibrate‐emulate‐sample framework, which performs efficient model calibration and uncertainty quantification with only model evaluations, compared with evaluations typically needed for traditional approaches to Bayesian calibration. We demonstrate the algorithm with an idealized GCM, with which we generate surrogates of local data. In this perfect‐model setting, we calibrate parameters and quantify uncertainties in a quasi‐equilibrium convection scheme in the GCM. We consider targeted data that are (a) localized in space for statistically stationary simulations, and (b) localized in space and time for seasonally varying simulations. In these proof‐of‐concept applications, the calculated information gain reflects the reduction in parametric uncertainty obtained from Bayesian inference when harnessing a targeted sample of data. The largest information gain typically, but not always, results from regions near the intertropical convergence zone.
Labels
- Topic
- Explainability, uncertainty & physics, Hybrid physics-ML & parameterization
- Type
- New model or method
- Method
- Classical ML (trees, GP, linear, kernels)
- Domain
- Global
- Data
- Other
- Status
- Research
- Identifiers
- arXiv:2201.06998 DOI:10.1029/2022ms002997 OpenAlex:W4283732814
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.