A framework to quantify uncertainty in simulations of oil transport in the ocean

Abstract

An uncertainty quantification framework is developed for the DeepC Oil Model based on a nonintrusive polynomial chaos method. This allows the model's output to be presented in a probabilistic framework so that the model's predictions reflect the uncertainty in the model's input data. The new capability is illustrated by simulating the far‐field dispersal of oil in a Deepwater Horizon blowout scenario. The uncertain input consisted of ocean current and oil droplet size data and the main model output analyzed is the ensuing oil concentration in the Gulf of Mexico. A 1331 member ensemble was used to construct a surrogate for the model which was then mined for statistical information. The mean and standard deviations in the oil concentration were calculated for up to 30 days, and the total contribution of each input parameter to the model's uncertainty was quantified at different depths. Also, probability density functions of oil concentration were constructed by sampling the surrogate and used to elaborate probabilistic hazard maps of oil impact. The performance of the surrogate was constantly monitored in order to demarcate the space‐time zones where its estimates are reliable.

An overview of uncertainty quantification techniques with application to oceanic and oil‐spill simulations

“…The articles of Gonçalves et al . [] and Wang 2016 focus on their application in oil plume and oil‐fate modeling while Iskandarani 2016 focus on quantifying uncertainties in oceanic forecasts in the Gulf of Mexico due to initial condition uncertainties.…”

“…[a, ], and Gonçalves et al . [] and in Wang 2016 and Iskandarani 2016 illustrate how the observational data are used to identify the uncertainty ranges of the input parameters. See also Webster and Sokolov [] for an interesting discussion on how expert opinion can be used to inform the specification of input PDFs.…”

“…Iskandarani 2016 and Gonçalves et al . [] rely on these EOFs to perturb fields. For the remainder of the paper, we assume that the user has capitalized on the available a priori information to specify the uncertain parameters and their input PDFs.…”

“…One solution is to decompose the field into Empirical Orthogonal Functions (EOFs) and to characterize the field's uncertainties as uncertainties in the amplitudes of a few EOFs. Iskandarani 2016 andGonçalves et al [2016] rely on these EOFs to perturb fields. For the remainder of the paper, we assume that the user has capitalized on the available a priori information to specify the uncertain parameters and their input PDFs.…”

“…Such a comparison is not feasible for more complex models due to the high cost of each run, but this model's relatively low cost permitted the use of traditional Monte Carlo simulations to obtain output statistics for comparison with those resulting from the inexpensive techniques. Applications of these different techniques to oil‐fate modeling [ Gonçalves et al ., ], plume modeling (S. Wang et al, Propagation of uncertainty and sensitivity analysis in an integral oil‐gas plume model, submitted to Journal of Geophysical Research: Oceans , 2016, Wang 2016 hereafter), and to circulation forecasting in the Gulf of Mexico (M. Iskandarani et al, Quantifying Uncertainty in Gulf of Mexico Forecasts Stemming from Uncertain Initial Conditions, submitted to Journal of Geophysical Research: Oceans, 2016, Iskandarani 2016 hereafter) will be presented in separate articles in the special JGR issue on “Physical Processes Responsible for Material Transport in the Gulf of Mexico for Oil Spill Applications.” The present article consolidates the methodology sections of these papers. It also provides the opportunity to present to oceanographers the rapidly evolving field of uncertainty quantification.…”