Study Quantifies Influence of Data Input on Confidence in Loop Current Forecasts

3981

Time evolution of the Sea Surface Height anomaly from AVISO altimetry data (color), with the Loop current edge, as defined by the 17 cm contour, in the Polynomial Chaos ensemble (black lines) and in the AVISO data (white line). Image provided by Iskandarani.

Researchers described in a recent study a surrogate-based technique to quantify the uncertainty in forecasting the oceanic circulation. The authors focused on the time period during the Deepwater Horizon oil spill when an extended Loop Current increased the risk of carrying the oil slick towards the eastern seaboard of the U.S.  The new methodology, which accounts explicitly for the inherent uncertainty in forecasts, may help improve the planning of emergency responses to weather and marine pollution events.  The authors’ paper was published in the Journal of Geophysical Research: Oceans: Quantifying uncertainty in Gulf of Mexico forecasts stemming from uncertain initial conditions.

The accuracy and usefulness of material transport models depend on the quality of oceanic and atmospheric forecasts. However, the input data needed to run the forecast models are incomplete because observations are limited in space and time and may include measurement errors. The uncertainties in the model input lead to uncertainties in the model output. Useful forecasts should include a quantitative assessment of these uncertainties to better inform decisions such as evacuations or deploying resources. Probabilistic forecasts allow policy makers and emergency responders to consider a range of possible scenarios instead of only one best-guess scenario whose certainty is unclear.

The study focused on the period from May – June 2010, when the major concern was whether or not the Deepwater Horizon oil would be entrained in the Loop Current. “The Loop Current can act like a conveyor belt capable of moving water swiftly from the northern Gulf of Mexico to the Florida Strait,” explained study author Mohamed Iskandarani. “The Gulf Stream would then carry it quickly farther north.” Therefore, quantifying the uncertainty in Loop Current forecasts was a major goal of this study.

The researchers focused on the impact of misjudging the strength of a frontal eddy that was influencing the Loop Current’s path.  Using a polynomial chaos technique, the team ran a small ensemble of 49 scenarios (with varying frontal eddy strength) to build a cost-effective but faithful surrogate of the full model. They used this surrogate to determine how the model output changed with different input data. “The surrogate provides more accurate statistics than a traditional ensemble and can be analyzed to decide which input is most responsible for the largest portion of the uncertainty,” said Iskandarani. “This provides guidance for additional observations that can reduce the uncertainty in the model output most effectively.”

Simulations showed that a weak eddy would have kept the Loop Current on a trajectory to enter the spill area, potentially increasing the risk of oil spreading, while a stronger eddy would have separated the Loop Current from its trajectory before reaching the slick. “Uncertainties in the frontal eddy strength strongly affected the prediction of a Loop Current eddy detachment 15-30 days later,” explained Iskandarani. The study suggested that the predictability limit of the forecast was three weeks, and assimilating new observations is necessary to improve the forecast beyond that time.

The study’s findings have the potential to support an operational observing system for the Loop Current. “There would be great, two-way benefits from pairing an observing system with a probabilistic forecast model,” said Iskandarani. “The forecast model could be used to propagate these uncertainties forward in time to estimate the output uncertainties. The biggest contributors to the output uncertainties can be identified and targeted observations can be planned to reduce the uncertainties in the inputs.”

This study’s data are publicly available through the Gulf of Mexico Research Initiative Information & Data Cooperative (GRIIDC) at doi:10.7266/N77H1GNF.

The study’s authors are Mohamed Iskandarani, Matthieu Le Henaff, William Carlisle Thacker, Ashwanth Srinivasan, and Omar M. Knio.

************
This research was made possible in part by a grant from the Gulf of Mexico Research Initiative (GoMRI) to the Consortium for Advanced Research on Transport of Hydrocarbon in the Environment II (CARTHE II). Other funding sources included the Office of Naval Research (Award N00014-101- 0498), the US Department of Energy (DOE), Office of Science, Office of Advanced Scientific Computing Research (DE-SC0008789), the NOAA Quantitative Observing System Assessment Program (NA15OAR4320064), and the NOAA Atlantic Oceanographic and Meteorological Laboratory.

The Gulf of Mexico Research Initiative (GoMRI) is a 10-year independent research program established to study the effect, and the potential associated impact, of hydrocarbon releases on the environment and public health, as well as to develop improved spill mitigation, oil detection, characterization and remediation technologies. An independent and academic 20-member Research Board makes the funding and research direction decisions to ensure the intellectual quality, effectiveness and academic independence of the GoMRI research. All research data, findings and publications will be made publicly available. The program was established through a $500 million financial commitment from BP. For more information, visit http://gulfresearchinitiative.org/.

© Copyright 2010- 2017 Gulf of Mexico Research Initiative (GoMRI) – All Rights Reserved. Redistribution is encouraged with acknowledgement to the Gulf of Mexico Research Initiative (GoMRI). Please credit images and/or videos as done in each article. Questions? Contact web-content editor Nilde “Maggie” Dannreuther, Northern Gulf Institute, Mississippi State University (maggied@ngi.msstate.edu).