Description
- Abstract:
- Abstract of "OSOM: Predictability, Variability, and Response To Parameterizations Quantified Using Information Theory" by Aakash Bhalchandra Sane, Ph.D., Brown University, October 2021. The Ocean State Ocean Model (OSOM) spans the Rhode Island waterways from the Long Island Shelf to the region around Nantucket, including Narragansett Bay, Mt. Hope Bay, major rivers, and the Block Island Shelf. In this work, the OSOM has been implemented in the Regional Ocean Modeling System to set up a forecasting system and understand the physical aspects of the regional oceanic circulation. Ensemble simulations have been performed by perturbing the initial conditions of the model. This data, when evaluated with information theory-derived metrics, quantifies the predictability of the OSOM to infer the temporal persistence of anomalies. The predictability timescales inform the effectiveness of a forecasting system and prepare for coupling with biogeochemistry and fishery models with extensively varying timescales. The predictability of the OSOM is ~ 7 to 40 days, varying with parameters, region, and season. In this work, a new metric is proposed using two quantities from Information Theory - Shannon entropy and Mutual Information - to measure grid point internal and forced variability in ensemble ocean, atmosphere, and climate models. The proposed metric delineates intrinsic and extrinsic variability by measuring the visited probability distribution, as opposed to a variance metric that captures only its second statistical moment. The proposed metric responds to correlated fields apply to any data without assuming its probability distribution, is insensitive to outliers and changes of units or scale. Additionally, Shannon Entropy and Mutual Information have been applied to measure the sensitivity of temperature and salinity affected by modifying the external forcing conditions. The OSOM has been used to compare three different turbulence models, two of which are standard one- or two-equation models, while the third is a modified one-equation model that includes effects of symmetric instability due to horizontal gradients of the buoyancy and Coriolis effects. The results from the two-equation model are the most statistically distant from the other two, as given by the Mutual Information measure, while those from the one-equation models are close. These results suggest that possible SI effects on turbulence parameterization are limited in the present context. This work highlights Information Theory as a useful tool by demonstrating its use in analyzing outputs from the ocean and climate models. These metrics rank the potential impacts of improving boundary forcings, mixing parameterizations, and forcing conditions across multiple variables.
- Notes:
- Thesis (Ph. D.)--Brown University, 2021
Citation
Sane, Aakash Bhalchandra,
"OSOM: Predictability, Variability, and Response To Parameterizations Quantified Using Information Theory"
(2021).
Fluid, Thermal, and Chemical Processes Theses and Dissertations.
Brown Digital Repository. Brown University Library.
https://repository.library.brown.edu/studio/item/bdr:g23278v8/
Relations
Collection:
-
Fluid, Thermal, and Chemical Processes Theses and Dissertations
Theses and Dissertations for the Fluid, Thermal, and Chemical Processes department....