Sensitivity Analysis of Limited Area Ocean Model
Abstract
The boundary data for a data assimilation problem for a limited-area ocean model can either be supplied by its global model or from observations interpolated from a conventional observing network. In this study, the sensitivity of the ocean model to variations in the boundary data is investigated. Additionally, the authors look at how predictions of wave height are affected by variations in basin depth. Two models are used to carry out the analyses: a large-scale ocean circulation model and a combined wave refraction diffraction model (Ref/Dif 1 model). The large-scale circulation model is used to implement the limited-area ocean model. Results show that the accuracy and availability of boundary data affect the model's predictions. For the limited-area model, error was observed in the solution even when all the boundary data were passed to the limited-area model. This hints that on top of the need for accurate boundary data, the initial data might need to be initialized by the bounded derivative method, for example, to ensure that error does not propagate into the solution. It has been shown that variability in the bottom topography does not affect the results obtained from Ref/Dif 1. The density plots of the wave height and water depth concur with the mean and variance plots of the wave height and bottom topography. Furthermore, a comparison of the mean solution to the deterministic solution shows that the random bottom topography had little effect on the Ref/Dif 1 results. (33 figures, 12 refs.)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2002
- Accession Number
- ADA417044
Entities
People
- Dubar K. Kamara
Organizations
- Johns Hopkins University