Record Details
Data assimilation for prediction of shallow water flows with uncertain bathymetry
ScholarsArchive at Oregon State University
| Field | Value |
|---|---|
| Title | Data assimilation for prediction of shallow water flows with uncertain bathymetry |
| Names |
Wilson, Gregory W.
(creator) Ozkan-Haller, H. Tuba (advisor) |
| Date Issued | 2013-04-15 (iso8601) |
| Note | Graduation date: 2013 |
| Abstract | A new method is introduced for incorporating bathymetric uncertainty into predictions of nearshore and river flows (i.e., unstratified flows primarily forced by pressure and radiation stress gradients). The method involves the use of the ensemble Kalman filter (EnKF) as a parameter estimation scheme, where the parameter to be estimated is the spatial field of bathymetry. That is, bathymetry is treated as a slowly varying uncertain parameter in the model, which can be corrected via the assimilation of other available observations. The reason bathymetry is targeted is, as we show, it is often a limiting factor for accuracy in real-world modeling applications. Results are shown using data from four field experiments. Two experiments involve measurements in the nearshore (surf zone) ocean at Duck, NC. There, we show that bathymetric uncertainty due to rapid bathymetric change (time scale of days), or simply lack of available measurements, can cause significant error in model predictions of waves and currents. We demonstrate the ability of the EnKF to reduce this error by correcting the bathymetry, which we then cross-validate using in-situ measurements. Specifically, the correction is achieved by assimilating in-situ observations of alongshore current and significant wave height, as well as (in a separate experiment) remote-sensing observations of alongshore current, wave celerity, and location of shoreline. Similarly in a river environment (Snohomish River, WA, and Kootenai River, ID), we demonstrate the EnKF using twin tests, assimilating pseudo-observations of currents from a variety of hypothesized platforms (fixed in-situ gages, passive drifters, and Doppler radar). Again, the EnKF is found to yield accurate estimates of bathymetry. |
| Genre | Thesis/Dissertation |
| Topic | nearshore oceanography |
| Identifier | http://hdl.handle.net/1957/38466 |
