Oceanic Data Assimilation Tests with a One-Dimensional Model.
Abstract
A data assimilation technique for incorporating relatively sparse ocean thermal structure profiles into the Garwood (1977) Oceanic Planetary Boundary Layer (OPBL) model is proposed. A summary of the data assimilation tests by Elsberry and Warrenfeltz (EW) is presented. The complete and perfect model generated verification data from the EW study were used to simulate incomplete and noisy data as might be expected in real data verifications. Random errors that are normally distributed about the mean mixed layer depth (MLD) and temperature (MLT), are added to subsets of the EW verification data during the summer and winter regimes. From these simulated tests, it was concluded that a data assimilation technique with a 1-D OPBL model can improve predictions of the ocean thermal structure even with incomplete and noisy verification data. Real bathythermographic temperature profiles from Ocean Weather Station PAPA are then inserted into the Garwood model to verify the EW data assimilation studies. The tests with real data demonstrate the necessity of defining the MLD in an observed profile that is consistent with the model output MLD. In addition, biases were observed that originated from the use of an imperfect model. After the elimination of the biases and the MLD descrepancies, it is suggested that a 1-D model used for data assimilation can improve predictions of the ocean thermal structure. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1981
- Accession Number
- ADA115175
Entities
People
- Dennis Glenn Larsen
Organizations
- Naval Postgraduate School