All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 160 30 0
PDF Downloads 15 5 0

An Operational Global-Scale Ocean Thermal Analysis System

R. Michael ClancyOcean Models Division, Fleet Numerical Oceanography Center, Monterey, California

Search for other papers by R. Michael Clancy in
Current site
Google Scholar
PubMed
Close
,
Patricia A. PhoebusOcean Sensing and Prediction Division, Naval Ocean Research and Development Activity, John C. Stennis Space Center, Mississippi

Search for other papers by Patricia A. Phoebus in
Current site
Google Scholar
PubMed
Close
, and
Kenneth D. PollakOcean Models Division, Fleet Numerical Oceanography Center, Monterey, California

Search for other papers by Kenneth D. Pollak in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

The Optimum Thermal Interpolation System (OTIS) is an ocean thermal analysis product developed for real-time operational use at the U.S. Navy's Fleet Numerical Oceanography Center. It functions in an analysis-prediction-analysis data assimilation cycle with an ocean mixed-layer model using the optimum interpolation formulation of Alaka and Elvander (1972). Thus, climatology serves as the first-guess field for the analysis of synoptic ship, bathythermograph, buoy, and satellite data, with the prediction from the mixed-layer model treated as a special class of data.

An operational test, involving comparisons of OTIS against an existing operational ocean thermal structure model, was conducted during February, March, and April 1988. Qualitative comparison of the two products suggests that OTIS gives a more realistic representation of subsurface anomalies and horizontal gradients. Quantitative comparison of the two products with over 9900 independent (i.e., unassimilated) bathythermograph observations to calculate apparent root-mean-square (rms) errors indicate that OTIS gives a more accurate analysis of the thermal structure, with improvements largest below the base of the mixed layer. Regional differences in the relative performance of the models, which are probably related to the validity of the prescribed statistical parameters required by OTIS, are also noted from verification against the bathythermograph data. OTIS performs best relative to the existing model in the Indian Ocean, and poorest relative to this model in the eastern midlatitude Pacific. In the latter area, however, both models exhibit their lowest error levels and perform best relative to climatology.

Subgrid-scale noise contaminates the apparent rms error statistics and obscures the relative grid-scale accuracy of the models. A general technique is presented to estimate grid-scale rms errors which are uncontaminated by subgrid-scale and instrumental noise in the observations. Application of the technique indicates that the grid-scale errors for OTIS are typically 20% less than those of the other model.

OTIS was designated as the Navy's new operational global-scale ocean thermal analysis product in July 1988.

Abstract

The Optimum Thermal Interpolation System (OTIS) is an ocean thermal analysis product developed for real-time operational use at the U.S. Navy's Fleet Numerical Oceanography Center. It functions in an analysis-prediction-analysis data assimilation cycle with an ocean mixed-layer model using the optimum interpolation formulation of Alaka and Elvander (1972). Thus, climatology serves as the first-guess field for the analysis of synoptic ship, bathythermograph, buoy, and satellite data, with the prediction from the mixed-layer model treated as a special class of data.

An operational test, involving comparisons of OTIS against an existing operational ocean thermal structure model, was conducted during February, March, and April 1988. Qualitative comparison of the two products suggests that OTIS gives a more realistic representation of subsurface anomalies and horizontal gradients. Quantitative comparison of the two products with over 9900 independent (i.e., unassimilated) bathythermograph observations to calculate apparent root-mean-square (rms) errors indicate that OTIS gives a more accurate analysis of the thermal structure, with improvements largest below the base of the mixed layer. Regional differences in the relative performance of the models, which are probably related to the validity of the prescribed statistical parameters required by OTIS, are also noted from verification against the bathythermograph data. OTIS performs best relative to the existing model in the Indian Ocean, and poorest relative to this model in the eastern midlatitude Pacific. In the latter area, however, both models exhibit their lowest error levels and perform best relative to climatology.

Subgrid-scale noise contaminates the apparent rms error statistics and obscures the relative grid-scale accuracy of the models. A general technique is presented to estimate grid-scale rms errors which are uncontaminated by subgrid-scale and instrumental noise in the observations. Application of the technique indicates that the grid-scale errors for OTIS are typically 20% less than those of the other model.

OTIS was designated as the Navy's new operational global-scale ocean thermal analysis product in July 1988.

Save