Dynamical Forecasting and Dynamical Interpolation: An Experiment in the California Current

View More View Less
  • 1 Center for Earth and Planetary Physics, Harvard University, Cambridge. MA 02138
  • | 2 Department of Oceanography, Naval Postgraduate School, Monterey, CA 93943
© Get Permissions
Full access

Abstract

In order to perform real-time dynamical forecasts and hindcasts, three high-resolution hydrographic surveys were made of a (150 km)2 domain off northern California, providing two sets of initialization and verification fields. The data was objectively analyzed and regularly gridded for model compatibility. These maps initially show an anticyclonic eddy segment in the northeast and part of another in the northwest. Two weeks later only the northwest anticyclonic eddy remained, with the domain center dominated by a 0.6 m s−1 jet. Two weeks after that only a larger northwest eddy with fairly weak velocities remained. Numerical forecasts with persistent boundary conditions and forecast experiments with boundary conditions linearly interpolated between surveys were performed. The real-time forecast successfully predicted the formation of the zonal jet prior to its observation. Dynamical interpolation shows unambiguously that the two anticyclonic eddies have merged and formed a single eddy. Even the forecast with incorrect boundary conditions demonstrates the internal dynamical processes involved in the merger event.

Two examples are given of four-dimensional data assimilation: direct insertion and a backward-forward combination technique. These results justify the use of the dynamical forecasts as synoptic time series. Parameter sensitivity experiments were performed to determine the sensitivity of the model to physical parameters such as stratification, to explore the dynamical balance, and to choose a reference level. The dynamics were found to be controlled by horizontal nonlinear interactions. A reference level of 1550 m was chosen. A set of energy and vorticity equations, consistent with quasi-geostrophic dynamics, were evaluated term by term for the forecast experiments. The evolutions of the streamfunction and vorticity fields are shown to be a three-phase (merging, expanding, and relaxation) process. Available gravitational energy increases due to buoyancy work; the merger event is interpreted as a finite amplitude barotropic instability process.

Abstract

In order to perform real-time dynamical forecasts and hindcasts, three high-resolution hydrographic surveys were made of a (150 km)2 domain off northern California, providing two sets of initialization and verification fields. The data was objectively analyzed and regularly gridded for model compatibility. These maps initially show an anticyclonic eddy segment in the northeast and part of another in the northwest. Two weeks later only the northwest anticyclonic eddy remained, with the domain center dominated by a 0.6 m s−1 jet. Two weeks after that only a larger northwest eddy with fairly weak velocities remained. Numerical forecasts with persistent boundary conditions and forecast experiments with boundary conditions linearly interpolated between surveys were performed. The real-time forecast successfully predicted the formation of the zonal jet prior to its observation. Dynamical interpolation shows unambiguously that the two anticyclonic eddies have merged and formed a single eddy. Even the forecast with incorrect boundary conditions demonstrates the internal dynamical processes involved in the merger event.

Two examples are given of four-dimensional data assimilation: direct insertion and a backward-forward combination technique. These results justify the use of the dynamical forecasts as synoptic time series. Parameter sensitivity experiments were performed to determine the sensitivity of the model to physical parameters such as stratification, to explore the dynamical balance, and to choose a reference level. The dynamics were found to be controlled by horizontal nonlinear interactions. A reference level of 1550 m was chosen. A set of energy and vorticity equations, consistent with quasi-geostrophic dynamics, were evaluated term by term for the forecast experiments. The evolutions of the streamfunction and vorticity fields are shown to be a three-phase (merging, expanding, and relaxation) process. Available gravitational energy increases due to buoyancy work; the merger event is interpreted as a finite amplitude barotropic instability process.

Save