Non-Linear Coupling of Scales of Ocean Variability and Implications for Downscaling

Abstract

The objectives of this study are to develop better methods for dynamically downscaling ocean conditions, and to improve understanding of the physical processes controlling variations on regional scales. Motivated by the study of Henshaw et al. (2003), the predictability of small scale variability from the time history of the associated large scales is investigated. This leads to a practical method for assimilating large scale information in order to recover small scale variability. The method is similar to spectral nudging of limited area atmospheric models but it is modified for ocean applications. The method is first applied to an idealized ocean model that captures the most relevant oceanographic non-linearity: baroclinic instability. Numerical experiments show that important features of the ocean circulation, including the position of a meandering mid-ocean jet and the associated eddies, can be recovered from the time history of a small number of large scale modes. A hybrid approach is then proposed for assimilating both the large scales and additional point observations from a limited number of locations that, alone, are too sparse to recover the small scales using traditional assimilation methods. It is shown that the assimilation of sparse observations improves significantly the accuracy of the recovered small scales after the correct large scales have been ingested by the model. A high resolution model of the Gulf of Maine, Scotian Shelf and adjacent deep ocean is then developed based on the Nucleus for European Modelling of the Ocean to downscale fields from a well known global operational system, specifically the US Naval Oceanography system based on the HYbrid Coordinate Ocean Model. The regional model is validated, using observations and results from previous studies, and weaknesses associated with simple one-way nesting are identified. The new downscaling method is shown to overcome these weaknesses. The implications for more realistic downscaling of ocean conditions are discussed. In the process of developing more accurate ocean downscaling methodologies, new regional oceanographic features of scientific interest and practical importance (e.g., seasonal variations of tidal current speed) are identified and explained.