Comment on os-2021-77

doi:10.5194/os-2021-77-rc2

Abstract

An effective and computationally efficient method is presented for data assimilation in a high-resolution (child) ocean model, which is nested into a coarse-resolution good quality data assimilating (parent) model. The method named Data Assimilation with Stochastic-Deterministic Downscaling (SDDA) reduces bias and root mean square errors (RMSE) of the child model and does not allow the child model to drift away from reality. The basic idea is to assimilate data from the parent model instead of actual observations. In this way, the child model is physically aware of observations via the parent model. The method allows to avoid a complex process of assimilating the same observations which were already assimilated into the parent model. The method consists of two stages: (1) downscaling the parent model output onto the child model grid using Stochastic-Deterministic Downscaling, and (2) applying a simplified Kalman gain formula to each of the fine grid nodes. The method is illustrated in a synthetic case where the true solution is known, and the child model forecast (before data assimilation) is simulated by adding various types of errors. The SDDA method reduces the child model bias to the same level as in the parent model and reduces the RMSE typically by a factor of 2 to 5.

Highlights

Due to inevitable approximations in the equations, numerical schemes, parameterisation and uncertainties in input data, ocean models tend to drift from reality
Local fine resolution models require initial and boundary conditions which can be obtained from good quality, but coarser resolution models run by major ocean modelling centres such as Mercator Ocean International (France) or Met Office (UK) via Copernicus Marine Service (CMEMS, 2021)
The Stochastic-Deterministic Downscaling (SDDA) data assimilation method detailed in the section Algorithm above was applied to the simulated child model forecast in order to create analysis for the forecasting cycle

Summary

Introduction

Due to inevitable approximations in the equations, numerical schemes, parameterisation and uncertainties in input data, ocean models tend to drift from reality. Fine resolution ocean modelling is becoming a ubiquitous practice to resolve important mesoscale and sub-mesoscale features such as eddies, fronts, boundary currents and localised upwellings which play important roles in ocean dynamics, see e.g. T. Meunier et al, 2012.) Such localised models can be run by relatively small groups due to availability of good quality ocean models such as ROMS or NEMO to the wider oceanographic community (ROMS, 2021; NEMO, 2021). Local fine resolution models require initial and boundary conditions which can be obtained from good quality, but coarser resolution models run by major ocean modelling centres such as Mercator Ocean International (France) or Met Office (UK) via Copernicus Marine Service (CMEMS, 2021).

Objectives

Results

Conclusion