The NRL Layered Global Ocean Model (NLOM) with an Embedded Mixed Layer Submodel: Formulation and Tuning

Abstract

A bulk-type (modified Kraus?Turner) mixed layer model that is embedded within the Naval Research Laboratory (NRL) Layered Ocean Model (NLOM) is introduced. It is an independent submodel loosely coupled to NLOM's dynamical core, requiring only near-surface currents, the temperature just below the mixed layer, and an estimate of the stable mixed layer depth. Coupling is achieved by explicitly distributing atmospheric forcing across the mixed layer (which can span multiple dynamic layers), and by making the heat flux and thermal expansion of seawater dependent upon the mixed layer model's sea surface temperature (SST). An advantage of this approach is that the relative independence of the dynamical solution from the mixed layer allows the initial state for simulations with the mixed layer to be defined from existing near-global model simulations spun up from rest without a mixed layer (requiring many hundreds of model years). The goal is to use the mixed layer model in near-global multidecadal simulations with realistic 6-hourly atmospheric forcing from operational weather center archives. A minimum requirement therefore is that there be no drift in yearly average SST over time. The dynamical layer densities are relaxed to climatology as a standard part of the NLOM model design, and this ensures that the temperature just below the mixed layer provided to the mixed layer submodel does not drift. The density relaxation below the mixed layer does not significantly dampen anomalies even on interannual timescales because the anomalies are largely defined by layer thickness variations. When combined with calculating the latent and sensible heat flux using model SST, this is sufficient to keep SST on track without any explicit relaxation to the SST climatology. The sensitivity of the global ocean model to the choice of free Kraus?Turner parameters in the bulk mixed layer model is investigated by undertaking a tuning exercise to find a single set of parameters that provides a realistic SST from realistic atmospheric forcing over as much of the global ocean as possible. This is done by comparing the monthly Comprehensive Ocean Atmosphere Data Set (COADS) SST climatology to monthly averages from the model using a set of statistical metrics. A single set of mixed layer parameters is reported that gives excellent agreement with the SST climatology over most of the global ocean. The actual parameter values are probably specific to this coupled system, but the same methodology can be used to tune any mixed layer model with free parameters.