Using NFLUX Analysis Fields for Regional Ocean Surface Bias Corrections

Atmospheric heat fluxes at the ocean surface are used to force operational ocean models. The required heat fluxes include latent heat flux, sensible heat flux, solar radiation, and longwave radiation. The turbulent heat fluxes are typically calculated within the model using state parameter input fields, namely surface air temperature, specific humidity, and wind speed. The state parameter and radiative heat flux forcing fields come from atmospheric numerical weather prediction (NWP) models, which often have biases at the surface. Monthly or yearly bias corrections can be identified and applied to these fields prior to starting the model forecast; however, identifying the bias corrections can be a rather time intensive process. As an alternative, we present a near-real-time bias correction method that uses satellite-based analysis fields produced by the Naval Research Laboratory (NRL) ocean surface flux system (NFLUX).

NFLUX is a complete end-to-end data processing, automated quality control, and 2D assimilation system that produces near-real-time three-hourly satellite-based global gridded analysis fields of surface air temperature, specific humidity, scalar wind speed, solar radiation, and longwave radiation. NFLUX primarily uses data from polar orbiting passive microwave sensors. These sensors provide brightness temperature data which is transformed into swath-level surface state parameter estimates using multi-linear regression algorithms. Satellite-derived atmospheric profile products from the Microwave Integrated Retrieval System (MIRS) serve as the primary inputs to the rapid radiative transfer model for global circulation models (RRTMG) to provide swath-level surface radiative flux estimates. The swath-level estimates undergo an automated quality control process and are then assimilated with a background atmospheric model field to produce the NFLUX 2D global gridded analysis fields.

The proposed near-real-time bias correction of NWP forcing fields for ocean models is applied during both the hindcast period and the forecast period. During the ocean hindcast period the NFLUX analysis fields are used directly to determine the surface bias correction to be applied to the NWP model forcing fields. The bias correction can then be extended to the forecast period by separating the bias correction into a long term (persistent) and short term (weather-dependent) correction. The long term bias correction is applied over the entire forecast time while the short term bias correction is applied with a decorrelation time of one to two days. To determine the averaging window for the long term bias corrections, the NFLUX analysis fields were compared to the Navy Global Environmental Model (NAVGEM). The NFLUX minus NAVGEM daily average time series over one year was used to determine a global long term bias correction averaging window separately for each parameter. The global long term bias averaging window is identified as 10 days for specific humidity, 11 days for air temperature, and 5 days for wind speed, longwave radiation, and shortwave radiation. The short term bias correction can then be determined by subtracting the long term bias (calculated using the global long term bias averaging window) from the NFLUX minus NWP daily average time series.

In this study, the NWP model of interest is the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS®), which provides surface forcing fields to regional ocean models, but does not presently include a bias correction. In particular, we run a twin experiment for a northeast Pacific domain to test the application of NFLUX analysis fields to determine and apply daily bias corrections as described above. The long term bias averaging windows calculated from the global model comparisons (NFLUX minus NAVGEM) are used for the regional experiments. We compare measures of forecast skill with a focus on upper ocean temperature and sound speed structure.