The GloSea5 Ocean and Sea Ice Analysis with Global Ocean 5.0

Andrew Peterson1, Jennifer Waters2, Chris D Roberts1, Laura C Jackson1, Ed W Blockley2, Maria Valdivieso3, Matthew Martin2 and Michael James Bell2, (1)Met Office Hadley Centre, Exeter, United Kingdom, (2)Met Office, Exeter, United Kingdom, (3)University of Reading, Department of Meteorology, Reading, United Kingdom
Abstract:
The Met Office seasonal forecast system, GloSea5 produces an ocean re-analysis for the period 1990-2014 using the same forecast ocean assimilation model (FOAM) system used for operational ocean forecasting at the Met Office. Using NEMOVAR (3D-VAR) ocean assimilation and making use of sub-surface observations of ocean temperature and salinity profiles (EN4), along with in-situ and satellite observations of sea surface temperature, satellite altimeter observations of sea level anomalies and satellite observations of sea ice concentration, an ocean and sea ice state estimation is produced. This ocean and sea ice analysis is then used to initialize these respective components of the coupled historical forecasts (hindcasts) of the seasonal forecast system. In turn, the real time FOAM analysis is used to initialize the ocean and sea ice components of the seasonal forecast along with coupled versions of the 7-day ocean forecast for MyOcean.

Besides its use in a leading edge seasonal forecast system, this ocean and sea ice re-analysis is an important tool in the estimations of past ocean states. The GloSea5 ocean and sea ice analysis participated in the GSOP/GODAE ocean re-analysis intercomparison project (ORA-IP) and continues to be active in other ocean analysis projects such as the evaluating ocean synthesis EU-COST action (EOS-COST). 

Since February 2015, the operational version of GloSea5 has used the joint NERC/Met Office global ocean 5.0 (GO5.0) and global sea ice 6.0 (GSI6.0) version of the ocean and sea ice analysis based on NEMO/CICE. In addition to improvements in turbulent kinetic energy vertical mixing schemes used in this version of NEMO, additional improvements have been made to the albedo and surface roughness parameters used in CICE. The NEMOVAR assimilation scheme has also been updated to include two length scale background error covariances to better represent the sparse profile observations in the analysis.

Here we will present some results from the GloSea5 ocean and sea ice analysis, GO5.0/GSI6.0, demonstrating some of the improvements of the system over past versions. In particular, its performance in observing the variability of the heat and salt content of the ocean, surface fluxes, the Atlantic meridional overturning, and meridional heat transport will be amongst the characteristics presented.