Improved High Resolution Ocean Reanalyses Using a Simple Smoother Algorithm. (24th November 2021)

• Record Type: Journal Article
• Authors: Dong, Bo; Haines, Keith; Martin, Matthew
• Is Part Of: Journal of advances in modeling earth systems. Volume 13:Number 12(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-11-24
• Subjects: ocean reanalysis -- data assimilation -- Kalman smoother -- AMOC -- ocean heat content -- salinity; Geological modeling -- Periodicals; Climatology -- Periodicals; Geochemical modeling -- Periodicals; 551.5011
• DOI: 10.1029/2021MS002626
• Languages: English
• ISSNs: 1942-2466

Abstract

Abstract: We present a simple smoother designed for smooth data adjustments in sequentially generated reanalysis products by utilizing knowledge of future assimilation increments. A decay time parameter is applied to the smoother increments to account for memory decay timescales in the ocean. The result is different from simply time smoothing the reanalysis itself as only the increments are smoothed so the reanalysis product can retain high frequency variability that is being internally generated by the model and the atmospheric forcing. The smoother is applied first to the Lorenz 1963 model and then to the daily Met Office GloSea5 Global ¼° ocean reanalysis during 2016. Results show significant improvement over the original reanalysis in the 3D temperature and salinity state and variability, as well as in the sea surface height (SSH) and ocean currents. Comparisons are made directly against temperature, salinity and SSH observations, as well as independent 15 m drifter velocities. The impact on the time variability of conservative quantities, particularly ocean heat and salt content, as well as kinetic energy and the Atlantic Meridional Overturning Circulation (AMOC), is also demonstrated. Plain Language Summary: The ocean observing system is sparse and ocean circulation is slow relative to the atmosphere, giving the system a longer memory of state properties. Therefore an ocean reanalysis could seek to use more "future data" to help produce the best historical ocean state reconstruction for a given day. Conventional sequential assimilation approaches used operationally make no use of "future" data outside the window of current data analyzed for forecasting. Here we introduce a new approach for application to large global ocean reanalysis systems that have only been run in "forward mode" (using past data), using the history of stored data increments to produce a more physically plausible time‐evolving ocean state with smoother temporal adjustments toward the available observations. Key Points: A novel temporal smoother algorithm is designed for improving ocean reanalysis products by utilizing the stored history of assimilation increments High frequency variability internally generated by the model and the atmospheric forcing is retained in the output The smoother significantly improves the reanalysis temperature, salinity, velocity, and sea surface height fields