CoAuthors

George Halliwell (NOAA/AOML, USA); Michael Mehari (UM/CIMAS and NOAA/AOML, USA); Jili Dong (NOAA/NCEP/EMC, USA); Villy Kourafalou (UM/RSMAS, USA); Robert Atlas (NOAA/AOML, USA); HeeSook Kang (UM/RSMAS, USA); Ioannis Androulidakis (UM/RSMAS, USA)

Altimetry is the backbone of the present global ocean observing system, and its components (namely, altimeter satellites) are constantly evolving. Altimetry data are routinely combined with data from other satellites and in situ platforms by operational monitoring and prediction centers. However, a quantitative assessment of the benefit of the current and future components of the observing network is necessary to assess and improve its performance. The ocean Observing System Simulation Experiments (OSSE) system developed by the Joint Ocean Modelling and OSSE Center (OMOC) of NOAA/AOML, CIMAS, and RSMAS, University of Miami is designed and rigorously validated to ensure that credible observing system impact assessments are obtained. This modeling system covers the North Atlantic hurricane region, and has been used to perform Observing System Experiments, where actual data are assimilated into the ocean model, to assess the performance of the existing observation network. It has also been used to perform OSSEs, in which the data that are assimilated are extracted from an independent simulation that represents the true ocean. These OSEs and OSSEs were aimed at quantifying the ability of the observation network to constrain the upper ocean structure and heat content, which is essential for correct hurricane prediction. These experiments illustrate the key role played by altimetry data for that purpose, in particular for its ability to constrain a large portion of the ocean mesoscale spectrum. OSSEs were also performed during Hurricane Gonzalo (2014) to investigate how the deployment of ocean gliders and/or airborne ocean profilers, in addition to altimetry and other components of the observation network, can improve the overall performance of the network. Ocean profiles that sample both temperature and salinity down to 1000m are more effective for correcting ocean mesoscale features than shallower profiles of temperature alone obtained from typical XBTs and thermistor chains. The error reduction provided by the assimilation of observations from any individual instrument is primarily confined to a diameter smaller than 2° surrounding the measurement location. Large spatial coverage with multiple instruments is, therefore, necessary to reduce ocean initialization errors over a region broad enough to potentially have a significant impact on storm intensity forecasts. A coupled ocean-atmosphere prediction system initialized by ocean analyses from our OSE-OSSE system during the same hurricane allowed investigating the impact of these observations on the hurricane forecast. Assimilation of observations corrected the upper-ocean heat content ahead of the storm, enabling the coupled model to more accurately predict the heat flux from ocean to atmosphere that fuels the storm. The OSE-OSSE results demonstrate that in situ ocean observations combined with altimetry will play an important future role toward improving intensity forecasts of tropical cyclones.