Importance of TOPEX/Poseidon/Jason data to improve the coupled ocean-atmosphere modeling of El Nino

Claire Perigaud⁽¹⁾ , Jean-Philippe Boulanger⁽²⁾ , and Serena Illig⁽¹⁾

⁽¹⁾ JPL, 4800 Oak Grove Drive, Pasadena, CA, 91109, United States
⁽²⁾ LODYC, 4 place Jussieu, 75252 Paris Cedex 05, France

Abstract

More than a decade of research where altimetric data have been used in ocean models that are either FORCED or COUPLED with the atmosphere have led to the following conclusions: 1) The events simulated in COUPLED experiments are sensitive to 1 cm sea level accuracy whereas they are not in FORCED. 2) It is not necessarily because of model errors that COUPLED experiments are sensitive, it can be because of physical reasons that matter in reality. We choose two examples to demonstrate these important conclusions.

The sensitivity of the simulated Nino3 SST index to off- equatorial sea level changes is always smaller than 1 Degree C in FORCED and reaches 4C in COUPLED. In FORCED experiments, the observed atmosphere which controls the model experiment has equatorial trade winds that relax/strengthen and make the ocean simulate warm/cold events regardless of what happens beyond the equator. But in reality, the off-equator plays a role in the recharge/discharge of the equatorial oceanic heat content for a (warm event/relaxed trade winds) to develop. Indeed, TP-Jason data show that the equatorial recharge by the off-equator is not symmetric as previously thought: a warm event charges the equatorial Pacific, it does discharge the South off-equator as known, but it recharges the North, making the sea level rise in the North because of the ITCZ equatorward migration. TP-Jason data are then used in the model outside the equator as the only source of data controlling the COUPLED behavior, and simulations from this "almost-data-free" experiment successfully reproduce the warm/cold events that have been observed since 1992.

Similarly, experiments are sensitive to the eastern boundary conditions of the model by more than 4C in COUPLED and less than 1C in FORCED. So TP-Jason data are then used in the COUPLED model to control the variations of the eastern boundary conditions only. This "almost-data-free" coupled experiment allows to reproduce the series of warm/cold events that have been observed since 1992 too. Experiments similarly controlled by altimetry at the western boundary are also performed, allowing the Indo- Pacific connection. The first experiment we performed was as successful as above when we used the one baroclinic version of the coupled model. But the experiment is not successful when performed with the model version which has two baroclinic modes and a vertical diffusion term added among other improvements. We need to improve our use of TP-Jason data to control the boundary conditions in case of 2 baroclinic modes, and also to refine the parameterization of the vertical mixing term which plays a key role on ENSO behavior. Absent in most common intermediate coupled models, this term deserves more attention. All these experiments demonstrate that we need altimetry with 1cm accuracy to make progress in understanding how ENSO is actually sensitive to COUPLED physical processes that have been neglected so far because they cannot be detected in FORCED ocean simulations.

Workshop presentation