Meeting Documents

Using an Ocean State Estimate to Improve Attribution and Prediction of Sea Level along the U.S. Gulf Coast

Abstract

Predictions of regional sea level (SL) variability on subseasonal to decadal timescales have become increasingly relevant for hazard mitigation and coastal planning. Yet physics-driven coupled seasonal prediction and climate models struggle with biases and secular drifts in SL, while data-driven approaches are limited by the spatiotemporal coverage of observations, especially in the subsurface ocean. The Estimating the Circulation and Climate of the Ocean (ECCO) state estimate provides a unique opportunity to use both physics-driven simulations and observations to improve understanding and predictions of SL variability (SLV). By assimilating ocean observations while conserving volume/mass, heat, and salt in the ocean interior, the ECCO state estimate avoids spurious drifts and reduces model bias. Furthermore, the adjoint capability of ECCO allows SL reconstructions that attribute SLV to local and remote forcings at the ocean surface, while also enabling synthesis with surface fluxes from seasonal and interannual/decadal prediction models. A case study in the Gulf of Mexico uses ECCO-based SL reconstructions to show that high-frequency monthly SLV is well explained by wind stress forcing on the Gulf Coast and adjacent Florida Straits. In contrast, interannual and decadal SLV has substantial contributions from wind stress in the Caribbean and along the U.S. Atlantic coast, and to a lesser extent from heat fluxes locally and in the Caribbean. By “forcing” the adjoint sensitivities with surface fluxes from seasonal (ECMWF SEAS5) and decadal (CESM) prediction simulations, not only are SL predictions often improved but biases in surface fluxes can be diagnosed, enhancing understanding of air-sea interactions that impact SL variability.