Evaluating the performance of an ensemble forecast system in predicting Loop Current Eddy separation in the Gulf of Mexico

Abstract

Forecasting mesoscale variability, such as the Loop Current (LC) growth and Loop Current Eddy (LCE) shedding in the Gulf of Mexico, poses challenges due to the large uncertainty in initial conditions and the development of non-linear instabilities from LC-cyclone interactions, which are crucial for LC/LCE separation. An ensemble forecast system may account for this uncertainty and filter out unconstrained scales and thereby extend the predictability of the mesoscale variability. A 32-member ensemble forecast system is employed to investigate the predictability of LC/LCE separation, with a specific focus on the December 2019 – March 2020 period. The forecasts demonstrated predictability of LC/LCE separation out to 7 to 13 weeks. During this period, significant changes occurred in the LC. The LC transitioned from being an extended LC to LCE separation on January 27, 2020. Subsequently, in March 2020, the LCE deformed and nearly split into two separate eddies. Detailed analyses of individual forecasts during this timeframe revealed that these transformations were influenced by two main interactions: (a) the interaction between the LC and a cyclone along its eastern edge, which caused the LC/LCE separation, and (b) the interaction between the LCE and a cyclone along its northern side, leading to the potential splitting of the LCE. These interactions were further intensified by the coupling between surface and deep cyclones. The validity of these findings is supported by a variety of observations, including drifters, current meters, and sea surface height, as well as verifying analysis.