.png){kind=icon}
Return
Loop current extension as an effective delayed dynamical system
.png){kind=icon}
PRE
AI
DOI：10.1007/s11071-026-13096-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The Loop Current is the dominant circulation feature of the Gulf of Mexico and exhibits pronounced variability through northward extension, retraction, and eddy shedding. We investigate whether this variability admits an effective delayed dynamical description using delayed-coordinate representations constructed from satellite-altimetry observations of Loop Current extension. Ridge regression, multilayer perceptrons, and Sparse Identification of Nonlinear Dynamics (SINDy) are used to learn delayed evolution maps. Forecast skill exceeds persistence at lead times of 30–90 days using only a few delayed coordinates. Ridge regression indicates that predictive skill rapidly saturates with delayed-state dimension, multilayer perceptrons provide only modest additional improvements, and delayed SINDy identifies sparse evolution maps with intraseasonal memory that remain stable under recursive iteration. These results show that much of the predictable variability of Loop Current extension can be represented by a small number of delayed observations and compact delayed evolution maps.
Keywords:
Loop current
Delayed-coordinate modeling
Machine learning
SINDy
Ocean predictability
Journal
IF:
6
Papers:
1.4W
Citations:
4.1W
Organization
No organization information available
Cited Papers
.png){kind=icon}
Discovering governing equations from data by sparse identification of nonlinear dynamical systems
OAAI
OAAIShareSave
ShareSaveShareSaveShareSave
View more