Internal variability and external forcings in the ocean–atmosphere multidecadal oscillator over the North Atlantic

• Published in: Climate Dynamics
• Main Authors: Ribera Rodríguez, Pedro, Ordónez, Paulina, Gallego Puyol, David, Peña Ortiz, Cristina
• Format: Article in Journal/Newspaper
• Language: English
• Published: Springer Link 2020
• Subjects: Atlantic multidecadal oscillation (AMO); SST reanalyses; Twentieth century reanalysis; Internal climate variability; Blocking; NAO; MTM-SVD; External forcings; North Atlantic
• Online Access: https://hdl.handle.net/10433/20055; https://doi.org/10.1007/s00382-020-05300-8

In this paper, we generalize the concept of “external forcing” to include any mechanism that modulates the long-term evolution of a meteorological variable but is not directly related to the internal variability of the climate system. Applying this concept, the corresponding ‘external forcings’ are removed from several long record datasets of oceanic and atmospheric variables at the surface in the North Atlantic. We perform a multivariate analysis in the frequency domain over both the original data felds and the new ‘internal variability’ felds. This multivariate analysis is based on a MultiTaper Method-Singular Value Decomposition (MTM-SVD). It is noteworthy that, after the removal of the external forcings, there is an almost perfect alignment of the main multidecadal oscillatory band (f=0.21 cycles/decade) with all the spectra of the analysed felds. This alignment was not observed before the external forcings were removed. Particularly striking is the case of the sea level pressure (SLP), which shows a notable variation in its oscillation period despite the fact that this variable has traditionally been considered to be unafected, at global scale, by any external forcing. The external forcing in the SLP records is very probably caused by the scarcity of the observed data during the frst hundred years of the record (most evident, during the earliest decades), by the spatial distribution of those observations and, possibly, by the assimilation model employed to build those long record datasets. When we analysed the relationship between the ocean and the atmosphere using this approach, we found strong evidence of a cyclic behaviour in which oceanic conditions modulate the atmospheric variability, with a lead time of up to about 10 years. Departamento de Sistemas Físicos, Químicos y Naturales