The Dominant Global Modes of Recent Internal Sea Level Variability

Abstract

The advances in the modern sea level observing system have allowed for a new level of knowledge of regional and global sea level in recent years. The combination of data from satellite altimeters, Gravity Recovery and Climate Experiment (GRACE) satellites, and Argo profiling floats has provided a clearer picture of the different contributors to sea level change, leading to an improved understanding of how sea level has changed in the present and, by extension, may change in the future. As the overlap between these records has recently extended past a decade in length, it is worth examining the extent to which internal variability on timescales from intraseasonal to decadal can be separated from long‐term trends that may be expected to continue into the future. To do so, a combined modal decomposition based on cyclostationary empirical orthogonal functions is performed simultaneously on the three data sets, and the dominant shared modes of variability are analyzed. Modes associated with the trend, seasonal signal, El Niño–Southern Oscillation, and Pacific decadal oscillation are extracted and discussed, and the relationship between regional patterns of sea level change and their associated global signature is highlighted.

Reconciling global mean and regional sea level change in projections and observations

“…This indicates the PDO decadal variability potentially accounts for a small fraction of the acceleration in GMSL over the altimeter period. This corroborates the findings of previous studies 27,28 , indicating that the decadal variability in the satellite GMSL records related to changes in TWS driven by PDO might obscure the anthropogenic acceleration of GMSL. The AR5 GMSL accelerations over the first 26 years (2007-2032) are 0.035 ± 0.052, 0.048 ± 0.048, and 0.067 ± 0.049 mm yr −2 under RCP2.6, 4.5, and 8.5, respectively ( Fig.…”

“…The observed GMSL change contains considerable interannual variability, at least partly due to changes of TWS driven by ENSO 25 , 26 , as well as decadal variability in TWS related to PDO 27 , 28 , thermal expansion, and ice-sheet mass loss 29 . The underlying mechanisms between GMSL and other large-scale climate modes (e.g.…”

Identifying ENSO-Related Interannual and Decadal Variability on Terrestrial Water Storage

“…The nested period's determination plays a critical role in the CSEOF analysis, especially when targeting specific frequencies, as in this study. Based on previous studies 21,37,38 about ENSO variability, we chose a 24-month nested period. This two-year nested period allows for the separation of ENSO's biennial oscillation (i.e., the transition between warm and cold events) and lower-frequency variability found centered in the Pacific Ocean.…”

“…EOF and CSEOF analysis decompose spatiotemporal data into the statistical modes in the order of each mode's variance. Various studies have applied CSEOF analysis to decompose climate datasets into independent modes representing natural variabilities 3,21,22,[26][27][28] . The most challenging CSEOF analysis issue is interpreting the decomposed statistical mode, and the previous studies have tried to link the decomposed mode to predefined climate indices.…”

“…While PH12 suggests a relationship between ENSO and TWS variability in the globe, questions remain regarding the degree to which the MEI represents ENSO and the appropriateness of the underlying technique itself (see Chao and Chung 20 ; hereafter CC19). PH12 was conducted on only eight years of data, and with the influence of longer-timescale natural variability during this period, the degree to which ENSO influences can be separated from other variability is challenging to assess 21 .…”