Climate-driven polar motion: 2003-2015

Abstract

Earth's spin axis has been wandering along the Greenwich meridian since about 2000, representing a 75° eastward shift from its long-term drift direction. The past 115 years have seen unequivocal evidence for a quasi-decadal periodicity, and these motions persist throughout the recent record of pole position, in spite of the new drift direction. We analyze space geodetic and satellite gravimetric data for the period 2003-2015 to show that all of the main features of polar motion are explained by global-scale continent-ocean mass transport. The changes in terrestrial water storage (TWS) and global cryosphere together explain nearly the entire amplitude (83 ± 23%) and mean directional shift (within 5.9° ± 7.6°) of the observed motion. We also find that the TWS variability fully explains the decadal-like changes in polar motion observed during the study period, thus offering a clue to resolving the long-standing quest for determining the origins of decadal oscillations. This newly discovered link between polar motion and global-scale TWS variability has broad implications for the study of past and future climate.

Keywords: Earth sciences; climatology; earth’s spin axis; hydrology; polar ice sheets; polar motion; space geodesy; terrestrial water storage.

Fig. 1. Observed pole position data.

Mean monthly polar motion excitations (black lines) derived from the observed daily values after removing semiannual, annual, and Chandler wobbles. Smoothed solutions (blue lines) reveal quasi-decadal variability in the corresponding component of the 20th-century linear trend (dashed red lines). Cyan shadows in the background cover our study period, over which the drift direction deviates (solid red lines) from the long-term linear trend.

Fig. 2. Climate-induced mass redistribution on Earth’s surface.

(A) Linear rate of change in mass (in WEH per year) during April 2002 to March 2015, derived from monthly GRACE observations and associated sea-level computations. Solutions are reproduced with different color scales for (B) the GIS, (C) the AIS, and (D) the oceans.

Fig. 3. Climate-induced polar motion.

(A) Polar motion excitations caused by four climate-related sources. (B) Total reconstructed (REC) and observed (OBS) excitations. We add global (nontidal) AOM-associated excitations (${\stackrel{.}{\mathrm{\chi }}}_1=-0.03$ and ${\stackrel{.}{\mathrm{\chi }}}_2=0.22$ mas/year) to the reconstructed solutions and remove the 20th century linear trends from the observations (see Materials and Methods). (For ease of comparison, minor smoothing is applied to the observed data.) Large positive gradients during 2005–2012 (cyan shadow), followed by negative trends, are apparent for χ₂(t), and it may be explained by analogous trends associated with TWS [see (A)].

Fig. 4. Spatiotemporal variability in TWS.

Linear trends in TWS mass redistribution (in WEH per year) during two periods (from January 2005 to December 2011 and from January 2012 to December 2014) derived from monthly GRACE observations.

Fig. 5. Origins of observed polar motion.

(A) Reconstruction and partition of polar motion during 2003–2015. Observed data have the 20th-century linear trends removed. Semimajor and semiminor axes of error ellipses are defined by the uncertainties in the magnitude and direction of the corresponding polar motion vector. For clarity, we do not show error ellipses for GICs, which have large uncertainties but very small amplitudes (see Materials and Methods) and AOM. (B) Observed (including the long-term linear trend) and reconstructed mean annual pole positions, in the excitation domain, with respect to the 2003–2015 mean position. Blue error band is associated with the reconstructed solution; red signifies additional errors that are related to uncertainty in the long-term linear trend.