The Nature of the Macroweather-Climate Scaling Break in Holocene Climate

Abstract:

There exists a large body of literature which demonstrates the existence of long-range memory scaling in climate time series, but there are very few papers that attempt to explain its origin. Huybers and Curry (2006) claim an apparent higher scaling exponent β>1 on time scales greater than a century, corresponding to a non-stationary time series. Lovejoy et al. (2013) also claim a break in the scaling, and denotes the stationary regime with β~0.1 on the time scales 10 days <τ<10² yr the macroweather regime and the non-stationary time-series on the scales τ>10² yr the climate regime.

By closer scrutiny, however, the break in scaling for ice-core records from Greenland and Antarctica seems to be a result of indiscriminate use of paleoclimatic records over periods that include both glacial and interglacial climate. By analysis of these time series for glacial and Holocene climate separately, the scale break in the Holocene is shifted to τ~10³ yr, while it remains at τ~10² yr during the last glacial maximum, and is associated with the DO events. Moreover, the scaling for 10³ < τ<10⁴ yr is still stationary with β ~0.8. The scale break observed in some multiproxy data at τ~10²  yr for global or hemispheric mean temperatures for the last two millennia is associated with an oscillation of period approximately 10³ yr, involving the MWP and LIA. This oscillation can be attributed to a combination of volcanic and solar forcing. If we correct for this forced oscillation we obtain a consistent scaling with β~0.8 up to τ~10³ yr, which match perfectly onto Holocene ice-core scaling for τ>10³ yr. The low β ~0.1 observed for ice-cores at scales τ<10³ yr is typical for local, continental temperature records, while the multiproxy scaling with β ~0.8 is typical for global temperature scaling in the Holocene. Thus, with exception of the millennium-period oscillation observed in global-scale multiproxy records there is consistent macroweather scaling in global temperature records with β~0.8 on all scales above the weather regime throughout the entire preindustrial Holocene.

In the industrial period an apparent scale break at τ~10¹ yr is due to the anthropogenic trend. After removal of this trend there remains an effect of a global oscillation of period ~60 yr associated with the AMO, but not really a break in scaling.