Links between annual, Milankovitch and continuum temperature variability

TLDR: The annual, Milankovitch and continuum temperature variability together represent the response to deterministic insolation forcing, providing insight into the mechanisms governing interannual and longer-period climate variability.

Abstract: Climate variability exists at all timescales-and climatic processes are intimately coupled, so that understanding variability at any one timescale requires some understanding of the whole. Records of the Earth's surface temperature illustrate this interdependence, having a continuum of variability following a power-law scaling. But although specific modes of interannual variability are relatively well understood, the general controls on continuum variability are uncertain and usually described as purely stochastic processes. Here we show that power-law relationships of surface temperature variability scale with annual and Milankovitch-period (23,000- and 41,000-year) cycles. The annual cycle corresponds to scaling at monthly to decadal periods, while millennial and longer periods are tied to the Milankovitch cycles. Thus the annual, Milankovitch and continuum temperature variability together represent the response to deterministic insolation forcing. The identification of a deterministic control on the continuum provides insight into the mechanisms governing interannual and longer-period climate variability.

Figures

Figure 2 | Patch-work spectral estimate using instrumental and proxy records of surface temperature variability, and insolation at 658 N. The more-energetic spectral estimate is from high-latitude continental records and the less-energetic estimate from tropical sea surface temperatures. Highlatitude spectra are estimated from Byrd, Taylor and GISP2 ice-core d18O; Vostok and Dome C ice-core dD; Donard lake varve thickness; the Central England Time-series (CET); and the Climate Research Unit’s (CRU) instrumental compilation. From low latitudes we use ODP846 marine sediment-core alkenones; W167-79, OCE205-103, EW9209-1, ODP677 and ODP927 calcite d18O; PL07-39 and TR163-19 Mg/Ca; ODP658 foram assemblages; Rarotonga coral Sr/Ca; and the Climate Analysis Center (CAC) and CRU instrumental compilations. Temperature spectral estimates from records of the same data type are averaged together (d18Oice and dDice are also averaged). Power-law estimates between 1.1–100 yr and 100–15,000 yr periods are listed along with standard errors, and indicated by the dashed lines. The sum of the power-laws fitted to the long- and short-period continuum are indicated by the black curve. The vertical line-segment indicates the approximate 95% confidence interval, where the circle indicates the background level. The mark at 1/100 yr indicates the region mid-way between the annual and Milankovitch periods. At bottom is the spectrum of insolation at 658N sampledmonthly over the past million years plus a small amount of white noise. The vertical black line indicates the 41-kyr obliquity period.