The enigmatic 1470yr cycle

Any effort to explain ocean-ice interactions during the last glaciation must take into account the puzzling and widespread ca. 1470-yr cyclicity found in glacial records from ice cores and in marine and terrestrial sediments in both hemispheres (see Alley et al., 2001). In records of S18O from Greenland ice, the cycle appears as a sharp peak in spectral energy that suggests a truly periodic process with a period of about 1470yr (Mayewski et al., 1997; Grootes & Stuiver, 1997). Recently, Rahmstorf (2003) demonstrated that the spacing of D-O cycles in the GISP2 ice core from Greenland was exactly 1470yr or exactly even multiples of 1470

yr, thereby tying the D-O oscillations to the 1470-yr cycle (Fig. 24.5a & b). Rahmstorf argued from those findings that D-O events are discrete events paced by the 1470-yr cycle rather than being discrete cycles themselves. He concluded that internal processes in a complex non-linear system such as the Earth's climate system would not likely produce such a regular oscillation.

If Rahmstorf (2003) is correct, the implication is that a regular, external process must underlie D-O cycles. Orbital cycles would meet that criterion, but there is no known orbital cycle with a period of 1470 yr. Regardless of the mechanism, however, the regularity of the 1470-yr cycle could provide the weak periodic forcing required by stochastic resonance models, such as proposed by Alley et al. (2001). It should be emphasized though that ice-core age models are undergoing revision, and improved chronologies in the future may reveal a much less regular oscillation within the 1000 to 2000yr band.

6420-2 0

1470 yrs

1470 yrs

2940

5880

8820 Time in years

11760

14700

17604

Figure 24.6 The record of haematite-stained grains from the Holocene in cores MC52 and VM29-191. Vertical lines mark out an exact 1470-yr cycle. The dashed line is the result of filtering the haematite-stained grain record with a Gaussian filter centred on 1470yr and bandwidth of 1000 to 2000yr. Note that although the Holocene cycle pacing is not exactly 1470yr, it falls within the same range as its glacial counterpart in Fig. 24.5c.

2940

5880

8820 Time in years

11760

14700

17604

Figure 24.6 The record of haematite-stained grains from the Holocene in cores MC52 and VM29-191. Vertical lines mark out an exact 1470-yr cycle. The dashed line is the result of filtering the haematite-stained grain record with a Gaussian filter centred on 1470yr and bandwidth of 1000 to 2000yr. Note that although the Holocene cycle pacing is not exactly 1470yr, it falls within the same range as its glacial counterpart in Fig. 24.5c.

Further insight into the nature of the 1470-yr cycle came from efforts by Bond et al. (1999) to produce a detailed record of IRD variations during the entire glacial cycle using measurements of two of the petrological tracers of IRD, haematite-stained grains and Icelandic glass (Fig. 24.2). They found that increases in both tracers (and, hence, increases in IRD) were coherent and occurred in an irregular non-periodic cycle with a mean value of 1470yr and a dispersion around that mean (at 1 o) of ± 500yr (Fig. 24.5c). A true 1470-yr period could be present, though, and obscured by the large errors in constructing age models for the marine records. The cycle was not confined to just the cold phases of D-O events; rather, it punctuated the entire glacial record, including the LGM and the interstadial phases of the D-O cycles. It would appear then that the IRD maxima were not tied to D-O cycles, but instead could be a manifestation of the underlying 1470-yr cycle identified by Rahmstorf (2003).

Was this article helpful?

0 0

Post a comment