Comment on "On sharp spectral lines in the climate record and the millennial peak" by Carl Wunsch

the record of calcium concentrations from the Greenland Ice Sheet Project 2 (GISP2) ice core [Mayewski et al., 1997] extending from the present (1985) back in time 110,000 years (110 kyr B.P.). Also shown is the band-pass component filtered from the series using a digital filter with central period of 1460 years applied to a uniform 50 year resampling of the series. This component represents 7% of the variance in the resampled calcium series. Figure l c exhibits a 6000 year segment of the calcium series dated from 30 to 36 kyr B.P. As can be seen, the segment includes a portion of the overall record containing some of the Dansgaard/Oeschsger (D/O) events [dohnsen et al., 1992] for which the 1460 year periodic component oscillates with an amplitude of 100 ppb. It is this feature of the record which Wunsch believes to be a mathematical artifact due to the aliasing of an annual signal at either a tropical year, 365.2422 days, an anomalistic year, 365.2596 days, or a combination of the two. We do not support this conclusion for both scientific and mathematical reasons. The mathematical foundation of the author's argument is clear. An oscillation with a period of, for example, a tropical year, when sampled at intervals containing a uniform number of common years (365 days), will appear (aliased) with a period near 1500 common years. Figure 2 illustrates this phenomenon. Figure 2a shows the result of sampling a tropical year periodic oscillation (amplitude of 100 ppb) at a uniform 50 year interval for 6000 years of an artificial record (gen