Heliomagnetic and geomagnetic activity cycles on millennial timescales

We report 17 new solar activity cycles with periodicities ranging from 1,700 to 18,000 years. Using Bayesian spectral analysis we determined for each cycle period, amplitude and phase for the past 145,000 years. These results were obtained by analyzing ¹⁰Be in the 2775 m long EDML ice core from Dronning Maud Land, Antarctica. In addition, we have augmented the ¹⁰Be data with synchronous ice accumulation and δ¹⁸O data obtained from the same ice core; with paleomagnetic dipole moment data derived mainly from ocean sediments; and with calculated global insolation data. There is a close agreement between the ¹⁰Be periodicities in the first and the last 70 kyr intervals. We have developed a new analytical technique we call the “frequency domain differentiation technique” (FDDT) that distinguishes between the periodicities due to (a) variations in the cosmic ray intensity and (b) climate effects related to the accumulation rate. In a first step we have calculated the ¹⁰Be flux and then selected all periodicities which are common within 1 % in at least 2 of the investigated parameters (¹⁰Be concentration, ¹⁰Be flux, accumulation rate, δ¹⁸O, and insolation). Using these data we identified a total of 42 statistically significant periodicities. 20 of them are found in the ¹⁰Be concentration which we separated into 3 groups of origins: Four due to ¹⁰Be production only (group 1: 18012, 6508, 5782, 3833 yr), thirteen being a combination of production and accumulation related (system) effects with the production components being generally larger than the system components (group 2: 15503, 13837, 8441, 7722, 7136, 4695, 4556, 4110, 3961, 3292, 3104, 2025, 1680 y), and three being related to system effects only (group 3: 23971, 21101, 11637 y). In a second step we used independent paleomagnetic information to distinguish between heliomagnetic and geomagnetic periodicities in the ¹⁰Be production. We find an overall good agreement between the ¹⁰Be flux and the dipole moment in the time domain. In the frequency domain there is evidence for system effects in the geomagnetic data. Due to inconsistencies between the paleomagnetic records we do not draw conclusions regarding a potential geomagnetic origin of the ¹⁰Be cycles found in the EDML ice core.