Channell et al (2009) have produced a new geomagnetic paleointensity record which can be used in ACE to examine the effect of past changes in geomagnetic intensity on cosmogenic nuclide production. The Earth’s magnetic field deflects cosmic rays, so that changes in field strength can significantly change production rates. A comparison of the new Channell et al paleointensity record (blue) with the well known SINT-2000 (red, Valet et al 2005) is shown in their Figure 7 and reproduced below:
Click on image to expand
For ages less than 41 ka, the Channell et al PISO-1500 record is about half that of SINT-2000, which suggests that the differences in these two records could be significant for cosmogenic nuclide dating. Unfortunately the PISO-1500 record only begins at 5 ka BP, and to use it we require a continuous record from the present day. Extrapolation is always problematic, however Channell et al cite a ‘recent’ intensity of 7.5 x 1022 Am2, so for this record we use this value as a present day intensity and then linearly interpolate over the 0-5 ka period of missing data. This approach gives the intensity values a similar shape to the recent intensity changes suggested by SINT-2000 (see the above figure).
To examine the effect of this new dataset, we have imported it into ACE and compare results for 36Cl using the original Guyodo and Valet 1999 dataset (described in Data Collections). After following the standard module creation procedures, we make two different experiments, both with Desilets and Zreda scaling (2003). This scaling accounts for changes in geomagnetic intensity:
click on image to expand
The difference between the two experiments shown above is the choice of geomagnetic intensity record. After calibrating these experiments, we find that for the Guyodo and Valet record the reduced chi-square is 1.32, suggesting a reasonable match of calibrated with independent inventories:
Click on image to expand
The same calibration of 36Cl HLSL production rates conducted using the Channell et al 2009 record has a reduced chi-square of 1.71:
click on image to expand
so it could be said that all other things being equal, the Guyodo and Valet paleointensity record produces sample ages of the 36Cl calibration data set closer to the independent ages than the Channell et al paleointensity dataset. More important is the fact that calibrated HLSL production rates are systematically lower for Channell et al reconstruction compared to the Guyodo and Valet and Channell reconstruction:
Click on image to expand
For Calcium Spallation and low-energy Neutron Capture production rates are 5% lower, and for Potassium Spallation the production rate is 11% lower. The Channell et al paleointensity record suggests a weaker geomagnetic field over most of the age span of the 36Cl calibration database (0-49 ka BP), which according to ACE means lower production rates. Why is this? The following plot uses the two experiments above and calculates ages of the calibration database used here to identify where major differences arise:
Click on image to expand
This plot has been created by dating the calibration database using the two experiments made above, sorting the output first by id and then by age, and then clicking plot selected attributes in the Sample Browser:
The different experiments then plot on top of each other, and only the differences will show up. In this case it appears that the major differences are for the Metor Crater samples MC-1, MC-3, MC-4 and MC-5. As these are the oldest samples in the calibration database (49 ka), it is reasonable to expect that they are most sensitive to changes in the paleointensity record.
These results suggest that to further explore whether the Channell et al (2009) paleointensity record should be used instead of the older Guyodo and Valet (1999) record, the sensitivity of results for other nuclides, scalings and calibration databases should be examined first. As with all other datasets used in ACE, details regarding the dataset publication, source and modification are available in the Data Collections section together with a csv file to download.