Statistical Analysis of Palaeomagnetic Data from the Last Four Centuries: Evidence of Systematic Inclination Shallowing in Lava Flow Records

Unbiased analysis of geomagnetic data sets and comparison of historical data with paleomagnetic and archeomagnetic records

“…This is because meaningless comparisons are excluded by the data selection and weighting procedure described in section 2. Similar results have been obtained in the past using a field model as the basis for comparison of direct and indirect records to overcome temporal and spatial mismatches [ Pavón‐Carrasco et al , ; Suttie et al , ]. While both approaches appear to be equally effective, subtle biases possibly introduced using field models for bridging mismatches can be excluded in our case.…”

“…Strong contribution of these materials—as well as from materials, which are not specified in the database—means that systematic inclination shallowing affects the entire data set, even if other materials, such as bricks, appear to be immune (Table ). Inclination shallowing in lavas has been previously reported [e.g., Pavón‐Carrasco et al , ], and possible causes are discussed in section 4.1.…”

“…While individual comparisons between indirect and historical records are evidently dominated by random error sources (e.g., Figure a), the whole database can be used to detect possible systematic biases through a statistical analysis of values. If such values represent independent realizations of a single random variable with given statistical distribution, deduced, for instance, from histogram plots [e.g., Pavón‐Carrasco et al , ], statistical tests can be used to make inferences about the expected value μ X of (e.g., a t test for the mean as in Suttie et al []). However, the conditions required for these tests to be valid are not necessarily fulfilled by geomagnetic data sets, which are notoriously characterized by an irregular distribution of data in space and time.…”

“…Suttie et al [] compared archeointensity and paleointensity data with the gufm1 field model [ Jackson et al , ] derived from historical records and found a systematic deviation for cooling rate corrected intensities. Pavón‐Carrasco et al [] found a systematic inclination shallowing in volcanic rocks when comparing these records with the historical field model over the last 400 years.…”

“…Inclination values are transferred to the reference site using the difference of inclinations generated by an axial dipole. After Pavón‐Carrasco et al [] with additional historical records.…”