Global Implications of Terrain-Induced Bias in Paleomagnetic Data From 3D Forward Modeling of Mt. Etna

Abstract

In paleomagnetism, records of paleosecular variation (PSV) on centennial timescales are often based on magnetic measurements from lava flows. However, samples from recent flows that cooled in known magnetic fields often fail to produce the expected paleomagnetic directions, and samples from the same flow often show large within- and between-site dispersion. Direct measurements of local magnetic anomalies in volcanic terrain showed that the topography of the underlying lava flow distorts the magnetic field. Such Magnetic Terrain Effects (MTEs) are subsequently recorded by overlying flows and can account for much of this dispersion. Here we present a data-independent numerical model capable of simulating topographies and predicting magnetic fields over highly magnetized rocks. Using this Python-based forward model and high-resolution Digital Elevation Models, we replicate field measurements taken on Mt. Etna. The computed anomalies often match the location and shape of the measured results, showing intensity deviations up to 14 $\mu T$, inclination deviations up to $12°$, and declination deviations up to $14°$, compared to the measured deviations of 13 $\mu T$, $19°$, and $21°$, respectively. Our findings suggest that MTEs must be considered in paleomagnetic interpretations: while distinguishing between MTE and PSV is nearly impossible in individual measurements when the underlying topography is unknown, MTEs introduce systematic bias that can be estimated using our model. Furthermore, we demonstrate that MTE-induced bias varies globally and persists despite improved sampling strategies, necessitating the development of correction methods to address the limitations of lava flows as paleomagnetic recorders.