DEFLECTION OF PALEOMAGNETIC DIRECTIONS DUE TO MAGNETIZATION OF THE UNDERLYING TERRAIN

In the summers of 1989 and 1991 we made 344 near-ground level measurem ents of the ambient geomagnetic field above recent basalts on the isla nd of Hawaii using a three-component fluxgate magnetometer. We studied 12 surface features, including a lava pond, lava channels, long tilte d blocks, smooth sloping surfaces, two fissures, and a deep U-shaped r oad cut. We observed substantial differences (up to 20 degrees) betwee n the observed and expected (International Geomagnetic Reference Field , IGRF) magnetic field directions over these features except those com posed of shelly pahoehoe and a flat (horizontally) thin lava pond. We also observed inclinations that were systematically shallower than the IGRF field by up to 5 degrees. We show that these shallower inclinati ons can be explained by the magnetization of the regionally sloping su rface of the southern side of the island. We found that all of the obs erved inclination deflections can be explained by simple two-dimension al models which assume uniform induced and remanent magnetization para meters in the local terrain. Our observations imply that the inclinati on deflections cannot be corrected without a complete knowledge of the preexisting terrain and the remanence in the underlying flows upon wh ich the lavas cooled. Since this information is rarely available, it i s difficult or impossible to discriminate between dispersion of paleom agnetic directions caused by the magnetic terrain effect and dispersio n due to other factors such as paleosecular variation (PSV). We theref ore conclude that PSV dispersion parameters cannot be accurately deter mined from paleomagnetic measurements on highly magnetic volcanic flow s. We also suggest that some of the geomagnetic excursions inferred fr om. measurements on volcanic rocks may be at least in part due to the magnetic terrain effect. It is unnecessary to invoke ad hoc mechanisms such as elastic, block, or crustal rotations, distortion of the top c rust, or flow deformation to explain the large between-site dispersion s or inclination anomalies observed in many of the paleomagnetic data from volcanic rocks. Our observations also bring into question the gen eral reliability of paleomagnetic pole positions inferred from volcani c rocks, as a systematic inclination deflection due to local and regio nal slopes and irregular terrain, such as those we observed, would lea d to a corresponding error is the inferred paleolatitude. The magnetic terrain effects also offer alternative explanations for anomalous pal eomagnetically inferred plate motions.