NORTHEASTERN PACIFIC MANTLE CONDUCTIVITY PROFILE FROM LONG-PERIOD MAGNETOTELLURIC SOUNDING USING HAWAII-TO-CALIFORNIA SUBMARINE CABLE DATA

We present results of a long-period magnetotelluric (MT) investigation of the electrical structure beneath the eastern North Pacific. The el ectric field data consist of similar to 2 years of continuously record ed voltages across an unpowered, similar to 4000-km-long submarine tel ephone cable (HAW-1) extending from Point Arena, California, to Oahu, Hawaii. The electric field measurements are coherent to some degree wi th magnetic field measurements from Honolulu Observatory at periods of 0.1 to 45 days. This coherence is enhanced at long periods over that observed with point electric field sensors due to horizontal averaging of the motional electric fields of spatial scale smaller than the cab le length, significantly diminishing their effect. Robust, controlled leverage MT response estimates and their jacknife confidence limits ar e computed for the HAW-1 to Honolulu data. An equivalent scalar MT res ponse obtained from Honolulu magnetic variations data is used to corre ct the HAW-1 MT response for static shift and to extend the MT respons e estimate to periods of 100 day's. The composite response function sa tisfies necessary and sufficient conditions for consistency with a one -dimensional conductivity structure and is most sensitive to structure between 150 and 1000 km. Inversion of the MT response reveals a condu ctive zone (0.05-0.1 S/m) between 150 and 400 km depth and a positive gradient below 500 km; these observations are consistent with previous MT studies in the North Pacific. This upper mantle conductivity is to o high to be explained by solid-state conduction in dry olivine using reasonable mantle geotherms. Calculations based on measurements of hyd rogen solubility and diffusivity in olivine indicate that Hf dissolved in olivine, possibly combined with a lattice preferred orientation co nsistent with measured seismic anisotropy, provide sufficient conducti vity enhancement to explain the inversion results. The high conductivi ty may also be explained by the presence of gravitationally stable par tial melt. Comparison of the HAW-1 results with long-period MT studies conducted on land reveals differences in upper mantle conductivity be tween different tectonic regimes. In particular, the upper mantle bene ath the Pacific Ocean is considerably more conductive than that beneat h the Canadian shield and similar in conductivity to that beneath the Basin and Range.