GP33A-3702

Preliminary Results of a Magnetotelluric Survey in the Center of Hawaii Island

UNIVERSITY of HAWAI ' I ·

~-a: Kawa'-. tDabMe of HILO Barry Lienert1, Erin Wallin 2 and Donald Thomas 1,2

, ffenJhRSi&s & PI'anetn[nll

~ HIGP/SOEST, Univ. Hawaii at Manoa, 2525 Correa Rd, Honolulu HI 96822, lienert@soest.hawaii.edu CSAV, Univ. Hawaii at Hilo, 200 W Kawili St, Hilo HI 96720

Abstract From 2013 up to the present we have been recording magnetotelluric (MT) data at 25 sites in a 35x25 km region (elev. 1943 m) on the saddle between the active volcano of Mauna Loa (4169 m) and the dormant volcano of Mauna Kea (4205 m) on HawaiOi Island. The MT data, particularly the electric fields, are frequently contaminated by spurious components that are not due to the plane-wave magnetic Signals required for derivation of the M T impedance tensor. We therefore developed interactive graphical software (MTPlot) to plot and analyze the MT signals in the field. MTPlot allows us to quickly ex­amine records in both the time and frequency domain to in order to judge their quality. It also transforms the data into estimates of apparent resistivity and their error in the frequency range 0.001-500 Hz. This has proved very useful for selecting suitable records for subsequent analysis. We then use multi-taper remote reference processing to obtain our final apparent resistivity estimates and their errors. We present prelimi­nary results of one and two dimensional modeling of these estimates to obtain the three-dimensional distribution of subsurface resistivities down to depths of 5 km. The results are compared to temperatures and properties of cores obtained when we drilled a research hole to a depth of 1760 m in this same region. We shall discuss how our re­sults relate to the extent of the fresh-water and geothermal energy reservoirs that we discovered during drilling.

Results So far, data for constructing models have been processed at 9 sites, 8 of which appear in Fig. 1. At some of the sites, data were biased by both cultural noise and perturba­tions that may be self-potential effects related to volcanic processes. After examining layered model inversions for both XY (red symbols) and YX (blue symbols) components of apparent resistivity derived from the MT impedance tensor at each of the 9 sites, the 2D model in Fig. 2 was obtained using a starting model consisting of a 100 ohm-m half space. Inversions were run with the strike of the two-dimensionality (along which the resistivities vary) in the NS, EW, NE-SW and NW-SE directions. The EW model was selected as it had a significantly lower misfit. However, the relatively small differ­ences between the misfits implied that 3D modeling will be required to establish the N-S limits of the resistivity low closest to the drill hole.

Discussion

'''I: ",

~I~~

'o'l~~;~"'"' ~ ::':':~ IpTA" 3 _"" " I!IlilllcDI 'o'l~. ;~""' ~ :~"" "'1 ""''''';;---3~· '''" 1!lI000IDI

102 • • •• • 102 .. ...

",

~I::;~ .N~ -N~. ~EI::"-~-:~~'~: :::o::ITrTTT-:-

'l~~:H: '~~'~ ' :j :,-,-*-::---,-~ , ':"--,-,, :.~-.-.-. .' ... ' +-'-" .'--'--.L •• -'-1-••• ~.

Background

Dry basalts typically have resistivities of >1000 ohm-m, so the low resistivities in Fig. 3 imply a combination of high temperatures, water- saturation and/or mineralization. Cores from the HGRP drill hole indicated that at depths of < 550 m, there were several thin perched acquifrers. At depths greater than 1000 m the cores consisted of basaltic lava that become successively compacted and mineralized with increasing depth. A previous audiomagnetotelluric (AMT) study 3 over the Puhimau thermal area south of Kilauea summit was modeled with a 5 ohm-m conductor at a depth of 200 m. These resistivities were proposed as being due to either partially molten magma or hot solidified magma containing water. The shallowest low resistivities in our 2D model to the east of the drill hole could be explained by mineralized and water saturated basalts at temperatures >300 C. At greater depths, where porosity is much lower, the low resistivities suggest the presence of magma or partially molten material. Data from more stations will be required to more accurately establish the extents of the two low-resistivity regions in Fig. 3, but our results suggest that their extents are> 5 km.

Remote Reference M T Site

The Humu'ula Groundwater Research Project (HGRP) drilled their first continuously­cored hole in the saddle region of the big island of Hawaii in March of 2013 achieving a depth of 1760 m (Fig. 1). Temperatures at the bottom of the hole were unexpectedly high and reached over 100 C. Since this suggested the presence of a potential geothermal resource, a magnetotelluric (MT) survey was initiated to establish the extent of this resource.

Methods Five-component (three magnetic, two electric) MT data were collected at the sites shown in Fig. 1 using five commercial MT systems (KMS Technologies) employing coils and Ag/ AgCI electrodes. Data were sampled for 12 hours at 1 ksamples/sec and 4 days at 40 samples/sec using the systems 24-bit data loggers.

We developed our own interactive Windows software package, MTPlot (Fig. 2) to initially examine the data and perform the following tasks: 1. Editing of header information, concatenation, filtering and sub-sampling. 2. Detrending (Fig. 2), windowing and Fourier transformation. 3. Spectral cross correlation and calculation of apparent resistivities and phases and their errors

Robust Remote Reference estimates of the MT impedance tensor were then determined using published Fortran routines1.2 . The resulting impedances were then inverted for one and two-dimensional resistivity models using WinGLink, a commercial software package.

'a. -5 1 -10

-15

'a. i

-5

7.5

'a. 5.0

~ 2.5 0

~2 .5

References 1 Egbert, G.D., and J.R. Booker, Multivariate analysis of geomagnetiC array data: 1. The Response Space, J. Geophys. Res., 94, BIO, pp 14,227-14,247, 1989 2 Egbert, G.D., Robust multiple-station magnetotelluric data processing, Geophys J. Int., 130, 475-496,1997 3 L. C. Bartel and R. D. Jacobson, Results of a control/ed-source audiofrequency magnetotelluric survey at the Puhimau thermal area, Kilauea Volcano, Hawaii, Geophysics, 52,665-677,1987

Acknowledgements Funding for this work was provided by Army Garrison Hawaii, and Pacific Command through the Cooperative Ecosystems Study Unit Grant W9126G-1l-2-0056, and by the Office of Naval Research through the APRISES (ASia Pacific Research Initiative for Sustainable Energy Systems) Program.

sms. I5osIidEsIirrloo~

r~ r zu r;:AmIlI~ r' Z'I"

ImI>edan«:EsIirrloo~-­

acoeffs!Uf)llled ~

Ov~, % ro--LNS!~esEstimates I PdynomiaiRttin;!~

r.;r ~ (; Awar~tR~mo.. ' ~....oes ~ f"3 .3 Ir __ r,- ",~,,·~ I "'~ r"".~

jj bi.,,!, ~t'!·'~ '~ ~'~~~I.'~-1.~~*~"I'~'~'" :~~ ~'4

.... d'

: ~~ :" ... ",;., ... "J., ", ;,,,,,,,",~l.,j~ .. ,, ~, ""'" ~ ... ,,~ •• " •• 'r :'t "'" ; .. , •• ~ 10/30/1316:01 :15 16:01 :30 16:01 :45 16:02:00

Figure 2. Main window of the MTPlot application showing one record (2560 points) of 5-component MT data and fitted 3rd order polynomials (red curves) used for detrending.

UTe

, .~.:.~-=-Zi~

, ,

W I ~ ~ •••••• ~

~ ~~ P Hz ~ ~'::;mg IPTAOl 3 P Ell ~

:I~ r~~~ .N~N~

':r~~--~~ 10'1 •••• ,;p

];1E-10- T:! 0(J312 sec:

Figure 1. Regional (left) and station map (center) showing location of MT stations on Hawaii Island and the location of the 2013 m HGRP drill -hole. The location of the Remote reference site is shown on the regional map. Apparent resisivity, phase (TE & TM modes) and tipper data dram selected sites are plotted around the central map as well as the curves predicted by the 2D resistivity model in Fig. 3.

~ 1.(1656 rs.rM-to [S olE- 10- T

OJ > ~ 0 ., '" ;t 0

OJ -'l L .p-o. .,

CJ

5

10

15

--

I

Figure 3. EW cross-section of the 2D resistivity model obtained by inverting the MT data in Fig. L The verti­cal and horizontal scoles are equal and identical to the horizontal scale in the station mop in Fig. L The resistivity color codes are shown on the right. 1D layered model TE mode inversions are also shown beneath each site.