Title: Multi-dimensional interpretation of electromagnetic data from Silali geothermal field in Kenya: Comparisons between 1-D, 2-D and 3-D MT inversions

Abstract

With development of MT interpretation codes and advancement in computer hardware, 3-D
electromagnetic data interpretation has become attainable. This study therefore seeks to compare
results of Electromagnetic data using different interpretational techniques in order to provide
reliable information about the presence, location, and size of geothermal systems in Silali field.
Resistivity study of the Silali area in Kenya was carried out by the combined use of TEM and MT
soundings. Joint inversion of the EM data was used to correct for static shifts in the MT data,
which can be severe due to large near-surface resistivity contrasts. Joint 1-D inversion of 102
TEM/MT sounding pairs and a 3-D inversion of a 97 sounding subset of the MT data were performed.
Additionally 2-D inversion of the same data set was done and results are compared with those of 1-D
and 3-D inversion models. The robustness of the final 3-D inversion models was tested by using
three different initial models, which gave similar results with RMS of between 1.5 and 1.7 for all
three models. Similarly 2-D inversion modelling was done using two different inversion codes,
REBOCC and WinGlink and their results compare fairly well.
The resistivity models resulting from the inversion were elevation corrected and smoothed and are
presented as planar maps and cross sections. The inverted model of electrical resistivity reveals
the presence of highly resistive near surface layer, identified as unaltered formations, which
covers a low resistivity cap corresponding to the smectite-zeolite zone. Beneath this cap a more
resistive zone is identified as the epidote-chlorite zone (the resistive core) and interpreted as
the host of geothermal reservoir. Further at depth of about 6 km an electrically conductive feature
has been imaged, and has been tentatively interpreted as a heat source for geothermal system in
this field.
The aim of modelling EM data using all the three interpretational techniques is to compare