Deconvolution
and Mapping of Soil Gas Anomalies in Surface Prospecting: A New Approach Based
on Bayesian Geostatistics
Goncalves, Felix T. T.1,
Fernando H. Pulgati1, Ricardo P. Bedregal1, Flavio L.
Fernandes1, Jason T. G. Carneiro1 (1) PGT-Petroleum Geoscience
Technology Ltd,
Geochemical prospecting methods use
surface or near-surface occurrences of hydrocarbons (micro seepage) provide
direct evidence of the existence of an active petroleum system, helping in the
identification of prospective areas and in the assessment and ranking of
exploration leads and prospects. During the last decades, a significant
advances in sampling and analytical techniques have allowed the detection of
minute traces of hydrocarbons in a more accurate and effective way. Conversely,
interpretative methods have been mostly limited to straightforward statistical
approaches that define anomalously high values relative to a background usually
by (a) arbitrarily assigning a background threshold, such as the mean or one
standard deviation from the mean, (b) separating groups or populations of
concentrations based on the plot of all the data as a histogram; or (c)
plotting the data in a cumulative frequency diagram to delineate a break in the
slope of the curve. This study presents a new integrated geostatistical
approach that allows an enhanced definition of surface hydrocarbon anomalies in
several areas from Subandean basins and Brazilian rifts. Analysis of variance
was used to identify and filter the influence of near-surface environmental
factors, such as the moisture content, grain size and land use, which can mask
any relationship with subsurface hydrocarbon accumulations. Bayesian modeling
with Monte Carlo-Markov Chain (MC-MC) simulation was applied to delineate areas
of anomalous concentrations of thermogenic hydrocarbon gases by calculating the
probability that gas concentration on each site is higher than the value
expected by the stochastic process, recognizing spatial clusters of anomalous
values, and assessing the risk of biogenic contamination based on molecular
parameters.
AAPG Search and Discover Article #90063©2007 AAPG Annual Convention, Long Beach, California