 |
|
|
|
| Matthews,
M. D., 1996, Importance of sampling design and density in target recognition,
in
D. Schumacher and M. A. Abrams, eds., Hydrocarbon migration and its near-surface
expression: AAPG Memoir 66, p. 243-253.
|
|
|---|
|
|
Abstract
The design and
density of surface  geochemical sampling programs can significantly influence
the interpretability of the survey. Testing hypotheses by purposeful sampling
is the most straightforward application of surface geochemistry and the
easiest to interpret but requires
a priori geologic knowledge and
provides limited information. Designing a spatial sampling program to produce
a map is a much more difficult problem. Two common techniques are line
profiles and areal surveys. The interpretation of these techniques, as
a function of sampling density, is simulated by examining the relationships
of artificial surveys to a known subsurface target. A deterministic model
of hydrocarbon migration is used to constrain the randomized mixing of
two nonoverlapping populations (anomalous and background). A high-resolution
grid of simulated surface measurements is decimated to create lower resolution
grids and line profiles. For regional high-grading, a grid with a minimum
of two samples across the narrowest expected surface signal (minimum subsurface
target width plus a dispersion zone) appears to be adequate. A higher density
of four samples is suggested for prospect high-grading. In addition, the
sampled area must include sufficient "background" measurements to recognize
the existence of an anomaly. About 80% of the samples should be obtained
outside the expected area of interest. The most cost-effective sampling
program consist of two stages: a low-density regional survey to high-grade
the area, followed by a higher density survey within the high-graded area.
Undersampling is probably the major cause of ambiguity and interpretation
failures involving surface geochemical studies. |
|---|
