Sampling and Variation: Application to Surface Geochemical Surveying

Joel L. Gevirtz, John B. Curtis, Tony W. Gorody, Stephen W. Brown

The objective of geochemical soil surveys is to gather surface and near-surfaced data that, when combined with conventional information, may provide information concerning type and distribution of hydrocarbons in the subsurface.

A well-planned sample arrangement is necessary if the resulting information is to be confidently applied throughout the sampled area. No amount of later statistical analysis or other sophisticated mathematical manipulation can overcome inadequate or improper sampling. Better estimates of probabilities of hydrocarbon potential and a more reliable depiction of probable regions for hydrocarbon reservoirs within a basin result when sampling design is coordinated with project goals.

In contrast to subsurface exploration, surface methods allow for a totally programmed sample design. Grids and various random arrangements are sample designs that estimate subsurface processes in three dimensions. Non-grid linear sample patterns, including selected locations along roads, fence lines, or seismic survey lines, estimate variance along these lines and, therefore, are limited to estimating subsurface processes in two dimensions.

A continuing tendency exists to plan surveys without a clear concept of purpose. Line transects are commonly performed when grids are required, or widely spaced grids are used where tightly spaced grids are more appropriate. Several optimal sampling designs are demonstrated to cover various terrestrial sampling situations.

Accurate interpretation of survey data requires accurate estimation of the relative contributions from each of three sources of variation: the surface environment, operational error, and the subsurface. Data can be related to the geologic framework, as determined by conventional subsurface methods, only within the limitations established by the magnitude of unwanted variation. Several pilot study designs that are useful for determining the relative magnitude of unwanted variation resulting from both operational error and from surface environmental effects are suggested, and should be planned as part of any large-scale program.

AAPG Search and Discovery Article #91030©1988 AAPG Annual Convention, Houston, Texas, 20-23 March 1988.