The effects of soil environment on the ability of surface geochemical surveys to detect underlying hydrocarbon traps

Abstract

Soil surveys represent an inexpensive potential technique for preliminary hydrocarbon exploration. However, large scatter and lack of repeatability have limited their use. To determine whether a signal from a reservoir is actually detectable, a series of closely spaced soil samples was collected along a traverse line across a gas-storage reservoir. Gases were desorbed from the soil by heating and were analyzed for their content of light hydrocarbons using gas chromatography. Additional information concerning the soil and its environment was also collected at each sample site to investigate the source of scatter in the data. Analysis of covariance (ANCOVA) was used to determine whether differences in ethane concentration from one site to another were significant and whether they could be correlated with the reservoir boundaries and/or other variables. The results do show systematic variations in ethane concentration and a strong correlation between ethane concentrations and the known reservoir boundaries. However, strong correlations were also found with some of the soil environment variables, specifically the moisture content and the land use (presumably related to the amount of disturbance). Unless these and other soil environment factors are taken into account, variability in gas content caused by them can mask any relationship with a hydrocarbon reservoir. Two other significant observations emerge from the study. First, sample spacing needs to be closer than that used in conventional surveys in order to detect boundaries and to separate the reservoir signal from surface noise. Second, the individual light hydrocarbons appear to behave differently in the soil. Methane is present in low concentrations despite its high concentration in the reservoir gas. Ethane is more abundant and was chosen for use in the analysis. Propane and butane have very low concentrations, reflecting their low concentrations in the natural gas in the reservoir. However, pentane, though dilute in natural gas, is very abundant in the soil. We believe that pentane is retained because soil temperatures are near its liquidus, particularly during cold weather. This makes pentane potentially unreliable as an indicator of underlying gas flux.