IMPROVED SOLUTE-SAMPLING PROTOCOL IN A SANDY VADOSE ZONE USING GROUND-PENETRATING RADAR

Soil solution samples taken from a vadose zone by current monitoring techniques are randomly obtained because it is assumed that the representativeness of the samples is dependent on the total number of samples instead of the location of the samples. In the presence of preferential flow (i.e., "macropore", "fingering", or "funnel" flow), however, the locations as well as the number of samples and not just the number of samples determine the representativeness. The purpose of this study was to: (i) demonstrate that current sampling protocols could underestimate solute breakthrough in a soil with preferential flow; and (ii) show that ground-penetrating radar (GPR) has great potential to nondestructively map the location of interbedded soil layers with textural discontinuities so that suction-cup samplers could be correctly installed along the preferential paths triggered by funnel flow. A 24 by 24 m plot at the Hancock Research Station, Wisconsin, was scanned with GPR in 1989. Based on the GPR images, four ceramic suction-cup samplers were installed at a depth of 1.85 m in locations with layering structures and abrupt textural discontinuities. Four other ceramic suction cup samplers were installed randomly at a depth of 1.5 m. The plot received a uniform pulse of Br solution. Breakthrough curves collected by samplers installed with the aid of GPR have higher concentrations than those collected by randomly installed samplers. This result indicates that random solute sampling in a sandy vadose zone exhibiting preferential flow could underestimate solute movement.