Natural leakage pathways through smectite clay: a hydrological synthesis of data from the Hudson Agricultural Trial Site on the Liverpool Plains. August 2002.

Abstract

This report provides a synthesis of available hydrological data for a site at Hudson on the southern footslopes of the Liverpool Plains. Agronomic processes at the site have been the subject of detailed study since it was suspected that significant groundwater recharge in the area was caused by deep drainage beneath short rooted crops. The data strongly indicate that soils cultivated on the smectite dominated clay and silt soils do not wet up uniformly. The presence of air filled fractures at depth is deduced from the barometric efficiency in water in piezometers installed at 15 m depth beneath the trial site. Investigation using the electrical image method indicated that changes in bulk resistivity occur to a depth up to at least 8 m. and it is deduced that this resistivity change must be caused by changes in soil moisture. If this is correct then lucerne seems to have changed the soil moisture status to as deep as 8m. The reduction in soil moisture is believed to have promoted deep cracking in soil column in depth greater than those normally associated with agriculture. There are strong indications that cracking associetd with growing lucerne add to bypass flow. At the Hudson site, deep drainage via fractures contributes to perched episodically saturated zones and is intercepted by caliche (Ca and Mg carbonates) horizons.Preferential flow appears to occur above at least two caliche horizons between 8 and 15 m in depth. A small proportion of deep drainage seeps through clay to the basalt aquifer. A more efficient local recharge source appears to be leakage from the base of an erosion gully crossing the trial site.Non-fractured and saturated Ca-smectite clay was characterized by a low hydraulic conductivity, but was not impermeable. Leakage was transmitted to the deeper aquifer with a delay of several hundred days following recharge events. The low specific storage means that relatively small quantities of recharge have led to increase in water level of several meters beneath the trial site. Hydro geochemical changes and incongruent isotope results indicated that water at 15 m depth was dominated by old matrix water contaminated by only a small proportion of recent recharge. The efficiency of recharge will be much greater in areas where weathered basalt is less than 10 m depth.

Conceptual models of a dual-porosity and compressible weathered basalt medium were best able to explain the field observations. The present work has been able to confirm that a small quantity of deep drainage to the saturated aquifer has occurred but is not able to quantify this. More accurate quantification will require extensive monitoring and the development of appropriate models of water movement through reformable smectite dominated silty clays.

The initial suspicion that deep drainage occurs beneath short rooted crops, based on a simple water balance, appears to have been unwarranted. However, the practice of trying to establish a large soil moisture deficit by growing a deep rooted crop like lucerne appears to cause too much drying with the result that fractures are established. At the Hudson Site, the water level in piezometers beneath the trial site was below 10\,m. A small quantity of recharge appears to have moved through the unsaturated profile to this depth. The quantity of recharge would be much greater over a shallower clay profile