ANALYSIS OF STEADY-STATE CHLORIDE TRANSPORT THROUGH 2 HETEROGENEOUS FIELD SOILS

Chloride transport was investigated in a loamy soil and a silty-loam s oil at the field scale under steady state flow conditions using a wate r flux of 2.84 cm d(-1) for the loamy and 1.5 cm d(-1) for the silty-l oam soil. The solute plume movement was recorded by means of horizonta lly installed time domain reflectrometry (TDR) probes at 5 depths up t o 90 cm below the soil surface and 24 locations along a transect of 8 m. The measurements consisted of solute resident concentrations every 2 hours in the loamy soil for a total period of 42 days and every 4 ho urs for the silty-loam soil for a period of 65 days. Parameters of the convective-dispersive equation (CDE) and the lognormal stochastic-con vective transport model (CLT) were determined using time-normalized re sident concentration breakthrough curves C-rt(z, t). In addition, tem poral moments of C-rt(z, t) were related to travel time moments and t ransport parameters for the two transport processes. At both sites the breakthrough curves at different depths were better described by the CLT than by the CDE. However, early solute breakthrough was underestim ated at most depths. Mean travel time and dispersivity were estimated using the temporal moments of C-rt(z, t) with the assumption of a sto chastic-convective transport process. In the loamy soil, solute was tr aveling from a heterogeneous, macroporous top soil toward a subsoil co ntaining significantly fewer macropores. The flow of solutes through t he macropores is not detected by the TDR probes, resulting in a larger observed mean travel time compared with the expected mean travel time based on the piston flow model and no increase in dispersivity. In co ntrast, in the subsoil the observed and expected mean travel times wer e in good agreement, and dispersivity increased with depth. In the sil ty-loam soil, mean travel times derived from concentration measurement s were larger than the expected mean travel times based on the piston flow model, implying temporal storage of solutes in stagnant water zon es. Dispersivity also showed deviations from the expected linear incre ase with depth, probably because of changing soil properties with dept h.