Root establishment of perennial ryegrass (L. perenne) in diesel contaminated subsurface soil layers.

The efficiency of rhizosphere biodegradation of petroleum hydrocarbons heterogeneously distributed in soils is dependent on the ability of plant roots to prospect into contaminated zones. Rhizobox experiments were conducted to study the influence of diesel contaminated layers on the spatial distribution and the development of the roots of perennial ryegrass. Root distribution and root and shoot development were monitored over time. The final root and above ground biomass and the final TPH concentration were determined. The spatial distribution of the contaminant as well as the irrigation method used affected root distribution, plant development and TPH degradation and therefore ryegrass remediation potential. The results show that roots colonise fully uncontaminated soil and grow preferentially between zones of contamination. Conversely, when no immediate uncontaminated soil is available, roots grow through contaminated zones in order to prospect for uncontaminated soil.

Determination of time-dependent partition coefficients for several pesticides using diffusion theory.

Diffusion-retarded partitioning of pesticides with aggregated soils results in a time-dependent partition coefficient (Kd') which is different at equilibrium from the partition coefficient derived from conventional 24-h batch studies (Kd) measured on dispersed soil. An experiment was undertaken to determine the importance of Kd' for the prediction of pesticide concentrations in solutions bathing artificial soil aggregates and to determine whether diffusion theory could accurately predict the concentrations. Two clay soils were mixed with polyacrylamide to create artificial aggregates of 0.8, 1.4 and 1.7 cm diameter when dry. After saturation, the aggregates were immersed in solutions containing isoproturon or a mixture of isoproturon, chlorotoluron and triasulfuron. The decline with time of the pesticide concentrations in the bathing solution was monitored and the results were compared with predictions from a diffusion-based model. The effective diffusion coefficients of the compounds were obtained by either fitting the non-linear diffusion model to the data (D(ef)) or by independent calculations based on the properties of the compounds and of the aggregates (D(ec)). The diffusion model was able to predict the temporal variation in pesticide concentrations in the bathing solution reasonably well whether D(ef) or D(ec) values were used. However, equilibrium concentrations in solution were sometimes overestimated due to increased sorption with time at the particle scale. Overall, the ratio between D(ef) and D(ec) ranged from 0.23 to 0.95 which was a reasonable variation when compared to the range of aggregate sizes used in the experiments and of the Kd values of the compounds.