No abstract

No abstract

This study addresses soil contamination in vegetated road shoulders with diffuse inflow of runoff. It aims (i) to characterize the spatial distribution of three metals (copper, lead, zinc) and PAHs, and (ii) to identify influencing factors for the inter-site differences. An extensive sampling campaign was carried out on forty road segments in the Paris region, targeting various distances and soil depths. Copper and zinc contamination was found to be mainly restricted to the first 30 cm from the road, where their concentrations ranged respectively between 40-270 mg.kg-1 and 150-950 mg.kg-1 (with a few additional extreme values related to the presence of a galvanized steel guardrail). Lead contamination was moderate (< 100 mg.kg-1 in more than 75% of the sites) and relatively uniform across the width of the shoulders. Conversely, highest PAH concentrations were found further from the road. These differences between contaminants were likely due to the varying importance of wet weather-related processes and atmospheric transport in their dispersion mechanisms. Copper and zinc concentrations correlated well with the traffic density, which alone explained ≥ 69% of the inter-site variability, whereas lead and PAHs did not exhibit such dependence. Soil organic matter was found to control the "ceiling" concentration of metals, thus limiting the maximum amounts that can be intercepted from road runoff. These results illustrate the feasibility of estimating contamination levels on road shoulders, and thus offer interesting perspectives for better consideration of the infrastructures' maintenance needs and improved routine operations.

Two methods for sampling aggregates in the soil surface under simulated rain were compared using two soil types. Results showed that aggregate size distributions obtained by spatula sampling were not significantly different from those obtained using rings buried in the soil surface, provided both were sampled to the same depth. The effect of transporting samples over a distance of 60 km was non-significant when samples were placed in bottles half-filled with rainwater and transported in an upright position. The per cent aggregates > 0125 mm was found to be the most suitable index of aggregate stability for both soils.

Two rainfall simulators of different plot sizes were used to test whether sediment in runoff could be used to measure aggregate breakdown in the surface of a cracking clay soil under rain. Plots were prepared with either levelled or furrowed surfaces. Samples of the soil surface under rain were taken from furrow ridges or levelled surfaces, and from areas of deposited sediment. These were compared with samples of sediment in runoff taken at the same times. On both furrowed and levelled plot surfaces and for both simulators, aggregate sizes were significantly finer in sediment in runoff than in samples of the soil surface taken with a spatula. No significant differences in surface aggregate size distributions were found between rainfall simulators, or between furrowed and levelled plot surfaces. Regression lines fitted to the data on size distributions of sediment or of aggregates in the soil surface showed no significant changes through time. The fitted lines showed sediment in runoff to be still significantly finer than aggregates in the soil surface after 50 min rain at 95 mm h ~ ', except for levelled plots under the rotating disc rainfall simulator, where extreme variability of data meant that even relatively large differences were not statistically significant. Size distributions of deposited sediment were similar to those of the surface of adjacent furrow ridges exposed to raindrop impact. This provides evidence that sampling the soil surface with a spatula gives a representative sample of the material available for rain-flow transport.