REDUCING SEDIMENT RUNOFF FROM UNSEALED ROADS

Road & Transport Research, Sep 2004 by McRobert, Jencie, Kemp, Alison

INTRODUCTION

The Unsealed Roads Stormwater Project has been investigating practical ways to control stormwater pollutants from unsealed roads. With funding by EPA Victoria as part of the Victorian Stormwater Action Program, the municipalities of Cardinia, casey and Mornington Peninsula initiated the project in an effort to protect stormwater quality in the Westernport catchment.

The research project aimed to compare the effectiveness of various road maintenance techniques in minimising sediment runoff from an unsealed pavement, particularly the interaction between the road profile, gravel quality and quantity, and surface treatment and their affects on sediment production and delivery rates.

Research input to the project was provided by the three councils, ARRB Transport Research and University of Western Sydney. The Cooperative Research Centre for Catchment Hydrology at Monash University also had input to the experimental design and results analysis for the project.

EXPERIMENTAL DESIGN

The experiments were conducted at Boes Rd located on the outskirts of the Hastings township in Mornington Peninsula Shire, Victoria. It has a moderate vehicle usage (approximately 350 vehicles per day) as a link between the nearby urban centres of Bittern, Hastings, Tyabb and Somerville.

Using rainfall simulation, runoff was sampled from three gravel surface treatments designed to observe the affect of the following variables on sediment generation: compaction, road surface materials and crossfall. The treatments were approximately 50m in length and a buffer of at least 20 metres separated each section. The treatments were as follows:

* Control - 100mm local gravel base course, dry compaction, 3% crossfall

* Treatment A - 100mm local gravel base course, wet compaction, 3% crossfall

* Treatment C - 100mm local gravel base course, wet compaction, additional 70mm wearing course, 6% crossfall

It should be noted that the wearing course material used in Treatment C was later tested and found to be not within specification due to an inadequate percentage of fines and limited plasticity.

Two experimental plots were constructed in each of the treatment sections. A drain constructed from sheet metal guttering was installed at the foot of each plot to intercept the runoff. The guttering was held in place by concrete and sealed to the road pavement using mortar, as shown in Figure 1.At the base of the plot, a sheet metal flume was used to measure the volume of water discharged from the plot during each experiment.

On each plot, three different sized rainfall events were simulated. Table 1 details the intended intensity and Average Recurrence Interval (ARI) for each rainfall event. A further 1 -in-50 year storm was conducted on one plot from Treatment A to examine the effects of an extreme storm on sediment runoff. Rainfall intensity data was sourced from an Intensity-Frequency-Duration (IFD) chart supplied by the Bureau of Meteorology. Each rainfall event ran for 15 minutes.

The storm events were run in order of lowest to highest intensity to allow for depletion effects. This is based on the assumption that sediment generated during the smaller storms would have been generated in the larger storms.

Samples were taken every 30 seconds, as shown in Figure 2. A selection of these were analysed for turbidity, pH, total suspended solids, and particle sizing.

RESULTS

Rainfall and sediment discharge

The rainfall intensity attained during some of the simulation runs did not meet the designed intensity. In each case the 1 year ARI was usually closely met (ranging between 0.8-1.8 year ARI) while the 5 year ARI was met within 3-5 years in most cases. All of the simulations targeting the 10 year ARI rainfall intensity were well in excess of this storm size ranging between 17 and 90 years. These differences in actual versus designed storms was likely to be due to variations in pressure, winds, nozzles and insufficient time spent on calibration prior to conducting the experiments.

Sediment concentration was measured as Total Suspended Solids (TSS). TSS is the amount of material suspended in the water column and is expressed in mass/volume (g/L). Figure 3 shows the mean TSS concentration for each rainfall event. The graph indicates that TSS increased with rainfall intensity; Run 3 (1 in 10 yr storm) had a peak concentration of 1.76g / L of sediment, which was two and a half times the peak concentration of Run 1 (1 in 1 yr storm). Sediment concentration tended to be highest after the first few minutes of rainfall, and then tapered off toward a steady state. This 'first flush' pattern is typical of the behaviour of stormwater pollutant runoff, and is likely to have been caused by the gradual depletion of material on the plots during each successive storm event.

The sediment discharge rate from the experimental plots, expressed in mass per second, is shown in Figure 4. The graph shows that with increasing rainfall intensity, the rate of sediment discharge increases. Sediment discharge was also initiated earlier during Run 3, commencing after only 1.5 minutes compared to 4 or 5 minutes during Run 1. This suggests that the road surf acebecame saturated more rapidly following each successive rainfall event, thus resulting in increased runoff.