Sampling can aim at documenting contaminant concentrations and fluxes during and after storm events (or other rain or snow melt), at mirroring the load and flux of contaminants over an extended period of time, or at characterizing an accidental discharge.
Road runoff presents specific characteristics that change significantly from place to place, depending on site characteristics and many other factors (Barbosa & Hvitved-Jacobsen, 2001). Therefore, the monitoring programme should also include the characterization of the most important factors such as: traffic characteristics (volume, speed, type of vehicles, fuel types, etc.), geographic location, climate, topography, drainage area and road design, pavement characteristics, right-of-way characteristics and adjacent land use. Although the quality of water is being influenced by several external characteristics, it is the duration and intensity of the pavement washing that determines the degree of dilution and transport of pollutants. The site selected for runoff sampling should be a drainage area (with significant and representative size) where the runoff from the paved and unpaved areas can be isolated from other sources within the selected highway system.
Runoff sampling may be used to characterize the influence of the mean daily traffic of a road section. For that purpose samples must be collected in a discrete way, in a surface drainage pipe (e. g. Fig. 7.1), during rain or snow thawing events, and be associated with a specific road drainage area. The runoff flow measurement devices are usually installed transversally in the outflow pipe from the system draining the area of pavement that is of interest and must include a calibrated component that links the observed water level to the corresponding runoff volume by knowing the shape and dimension of the pipe. Runoff sampling depends on the occurrence of precipitation, and therefore, sampling is facilitated by the use of sampling equipment that is activated by a sensing device which automatically starts the sampling whenever the flow rate exceeds a previously defined value. Such sampling equipment also has the benefit of allowing the sampling of the first flush of a storm event.
This type of monitoring gives discrete information about the changes in runoff pollution during a certain period and allows the determination of the Event Mean Concentration (EMC) and the Site Mean Concentration (SMC). An EMC is
calculated for an individual storm event as the total mass load of a pollutant parameter (CV) divided by the total runoff water volume (V) discharged during the storm (Hvitved-Jacobsen & Vollertsen, 2003):
EMC = ZCV/ZV (7.1)
A SMC is a characteristic runoff annual pollution loads for a specific site, typically defined by the arithmetic mean value of the EMC’s measured at one site (Hvitved — Jacobsen & Vollertsen, 2003).
To evaluate the EMC it is advisable to sample the whole event in a way that is time constant and volume proportional. In this method, discrete samples are collected at equal time increments and composed proportional to the varying flow rate during the sampling period (FHWA, 1987, 1988). In order to mirror seasonal variations, a monitoring programme should, ideally, include several discrete storm events for a one-year period. Authors differ in their opinions about the minimum number of storm events needed. Some authors consider 10, others 6 storm event episodes, each one characterized by a minimum of 5 samples (Barbosa, 1999, Burton & Pitt, 2002). Probably, the choice is influenced by storm severity and other climatological factors that may vary widely between countries and regions.
To allow the characterization of road runoff throughout a specific period, runoff monitoring devices should be:
• automated;
• include a rainfall device;
• include a runoff flow measurement device; and
• include sampling and recording equipment.
All these devices should operate in phase with each other in order to allow the determination of lag time from rainfall to runoff and the determination of pollutant load in each sample.
Figure 7.2 illustrates an alternative approach, common for measuring lower flow volumes than observed in runoff collection pipes — e. g. rainfall and seepage water. Water is fed into a bucket (Fig. 7.2a) that tips alternately to left and right when it fills. The number of tips may be counted electronically. Figure 7.2b shows a typical installation, in this case with the possibility of some of the water being sampled for later analysis. By integrating the output of the tipping buckets with the automatic water sampling equipment (as shown in Fig. 7.3) it is possible to sample water on a volumetric basis rather than a time basis.
Where discrete sampling is not feasible, runoff sampling can be accumulative. For that it is possible to arrange special sampling devices, in the form of a gutter installed along the sealed pavement’s edge to capture runoff water and then to conduct it to a type of tank, such as a rainwater settling tank or a retention tank installed close to or away from the road. Afterwards the tank can be sampled by abstracting a proportion into a vessel installed in a chamber dug into the roadside soil. These devices can give a broad idea about the road pollution but are not appropriated for load calculations if their draining section is not easily delineable. These sampling
methods are less informative since they give information about the road pollution integrated over a period of time. Also, other sources than the road and its traffic might contribute to the contaminants collected in the tanks. Furthermore, change in water condition can take place between the time that water arrives at the tank and the time at which a specimen is collected. Sampling from tanks, therefore, only gives a qualitative overview of road pollution. In the case of an overall sample taken in
Fig. 7.3 Automatic water sampler with refrigerated cabinet (Courtesy of Hach. ©Hach Company, 2007)
a settling tank or retention tank, water collection must be made at several locations and depths of the tank so as to provide a composite sample representative of the water in the entire tank.
