Turbidity dynamics during spring storm events in an urban headwater river system: the Upper Tame, West Midlands, UK.
Turbidity is an important water quality variable, through its relation to light suppression, BOD impact, sediment-associated contaminant transport, and suspended sediment effects on organisms and habitats. Yet few published field investigations of wet-weather turbidity dynamics, through several individual and sequenced rainstorms in extremely urbanised headwater basins, have emerged. This paper aims to address this gap through a turbidity analysis of multiple storm events in spring 2001 in an urban headwater basin (57 km2) of the River Tame, central England, the most urbanised basin for its size in the UK ( approximately 42%). Data were collected at 15-min frequency at automated monitoring stations for rainfall, streamflow and six water quality variables (turbidity, EC, temperature, DO, pH, ammonia). Disturbance experiments also allowed estimates of bed sediment storage to be obtained. Six important and unusual features of the storm event turbidity response were apparent: (1) sluggish early turbidity response, followed by a turbidity 'rush'; (2) quasi-coincident flow and turbidity peaks; (3) anti-clockwise hysteresis in the discharge-turbidity relationship on all but one event, resulting from Falling-LImb Turbidity Extensions (FLITEs); (4) increases in peak turbidity levels through storm sequences; (5) initial micro-pulses (IMP) in turbidity; and (6) secondary turbidity peaks (STP) or 'turbidity shoulders' (TS). These features provided very little evidence of a true 'first-flush' effect: instead, substantial suspended solids transport continued right through the flow recessions, and little storm-event sediment exhaustion was evident. A new, dimensionless, hysteresis index, HI(mid), is developed to quantify the magnitude and direction of hysteresis in a simple, clear, direct and intuitive manner. This allowed the degree of departure from the classic 'first-flush', clockwise hysteresis models to be assessed. Of the 15 turbidity events considered, 10 coincided with ammonia spikes of up to 6.25 mg l(-1) at Water Orton (the downstream station): this suggests that spills from combined sewer overflows (CSO) or waste water treatment works (WwTWs) are significant in the throughput of turbid waters here. Substantial ammonia peaks related most strongly to total storm rainfall receipt, of four rainfall variables considered, and significant ammonia peaks were generated even from low-magnitude storms (rainfall totals <4 mm), indicating that spills are a frequent occurrence. Local bed sediment stores appear to be limited, suggesting that other distal sediment sources, such as road networks and old mineworkings are possibly more important. Biofilms may also play a part in delaying sediment release until late in the hydrograph, and in suppressing late spring turbidity levels. Existing first-flush models appear to be an oversimplification here. Such urban headwater basin responses can provide useful insights into the generation of contaminant waves, and offer vital early-warning systems for pollution events propagating downstream.