Characterising and modelling urban runoff quality for improved stormwater management (2016)
Type of ContentTheses / Dissertations
Thesis DisciplineCivil Engineering
Degree NameDoctor of Philosophy
PublisherUniversity of Canterbury
AuthorsCharters, Francesshow all
Runoff from impermeable urban roof, road and carpark surfaces are key contributors of sediment and heavy metals to urban waterways, causing acute and chronic adverse effects on the aquatic ecosystem. Characterisation of the untreated runoff quality is necessary to guide the selection of effective and efficient stormwater management options that can reduce the pollutant load. Rainfall characteristics, such as intensity, storm duration, rainfall pH and the length of antecedent dry periods, are also known to be key drivers of stormwater pollution build-up and wash-off processes. However, there is limited knowledge of how low intensity rainfall climates, such as is found in Christchurch, New Zealand, influence pollutant generation. Current stormwater pollutant load models typically aggregate the contributing surface areas by land use or are annual load models that use per area pollutant load factors. While annualised load models are useful in quantifying the cumulative effects of pollutants from stormwater discharges on the receiving environment, storm-event based models are needed to identify the peak concentrations responsible for acute toxicity effects as well as informing design criteria of any stormwater treatment system based on pollutant characteristics. As pollutant build-up and wash-off processes are known to differ for various surface materials, load prediction from an individual surface enables targeting of ‘hotspot’ surfaces and assists with selecting appropriate management options for that particular surface’s characteristics. Thus, the main objective of this research was to characterise pollutant generation in a low intensity rainfall climate from different impermeable urban surface types and then develop an event-based pollutant load model to predict pollutant loads from those different surface type within a catchment. The research therefore had the following elements: (1) characterisation of sediment and heavy metal concentrations in untreated urban runoff from specific impermeable urban surfaces, (2) characterisation of particle size distribution (PSD) variance in the runoff, (3) development of an event-based model for total suspended solids (TSS), copper and zinc event loads using rainfall characteristics as predictor variables, and (4) application of the model to case study catchments. Untreated runoff samples were collected from 25 rainfall events from four impermeable surfaces (concrete tile, copper, and galvanised roofs and an asphalt road) located within 320 m of each other in a residential/institutional catchment in Christchurch. Pollutant concentrations were found to be significantly different between surfaces, confirming that quantification and prediction of pollutant loads from urban surfaces should take account of the different surface materials. The highest concentrations of TSS were seen in the asphalt road runoff under both initial and steady state conditions. As the road TSS was substantially higher than roof TSS, treatment of road runoff prior to it mixing in the kerb and channel with roof runoff may be warranted to reduce the ‘treatable’ volume for TSS. Substantial PSD variation was observed for each surface and between events, particularly for coarser road and concrete roof surfaces. Implications of this variation result in a wide range in predicted treatment performance since most sediment treatment is dependent on particle size and retention time. This suggests that short-retention treatment devices carry a high performance risk of not being able to achieve adequate TSS removal across all rain events. Copper and galvanised roof runoff had the highest copper and zinc concentrations, respectively, followed by road runoff. The majority of the copper in the copper roof runoff was in dissolved form (average of 77%), while only 28% of the road runoff copper was dissolved. Likewise, almost all (average of 99%) the zinc in the galvanised roof runoff was in dissolved form, while only 42% was dissolved in road runoff. As well as contributing to ecotoxicity in the receiving environment, dissolved metals in stormwater runoff can be more difficult to treat as majority of the standard stormwater treatment systems are based on filtration or settling processes that primarily aim to remove sediment. Therefore, source reduction of roof-contributed copper and zinc should be targeted via roof material replacement or painting. Road runoff treatment systems should consider processes that facilitate both dissolved and particulate metals, as removal of particulate-associated metals via settling or filtration may not adequately reduce metals loads entering urban waterways. The event-based pollutant load model, Modelled Estimates of Discharges for Urban Stormwater Assessments (MEDUSA), developed as part of this research, was found to be effective at modelling TSS, and total and dissolved copper and zinc loads under a low intensity rainfall climate. MEDUSA was calibrated against observed data and applied to two case study catchments in Christchurch, New Zealand. The model clearly identified the spatial distribution of pollutant generation across each catchment’s individual roof, road and carpark surfaces, and was found to be most sensitive on an eventto- event basis to rainfall intensity and duration, both factors which are expected to change under future climate change scenarios for Christchurch. The MEDUSA model can be further used to explore the effectiveness of different management scenarios on reducing pollutant loads and also employed for guiding the prioritization, location and selection of stormwater treatment systems to ultimately improve urban waterway health through reduction of untreated stormwater-generated pollutants. Enhancements to the MEDUSA framework can be advanced by incorporating other pollutants of concern, such as nutrients, emerging contaminants and other metals of concern. Overall, this research contributes to scientific understanding of both at-source stormwater character and the effectiveness of using rainfall characteristics to predict pollutant loads based on simulating build-up and wash-off processes. Specifically, the research has identified how urban surface types differ in their pollutant generation, (i.e. the relative influence of rainfall and material characteristics in generation of both sediment and metal pollutants); how heavy metals partition between particulate and dissolved state in untreated runoff from different urban surfaces (with implications for metals treatment selection); how particle size fractionation differs during and between rain events from different urban surfaces (with implications for sediment treatment system performance); and the importance and effectiveness of using a disaggregated model (i.e. individual surface-based modelling) as the pollutant generation processes differ significantly between different urban surface types.