Beyond the design event : sediment pollution movement in SuDS
Abstract
Sustainable urban Drainage Systems (SuDS) present a ‘blue-green’ method of urban
stormwater management that is increasingly implemented in the UK and worldwide.
SuDS mimic natural vegetated flow paths and are designed to manage the increase in
stormwater quantity and degradation of stormwater quality resulting from development
(urbanisation). They have been widely implemented across the UK over the last 15
years, to aid compliance with the EU Water Framework Directive (2000) standards for
river water quality. Given the increasing maturity of UK SuDS, there is growing
concern over the long-term performance efficiencies of these assets/networks,
particularly the variability of treatment efficiency over multiple flow events. Providing
the field monitoring evidence base to address this concern forms the aim of the present
thesis. Emphasis is placed upon understanding SuDS asset/network sediment detention
efficiency, as the majority of urban pollution is adsorbed to sediment material rather
than transported and treated as solute. A novel tracer method is, therefore, developed
and employed to identify and quantify sediment processes for mature UK case study
SuDS.
SuDS design manuals (CIRIA 2015, Water by Design 2006, USEPA 2009) present
expected or reported sediment detention and pollution mitigation levels for specific
SuDS assets. For example, the expected Total Suspended Solids (TSS) removal for a
swale has been defined as 90% in the WSUD (Water Sensitive Urban Design) technical
manual (Leisenring et al. 2013). Yet, these treatment efficiencies are based on a single
‘design’ rainfall-runoff event through the system; hence, fail to consider the sensitivity
of SuDS performance to non-design and multiple repeat events over the long term
design life of the SuDS asset. As natural variability in rainfall affects pollutant washoff,
shear stress for entrainment, conveyance, deposition, loss of treatment capacity etc., the
research presented in the present thesis intensively monitored four established UK
SuDS networks for 6-12 month timeframes. Bulk sampling data highlight that TSS
treatment is highly variable, ranging from highly effective (>80%) to inefficient
(<20%). Similar variability is found in sediment deposition rates (on average: 0.4-1.1
kg/m2/yr), providing insight on temporal and asset dependency of fine sediment
detention, including related treatment efficiency and long term loss of capacity.
Wetlands illustrate the most effective (mass) sediment detention per area (>1kg/m2)
while the swales detain the least (<0.8kg/m2). To advance the volumetric data noted above, source-sink routing of diffuse fine
sediment pollution required development of tracer methodology appropriate to use in
SuDS. This dictated use of Rare Earth Oxides (REO) as fine sediment tags; although
their use in an urban environment is new, it provides long term trace and experimental
replicability results without loss of provenance, signature degradation or loss of tag
material. Thus, unique time-stamped and source-specific identifiers have been used
monitor their movement into and through each SuDS over a 6-12 month period. Use of
REO tagged sediment data permits mass balance analysis of fine sediment through the
monitored SuDS assets and networks. Data clearly illustrate that sediment is not fully
detained (as assumed in SuDS design); rather, sediment is re-entrained and re-deposited
multiple times over multiple flow events. Residence time of sediment within a full
SuDS network is found to be as short as 12 weeks, raising concern over treatment
capability. Reviewing this at finer asset-based resolution, detention efficiency and
conveyance rates appear unique to each asset. Generally linear wetland and swale assets
demonstrate the greatest (tagged) sediment detention efficiency (>70%) while (the
monitored) wetland assets decline to below 50% efficiency over the first 12 months and
ponds demonstrated negligible sediment detention efficiency (<10%).
As 80% of urban pollution is conveyed adsorbed to fine sediment, the sediment
conveyance pattern through SuDS assets has been analysed to define the pollutant
concentration levels and trends of detained sediment. Pollutant levels show no
consistent trend across SuDS assets. Results illustrate that sediment pollutant
contamination shows an influence from particle size distribution and mass deposition as
well as asset design. Analysis shows the most numerous significantly elevated sediment
pollutant concentrations within the linear wetland, with Fe, Ba, Cr, Cu, Zn, K and P
demonstrating average concentrations above contaminated land thresholds. Enrichment
and geoaccumulation indexing of pollutants illustrates Fe, Zn, Cr, Ba, Cu and P to be
pollutants of concern, with Fe, Pb, Ni, Cr, Mn, Zn, Ba, Cu, Ni and P identified as
hotspot pollutants in one or multiple SuDS assets.
Cross-correlation of rainfall and flow characteristics with asset/network detention
efficiencies were used to define key drivers of multi-event sediment conveyance.
Outcomes highlight three variables of strong influence: the number of rainfall and flow
occurrences; the antecedent dry days; rainfall clustering characteristics. Weaker
correlations are found with flow characteristics (number of flow events, depth and
velocity – leading to Fr and Re values) and modal particle size. These influential forcings have then been considered with respect to a selected standard SuDS pollution
treatment process (MUSIC k-C*) to identify the compatibility for multiple rainfallrunoff
event SuDS fine sediment and pollutant simulation. The research provides
multiple event SuDS stormwater treatment efficiencies that can inform improved SuDS
design and maintenance planning by engineering consultants, Local Authorities,
environmental regulators and SuDS asset managers.