The pervasive catchment-scale impacts of stormwater runoff on urban streams are widely recognised; yet restoration activities remain targeted at reach scale features, oft with limited ecological success. Research has shown stormwater control measures (SCMs) can reduce pollutant loads and peak flows at the scale of individual systems. Few studies have, however, empirically assessed their effects on runoff or receiving water quality and quantity when dispersed across an entire catchment.
We aimed to restore the hydrology and water quality of several small urban streams via the catchment scale retrofit of stormwater infrastructure. We selected 11 first-order streams located in and around the Dandenong Ranges, Melbourne. We monitored stream hydrology and water quality in 3 forested reference streams, 2 urban control streams and 6 ‘intervention’ streams before, during and after SCM installation. We assessed stream water quality including total and dissolved nutrients, electrical conductivity, pH and discharge in all catchments. Between 2009-2017, we installed over 620 dispersed SCMs projects in the intervention catchments; including raingardens, biofilters, swales and rainwater tanks.
We modelled the performance of SCM projects and their nested spatial and temporal installation to quantify their hydrological performance as a predictor of our intervention. We calculated several indicators of hydrological performance and used these in a Bayesian hierarchical model to predict the observed changes in water quality through time. SCMs achieved sufficient runoff retention to result in measurable improvements in water quality in only some experimental catchments. Our ability to install sufficient SCMs to elicit larger changes was limited by a lack of sufficient public open space for the SCMs, and a lack of demand for the excess stormwater. Our study suggests that dispersed stormwater control is only likely to successfully protect or restore urban streams where there is sufficient space for SCMs and demand for retained stormwater.