You are here:
Catchment-scale hydrologic implications of parcel-level stormwater management (Ohio USA) (journal)
Citation:
SHUSTER, W. D. AND L. K. RHEA. Catchment-scale hydrologic implications of parcel-level stormwater management (Ohio USA) (journal). JOURNAL OF HYDROLOGY. Elsevier Science Ltd, New York, NY, 485:177-187, (2013).
Impact/Purpose:
To inform the public.
Description:
A major challenge to understanding the impact of environmental management is lack of data support and techniques to quantify the effectiveness of management treatments. We address aspects of this challenge in the context of a catchment-scale study of the hydrologic, ecological, and water quality impacts due to adding parcel-level detention (as rain gardens and rain barrels) as a means to reduce stormwater quantity in separated sewer systems. We performed a pilot study in the Shepherd Creek watershed located in Cincinnati, Ohio to evaluate the practicality of voluntary incentives for stormwater quantity reduction on privately owned suburban properties. We elicited participation in a stormwater mitigation program by owners of suburban homes using two successive reverse-auctions. Two auctions were held (2007, 2008) resulting in the installation of a total of 170 rain barrels and 83 rain gardens. We monitored runoff (as stream discharge) and precipitation in two phases, three-years before and after implementation of the treatments; and at six locations within the watershed, including two control locations that received no stormwater management. To structure the experimental design we initially used a before-after-control-treatment (BACT) approach, dividing the catchment into sub-basins of relatively similar land uses though analysis of variance (ANOVA), though these linear models could not account for variation in changing rates and amounts of precipitation among storms, such that most variation in discharge was apportioned to experimental error. An autoregressive, moving-average (ARIMA) model was used to account for autocorrelation in the discharge, lagged cross-correlation of the discharges with precipitation. We found that the model was effective at partitioning variance, with the vast majority of the variance in the discharges explained by autoregressive terms (p<0.0001) and (lagged) precipitation (p<0.1000)
