Evaluation of street sweeping effectiveness as a stormwater management practice using statistical power analysis

Although street sweeping is commonly regarded as a cost-effective stormwater best management practice, there is little quantitative evidence that street sweeping directly improves runoff water quality. In this paper, several previous street sweeping studies were re-evaluated using statistical power analysis. Two-group, independent-sample one-sided t-test power analyses were performed using log-transformed event mean concentrations (EMCs) of total suspended solids, suspended sediment concentration or chemical oxygen demand. The effect size between the two groups was estimated using the sweepers' pick-up efficiency, which showed that the failure to detect the difference between mean EMCs of the two sample groups (i.e., unswept and swept groups) is likely due to limited sample numbers. Too few samples, which also resulted in a high coefficient of variation, were analysed to detect the likely difference between swept and unswept observations. In addition, the temporal gap between street sweeping and subsequent storm events was not controlled to improve statistical power.