The first flush load of urban surface runoff

The treatment of urban surface runoff is steadily becoming common practice. Good management of treatment works requires an understanding of the first flush phenomenon of wet weather flow in urban drainage systems. The objective of this paper is to analyse evidence for the existence and nature of the first flush load of pollution input into drainage systems, using data on suspended solids, conductivity, pH, and temperature of storm surface runoff. The data were collected continuously for almost a year for single road inlets at two urban asphalt catchments (in Belgrade, Yugoslavia and Lund, Sweden). Cumulative load curves were constructed for all monitored water quality characteristics, and the pollution load carried by the first 20% of runoff, FF20, was calculated for each recorded event. FF20 values were analysed using standard statistical methods, including multiple regression. Only slight first flush effects for suspended solids and conductivity of storm runoff were observed at both catchments. No first flush effect was recorded for pH or temperature. The rainfall and runoff characteristics which influence the first flush phenomenon are different at each catchment studied, although the catchments have very similar characteristics. It was concluded that: (1) the first flush, if strongly present at the end of a drainage system, is not generated by the first flush of pollution input; (2) regression curves are not very reliable for prediction of the first flush load of pollution input into drainage systems.