The “gold standard” of stormwater control measure (SCM) performance monitoring is to collect and analyze water quality data as flow-weighted event mean concentrations (EMCs). EMCs may be generated by post-storm compositing of discrete samples as an alternative to collecting flow-weighted composite samples using automated equipment with integrated flow meters. Coincidentally, a standard procedure is not documented in industry-relevant literature to perform the requisite post-storm flow-weighting calculations. Lack of a documented procedure may deter data collection for agencies new to monitoring, or for complicated site conditions. It also leads to different approaches to post-storm flow-weighting that may influence resultant EMCs and mass loads. An open-source python script and related web application have been developed to enable consistent, transparent, easily applied calculations for post-storm flow-weighting and compositing and/or to generate an EMC from a pollutograph. The python script is available in a Github repository so that users may integrate the code with other bespoke data processing tools. The web app provides flow-weighted compositing instructions based on a user-uploaded hydrograph and times of sample collection, or returns an EMC based on a user-uploaded hydrograph and pollutograph. Total hydrograph volume is also returned so that users may determine a mass load from the EMC. A sensitivity analysis was performed examining the influence of post-processing assumptions on the calculated EMC value. Representative flow attribution (i.e., whether a discrete sample represents the beginning, middle, or end of a flow interval), volume calculation from flowrate integration (e.g., Reimann sum vs trapezoidal approximation), and relative resolution of discrete sample and flowrate data were assessed for relative impact on the EMC. Preliminary results indicate that various post-processing schemes can yield significantly different EMC values, on the order of 60% error from baseline. EMC agreement was most impacted when the resolution of flow data was low relative to discrete samples.
