Shafii, M., Slowinski, S., Bhusal, Y., Kaykhosravi, S., Hitch, C., Withers, W., Rezanezhad, F., Van Cappellen, P. (2022) Statistical prediction of phosphorus export and speciation in urban catchments: model development and potential applications for stormwater management. In Fall Meeting 2022. AGU. https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1170188

Quantifying the export of urban nutrients is challenging due to differences in watersheds physio-climatic conditions and anthropogenic factors, including stormwater management. As a result, accurate assessment of nutrient transport in urban landscapes at different spatio-temporal scales requires local stormwater monitoring and modeling studies. Our research focuses on phosphorus (P) as a major limiting nutrient driving eutrophication of freshwater environments. We collected discharge, total suspended sediment (TSS), total P (TP), and dissolved reactive P (DRP) data in three sewersheds within the greater Toronto metropolitan area in southern Ontario, Canada. The data were used to develop multiple linear regression (MRL) models to predict loadings of these P fractions. As an application of these models, we employed the MLR models to assess P retention performance of a stormwater pond located downstream of one of our study’s sewersheds. MLR models properly forecasted event-scale loadings during the monitoring period, with Nash-Sutcliffe Efficiency (NSE) values of 0.5-0.82, 0.41-0.75, and 0.46-0.8 for TSS, TP, and DRP, respectively. Expanded loading predictions for the past decade indicated one order of magnitude difference in the export of TSS and TP among our three study sewersheds. We attributed this variability to sewershed age and landuse where the oldest mixed industrial-residential sewershed exported the largest amount of TSS and TP, and the younger fully-residential sewershed exported intra-year invariant DRP loads. Model predictions also revealed that the pond was a significant sink of TSS and P, which was corroborated by the analysis of pond sediment cores taken concurrently with the stormwater monitoring. As our MLR models quantify sewersheds annual loadings of TSS and P fractions based on precipitation and temperature records, they are transferrable to other similar data-scarce catchments, enabling stormwater management teams to make predictions needed for infrastructure design or the prioritization of nutrient control in urban catchments.

https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1170188