The importance of events for total phosphorus export from agricultural catchments in southern Ontario

Eutrophication of freshwater ecosystems is symptomatic of declining water quality and is often constrained by the availability of phosphorus commonly a limiting nutrient of biomass and algal growth. In the Lake Erie basin, water quality issues including eutrophication have re-emerged after abatement achieved through management strategies carried out in Canada and the United States. In Ontario, new methodologies for accurately estimating phosphorus export from streams draining into Lake Erie are needed to inform related monitoring and management programs. Of particular interest are surrogate models for predicting stream phosphorus fluxes that account for water quality dynamics during transitory runoff events that export more solutes than baseflow conditions. Typical surrogate models use relatively sparse data and linear regression to predict phosphorus fluxes using discharge, turbidity, and seasonality. These approaches do not consider shorter time scale water quality dynamics (e.g., hysteresis) in predictions and under-utilize sensor data. This study used generalized additive models (GAMs) to predict total phosphorus (TP) export from 11 agricultural catchments in southern Ontario that were monitored from 2015-2020 by the Ontario Ministry of Environment, Conservation, and Parks. Besides familiar covariates, these GAMs considered stream baseflow proportion, antecedent wetness conditions, and hysteresis. This study assessed the potential benefits of this modelling approach and quantified the proportion of seasonal and annual TP budgets related to fluxes from runoff events. The GAMs are a flexible alternative to linear models that may appeal to practitioners. The inclusion of novel covariates improved TP flux predictions. Preliminary results indicate that most of the seasonal and annual TP budget is associated with runoff events. In some cases, more than 95% of seasonal TP fluxes were associated with a few large runoff events. More work is needed to characterize the size and intensity of extreme events that account for such a dramatic proportion of the annual TP budget.