An Integrated Modeling Approach for Evaluation of Phosphorus Loading in Rural Nova Scotia Watersheds

Abstract

Residential on-site wastewater systems (OWS) are a potential source of phosphorus (P) which can negatively impact surface water quality in rural watersheds. The magnitude of P loading from OWS is typically not monitored, and is further complicated when agricultural land-uses are intermixed with residential dwellings. Watershed-scale computer simulations are commonly used tools for evaluating the impacts of land-use changes on P loading. Existing models simulate OWS P treatment via vertical flow transport in native soils. However, in Nova Scotia (NS) OWS designs rely pre-dominantly on lateral flow and imported sand filter media. In this thesis, a watershed-scale computer modeling framework for simulating P loads from agriculture and lateral flow OWS designs was developed and tested. The framework consists of the P on-site wastewater simulator (POWSIM), designed specifically for this study, which is used in conjunction with the Soil and Water Assessment Tool (SWAT) model. The POWSIM loading tool has three computational components: (i) OWS disposal field design type selection and treatment media mass calculation; (ii) disposal field P treatment dynamics; and (iii) soil subsurface plume P treatment dynamics. The active P treatment media mass and dynamics equations were developed from numerical modeling (HYDRUS-2D) and lateral flow sand filter (LFSF) OWS disposal field experiments. A 2-part piecewise linear model was found to best represent LFSF P treatment processes. Testing of the modeling framework in the mixed land-use Thomas Brook Watershed (TBW) in NS demonstrated improved simulation of baseflow total P (TP) loads in both a predominantly residential subcatchment and one dominated by agriculture over the SWAT model without POWSIM. Different residential and agricultural development and beneficial management practice (BMP) scenarios were evaluated in the TBW. Agricultural BMPs were most effective at reducing cumulative TP loads while OWS BMPs were best at mitigating in-stream eutrophication impacts. The 50 year simulation period for the various scenarios found peak OWS TP loading occurring between 25 and 50 years, suggesting that modeling for many decades is required for proper evaluation. This study highlights the importance in identifying specific water quality issues that need to be targeted prior to implementing a BMP strategy.