Development of Statistical and Mass Balance Approaches for Assessing and Predicting Chloride Concentrations in Halifax Lakes

Abstract

Lake chloride concentrations ([Cl]) in Halifax and other urban centres where road salt is used is a growing environmental issue. This thesis used Cl data available seasonally from 57 Halifax lakes to model lake [Cl] using both linear regression methods and a statistically averaged mass balance model. The watershed variables found to be most predictive of mean lake [Cl] were the % urban coverage (R2 = 0.64), road density (m/m2) (R2 = 0.55), and stormwater pipe density (m/m2) (R2 = 0.51) in a watershed. The combination that best predicted mean lake [Cl] was stormwater pipe density, ramp density, and % rock (of the watershed’s surficial geology) (R2 = 0.71). A mass balance modeling approach was developed and used to estimate Cl loading rates for four land use categories (rural 1-3 g/m2/yr, commercial 0-9 g/m2/yr, residential 97-162 g/m2/yr, and road 804-964 g/m2/yr). These findings further support a growing body of literature that associates roads and development with elevated and increasing freshwater [Cl].