Document Type

Thesis

Degree Name

Master of Science (MSc)

Department

Chemistry

Program Name/Specialization

Biological and Chemical Sciences

Faculty/School

Faculty of Science

First Advisor

Scott Smith

Advisor Role

Supervisor

Second Advisor

James McGeer

Advisor Role

Co-supervisor

Abstract

The biotic ligand model (BLM) is a tool used to quantitatively evaluate how receiving water chemistry affects the bioavailability of metals. Sensitivity testing can be used to understand how the model outputs vary in response to systematic changes in water chemistry inputs. This will allow users of such models to understand how accurate their input parameters must be for a specified level of confidence in the output. Our focus was on dissolved organic carbon (DOC), which is often the most limiting data for application of BLM approaches to metals risk management. To potentially address DOC data limitations remote sensing can be explored as a tool to measure DOC, but it is necessary to understand DOC data requirements to produce water quality criteria outputs within the usually accepted prediction variance. This study begins with inputting average water chemistries to a copper BLM model for both cold and warm water regions with 1%, 10%, 25% and 50% variations in the mean values for all parameters, without considerations of correlations among parameters. The variation in the model output criterion continuous concentration (CCC) for copper as a function of DOC, and other model inputs, allows estimation of how well DOC needs to be estimated using remote sensing in a theoretical sense. It was discovered that 20% error allowance in the theoretical simulations gave accurate BLM results. To address if the 20% error holds up in practice, similar testing was completed with a real data set from 100 lakes in Ontario. Interestingly, it was discovered that DOC did not have as much influence on the CCC, and pH was the most sensitive parameter. This launched an investigation as to if the 20% error allowance would differ based on different pH levels, as results showed higher pH values having higher CCC values. The modelling then showed that although CCC is dependent on pH, the error allowance is independent of pH. Therefore, the 20% error allowance of DOC measurements through remote sensing is still a reasonable guideline, regardless of pH value.

Convocation Year

2021

Convocation Season

Spring

Available for download on Wednesday, May 11, 2022

Share

COinS