Forest harvest and water treatability : analysis of dissolved organic carbon in headwater streams of contrasting forest harvest history during base flow and storm flow

Dissolved organic carbon (DOC) is important to ecosystem functioning and of interest to drinking water treatment as it is costly to remove from source water. The effects of land use change (e.g. forest harvest) on DOC are not fully understood. Past studies of the effects of forestry on DOC have found mixed results, and covariate effects complicate mechanistic understanding and may underlie regional differences. DOC is typically highest during storm events, and detailed time-resolved measurements of DOC over longer periods are just becoming feasible with advances in in situ sensors. In this thesis, I investigated DOC concentration and character for headwater streams of contrasting forest harvest history in the UBC Malcolm Knapp Research Forest near Maple Ridge, British Columbia, and discussed implications for water treatment through expert interviews with water treatment engineers at Metro Vancouver. I took a two-pronged approach to the field study by measuring not only temporally detailed DOC concentration at two contrasting sites, but also spatially broad DOC concentration and character at 24 sites for six time points during the same study period. From the temporal analysis, I found base flow DOC concentration to be lower in the clear-cut stream (Clear cut: 2.26 ± 0.43 mg/L; Forested: 4.30 ± 0.83 mg/L), but storm response to be stronger (mean increase DOC: Clear cut: 2.42 mg/L, Forested: 1.99 mg/L). Lower DOC concentration after cutting may be due to reductions in carbon inputs like leaf litter, though differences in drainage area and the presence of a bog lake may also account for some of the observed difference. Elevated storm responses may be due to changes in flow paths related to forest harvesting. From the broad sampling analysis, I found differences in DOC character between the two land use types: DOC in clear-cut sites was more protein-like whereas DOC in forested sites it was more humic-like, although once random factors were taken into account in generalized linear mixed models, the relationship was no longer significant. Using categorization and regression tree modeling, I found drainage area to be the most important covariate related to DOC followed by time of year (i.e. seasonality).Science, Faculty ofResources, Environment and Sustainability (IRES), Institute forGraduat