1 School of Environment and Sustainability, Global Institute for Water Security, University of Saskatchewan, Saskatoon, Saskatchewan, Canada. Electronic address: danielle.spence@usask.ca.
2 School of Environment and Sustainability, Global Institute for Water Security, University of Saskatchewan, Saskatoon, Saskatchewan, Canada.
3 Department of Building, Civil, and Environmental Engineering, Concordia University, Montréal, Quebec, Canada.
4 Department of Geography and Environmental Studies, Wilfrid Laurier University, Waterloo, Ontario, Canada.

Drinking water treatability is defined by multiple parameters that are strongly impacted by climatic and anthropogenic drivers. Working in a shallow reservoir in the Canadian Prairies, generalized additive models (GAMs) were applied to a 33-year dataset to identify drivers of interannual variability in multiple indicators of drinking water treatability. Interannual variability in treatability indicators was substantial. In the most extreme years, annual means were 2.9, 2.4, 1.5, and 1.7 times higher than the long-term averages for odour, turbidity, dissolved organic carbon (DOC), and total dissolved solids (TDS), respectively. GAMs showed that these treatability indicators are highly responsive to two modes of climate variability: the El Niño-Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO). Specifically, cool/wet cycles contributed to elevated turbidity, odour, and DOC, while warm/dry cycles contributed to higher levels of TDS, DOC, odour, and turbidity. The effects of climate variability equate to 0.5-1.7 times the long-term average for each treatability indicator. Hydrological management and nutrients also play a key role, with effects equating to 0.10-1.1 times the long-term average in treatability indicators. Together, these findings show these predictors contributed to substantial variability in water treatability. Although shallow systems in dryland regions may represent extreme examples of climate sensitivity, extreme climatic conditions are expected to become more common, posing substantial risks to water treatment. This study is the first to use GAMs to provide long-term evidence of impacts of natural climate variability and water management to drinking water treatability, potentially offering early warning about changes to source water quality.