Estimating the fraction of young water in streamflow in mesoscale Precambrian Shield catchments in Northeastern, Ontario, Canada

Abstract

Understanding how catchments respond to precipitation under rapidly changing climate and land-uses, is key to managing water resources particularly in the understudied northeastern Ontario Canadian shield landscape. Despite the importance of understanding how Precambrian Shield catchments generate streamflow, our understanding beyond the small headwater scale is limited. The complexities of water storage and movement in large basins of heterogeneous landscape, has called for the application of a variety of enhanced techniques that give a better understanding of the flow paths and storage dynamics, which conventional modelling approaches commonly fail to capture. This study focuses on investigating catchments response to precipitation, by estimating the fraction of young water (𝐹𝑦𝑤) and corresponding river isotope damping ratios using flow-weighted and unweighted isotope data. We use 5 years of modeled precipitation isotope (δ18O) and 3 to 5 years of streamflow isotope (δ18O) records from 16catchments, ranging in size from 27 km² to 6,919 km². We employed Pearson correlation and hierarchical partitioning regression (HP) to identify the relationship and the contributory effect of catchment predictors on 𝐹𝑦𝑤. Our results showed 𝐹𝑦𝑤 values were between 18 % and 51 %,with no significant difference between unweighted (𝐹𝑦𝑤) and flow-weighted (𝐹𝑦𝑤∗ ) estimates (r= 0.99, p < 0.001). The average values of 𝐹𝑦𝑤∗ for 7 catchments and 𝐹𝑦𝑤 for the 16 catchments are 31 and 33 %, respectively. This indicates that around one-third of streamflow consists of water aged between 2 – 3 months old. Estimated 𝐹𝑦𝑤 was found to be associated with different landscape, landcover and hydrometric variables. Hierarchical partitioning analysis identified sparse treed area, mean slope, drainage area, lake area, and forest area as the key predictors of𝐹𝑦𝑤, accounting for 71.2 % of total variance of response variable 𝐹𝑦𝑤. Predictors such as mean slope, drainage area, lake area, and forest area showed a negative relationship with 𝐹𝑦𝑤,suggesting a reduction in 𝐹𝑦𝑤 as these four predictors increases. Conversely, 𝐹𝑦𝑤 increases as dispersed tree (sparse treed) area increases. This study revealed 49 to 82 % of the streamflow in the basin was contributed from slow lateral subsurface baseflow. These results set the foundation for a more focused watershed hydrological study in northeastern Ontario, revealing how responsive each catchment is to precipitation and identifying the important predictors and their effects on 𝐹𝑦𝑤 across SNF basin.