Bedrock geology controls on catchment sensitivity and resilience to climate change

To anticipate the consequences of climate change on hydrologic extremes, such as droughts and floods, there is a pressing need for a better mechanistic understanding of how altered atmospheric circulation patterns and an intensified hydrologic cycle translate into streamflow generation. Hydrologic tracers, such as the stable isotopes of O and H in water are key in this regard, because they allow to track the time precipitation takes to transit to the streams. Long-term datasets can thus unveil potential shifts in streamflow generation processes, although progress remains stymied by the limited availability and temporal extent of δ18O and δ2H records in precipitation and in stream water (~40 years at best). This PhD project investigates new ways to assess the impact of climate change on streamflow generation across nested catchments of different sizes and with differing physiographic characteristics.
To overcome the issue of short and truncated isotope datasets, freshwater bivalve shells are proposed as innovative long-term stream water δ18O recorders to test the hypothesis that shifts in a catchment’s hydraulic regime (as triggered by climate change) translate into a modification of the isotopic signature in stream water – with the magnitude of the isotopic signal being modulated by bedrock geology. For achieving that overarching objective, the thesis chapters address: (i) the local effects of atmospheric circulation patterns on 6 years of sub-daily precipitation δ18O and δ2H values from a sampling station located in Belvaux, Luxembourg (L), (ii) non-stationary streamflow responses in 12 contrasting nested catchments using Ensemble hydrograph separation to derive fractions of new water travelling to the streams in less than ~16 days as a new transit time metric in the Sûre river basin with 13 years of fortnightly δ18O and δ2H in streams and precipitation, and (iii) long-term trends in stream water δ18O records, reconstructed over 200 years from freshwater pearl mussel shells in four catchments in northern Sweden (S).
The results from the first chapter suggest that even if atmospheric inferences with isotopic signatures in precipitation are observed, including the air mass trajectories as an input for δ18O predictions might not provide a decisive advantage. In the second chapter, the streamflow responses to precipitation in the nested catchments were found to be highly variable, with the fraction of new water being strongly related to differences in bedrock geology. The fraction of new water (Fnew) was highest in impermeable bedrock catchments (i.e., with a dominance of marls and claystone), increasing with higher specific daily streamflow. Moreover, Fnew exhibited high variability in catchments with an important fraction of permeable sandstone and conglomerates, while it was very small in catchments with weathered bedrock (i.e., dominated by schists and quartzites). In the third chapter, the ~200 years of shell-based δ18O reconstructions in stream water revealed that summer contributions to streamflow were negatively correlated with baseflow. Seasonal contributions to streamflow also varied with catchment features, suggesting a buffering of summer low flows by lakes and peatlands over nearly two centuries in northern Sweden.
With the transferable and scalable approaches presented in this PhD project, isotope baselines in stream water can be extended into pre-instrumental times. The information gained from these timeseries is further shown to be valuable for a better anticipation of water storage and release functions under changing climate conditions, i.e., long dry spells and high-intensity precipitation events, supporting adaptive water-resource management in data-scarce regions.