PAUL: A Novel Autosampler for High-Resolution Isotopic Monitoring in Catchment Hydrology

Understanding how streamflow is generated is essential for managing water quantity, quality, aquatic ecosystems and drinking water resources. Stable water isotopes are widely used to separate streamflow into event and pre-event water, offering valuable insights into catchment storage dynamics, water ages and transit times of water. Typically, precipitation or stream water for isotope analysis are sampled using autosamplers to reduce effort and ensure an adequate temporal resolution.

However, hydrograph separation using stable isotopes is often limited by reliance on single-point rainfall sampling, which therefore assumes that precipitation inputs are spatially uniform across the catchment. This can introduce substantial errors, as spatial variability in the isotopic composition of precipitation—even within small catchments—may lead to misestimations of event water endmember contributions. Also, the various transit time models may experience biases due to an incorrect precipitation input time series of stable isotopes. Furthermore, typical autosamplers are susceptible to evaporative losses from the stored water samples, resulting in isotopic fractionation and compromised data integrity.
To solve these difficulties, we have developed and deployed the low-cost, evaporation-proof Portable Autosampler for Liquids (PAUL). Nine PAUL units were distributed across the 1.5 km² Krummenbach sub-catchment of the Brugga watershed, a mountainous headwater catchment located in the Black Forest, Germany. Eight units measured precipitation and one sampled streamflow, with biweekly collection over the course of one month.

Our findings show that spatially distributed, evaporation-secure sampling significantly improves the characterization of event water inputs and reduces uncertainty in hydrograph separation. The PAUL system provides a robust and accessible solution for high-resolution, catchment-scale isotope monitoring, providing spatial and temporal coverage that was previously unfeasible with standard autosamplers. This approach advances process-based hydrology by increasing the accuracy and reliability of isotope-based precipitation and streamflow analyses.