Stable isotopes of oxygen (O) and hydrogen (H) in streams and precipitation are cardinal tools to assess water sources, flow paths, and age. However, their spatial and temporal variability remain largely unknown &ndash; essentially due to the limited availability of long-term records of O and H isotope signatures in precipitation. These limitations have stymied their use in studies investigating catchment response to climate change. In this study, O and H isotope signatures are conjectured to be determined by the interplay of physical processes both at synoptic and local scales. Our hypothesis is that large-scale atmospheric circulation patterns (CPs) influence well-known physical effects (e.g., temperature effect) on &delta;¹⁸O and &delta;²H values in precipitation. To test this hypothesis, high-resolution (i.e., sub-daily) &delta;¹⁸O and &delta;²H data were collected in Belvaux (Luxembourg) from 2017 to 2022 to investigate potential relationships with large-scale atmospheric circulation patterns and local-scale meteorological variables.</p> <p>Our results demonstrate CP controls on precipitation &delta;¹⁸O and on the relationship between &delta;¹⁸O and air temperature across Western Europe, with &delta;¹⁸O-temperature (&delta;¹⁸O-T) correlation slopes being larger (0.24 to 0.69) than previously reported (below 0.40). We leveraged CP-specific &delta;¹⁸O-T functions to build a multiple linear regression model for long-term &delta;¹⁸O reconstructions in Europe, based on geospatial attributes. The model was fitted on GNIP &delta;¹⁸O data, with a performance expressed by a root mean squared error (RMSE) of 2.3 &permil;, comparing to other models focused on the regionalisation of seasonal &delta;¹⁸O variability over Europe. With this novel method, intramonthly &delta;¹⁸O in precipitation can be reconstructed for ~150 years by solely relying on historical temperature records (e.g., as obtained from WMO&rsquo;s Climate Explorer web interface (<a href="https://climexp.knmi.nl" target="_blank" rel="noopener">https://climexp.knmi.nl</a>)). Our model offers an assumption-lean and thus robust alternative to reconstruct pre-instrumental precipitation &delta;¹⁸O chronicles.