Daily timeseries for stable water isotopes (18O) in rainfall and snowfall (precipitation). The dataset is based on precipitation sampling at the Pallas research catchment. This data introduces gap filling for days where a sample for the stable water isotopes is not available. The gaps are filled based on linear regression models, which relate stable water isotope composition in precipitation 18O samples with air temperature and atmospheric water vapor 18O composition recorded at the site at hourly resolution. The dataset for daily precipitation sampling is described and available: to ADD DATASET and DOI The dataset for atmospheric water vapor 18O composition is described and available: to ADD DATASET and DOI The dataset for air temperature used in the analysis is provided by the Finnish Meteorologial Institute, Kenttärova station (ID 101987), available: https://www.ilmatieteenlaitos.fi/havaintojen-lataus. Detailed description of how 18O for days without a precipitation sample were estimated: The daily precipitation samples were distributed to hourly values based on the collection period. The most typical collection period was 24 hours between 8am - 8am local time. Some samples were accumulated over longer times. Hourly timeseries of vapor 18O and air temperature were used as the explanatory variables for linear regression models. The hourly values for vapor 18O and precipitation were averaged over the 24 hours when the precipitation 18O was sampled to match the timing of the variability in the explanatory variables, and the sampled precipitation event. The events used in the regression model are labeled "EVENT" in the datafile. The precipitation samples accumulated over longer that 24 hour periods were not used in the regression, and were labeled "COMPOSITE". The days that did not have any samples, were labelled "MISSING". The daily O18 value was assigned with the following priority: (i) if there was a precipitation sample for the day, the sample 18O value was used. Applies to days labelled both "EVENT" and "COMPOSITE" (ii) if a precipitation sample was missing, the 18O value was estimated based on linear regression of vapor 18O or air temperature. For the days with vapor 18O data was available, vapor data and model was used.

2023