Analysis of Water Balance Based on Long-term Observations and ERA5-Land Reanalysis in the Podkamennaya Tunguska River Basin

The sparse ground-based network for precipitation and river runoff observations in Siberia and the heterogeneity of observation series necessitate the use of reanalysis data. The study investigates the accuracy of reproducing the water balance components of the Podkamennaya Tunguska River basin for 1951–2022. These components were calculated based on ERA5-Land reanalysis data as compared with ground-based observations. The coefficients of correlation between the annual runoff according to reanalysis data and observations at gauging stations are 0.57–0.80. The normalized root-mean-square error nRMSE of annual runoff according to the reanalysis data for three catchments varies from 19 to 22%, increasing from east to west. For the catchment area limited by the Baikit–Kuz’movka factory, it reaches 40%, with the minimum correlation coefficient r = 0.57. A comparison of observational and reanalysis precipitation data showed that the modeling used values from the only reference weather station, which significantly reduced the model accuracy.