G2DC-PL+: a gridded 2 km daily climate dataset for the union of the Polish territory and the Vistula and Odra basins

G2DC-PL+, a gridded 2 km daily climate dataset for the union of the Polish territory and the Vistula and Odra basins, is an update and extension of the CHASE-PL Forcing Data - Gridded Daily Precipitation and Temperature Dataset - 5 km (CPLFD-GDPT5). The latter was the first publicly available, high-resolution climate forcing dataset in Poland, used for a range of purposes including hydrological modelling and bias correction of climate projections. While the spatial coverage of the new dataset remained the same, it has undergone several major changes. (1) the time coverage was increased from 1951-2013 to 1951-2019; (2) its spatial resolution increased from 5 to 2 km; (3) the number of stations used for interpolation of temperature and precipitation approximately doubled; and (4) in addition to precipitation and temperature, the dataset consists of relative humidity and wind speed data. The main purpose for developing this product was the need for long-term areal climate data for earth-system modelling, and particularly hydrological modelling. Geostatistical methods (kriging) were used for interpolation of the studied climate variables. The kriging cross-validation revealed improved performance for precipitation compared to the original dataset expressed by the median of the root mean squared errors standardized by standard deviation of observations (0.59 vs. 0.79). Kriging errors were negatively correlated with station density only for the period 1951-1970. Values of the root mean squared error normalized to the standard deviation (RMSEsd) were equal to 0.52 and 0.4 for minimum and maximum temperature, respectively, suggesting a small to moderate improvement over the original dataset. Relative humidity and wind speed exhibited lower performance, with median RMSEsd equal to 0.82 and 0.87, respectively. The dataset is openly available from the 4TU Centre for Research Data at https://doi.org/10.4121/uuid.a3bed3b8-e22a-4b68-8d75-7b87109c9feb (Piniewski et al., 2020).