This dataset consists of estimates of standard errors of combined measurement and sampling uncertainties of land surface air temperature anomalies, corresponding to the DCLSAT anomalies used in the estimation of DCENT_MLE_v1.0. Estimates are provided for each month from 1850 to 2023 and for each 5° latitude by 5° longitude grid cell with available data. However, available estimates are not spatially and temporally complete since the observational coverages of source datasets used are not spatially and temporally complete.
Measurement and sampling uncertainties were estimated using the approach of Brohan et al. (2006). Intersite correlations used were inferred from CRUTEM5 data. Wherever Chan, Gebbie and Huybers (2024) data was available, the number of observations used corresponds to the ensemble mean of Chan, Gebbie and Huybers (2024), excluding observations that did not pass quality control checks according to the unhomogenized GHCNv4 data. Otherwise, the number of observations used corresponds to unhomogenized GHCNv4 data after quality control checks. For "tier-1" stations as defined by Chan, Gebbie and Huybers (2023), uncorrected sample variances by calendar month during the climatological period were calculated using the ensemble mean of Chan, Gebbie and Huybers (2024). For other stations, uncorrected sample variances by calendar month during the climatological period were calculated using homogenized GHCNv4 data. A preference of east over west and north over south was applied when assigning weather stations to grid cells.