Postglacial land uplift in Estonia based on four precise levelings

•Vertical velocities from four precise levelings in 1933-2010 were computed.•Uncertainties of first three levelings are notably larger than a priori assumed.•Modelled uplift values range from − 0.6 to + 2.0 mm/yr, uncertainty is ± 0.6 mm/yr.•New model fits with the land uplift model NKG2005LU within ± 0.2 mm/yr.

The vertical velocities of the fundamental benchmarks of Estonian 1st order leveling network were estimated, based on the four precise leveling campaigns from 1933 to 2010. The kinematic least squares adjustment of the network was used, where heights and velocities were introduced as unknown parameters. For detection of outliers, Baarda's data snooping method was applied. Estimation of variance components by the Helmert's and the IAUE methods provided realistic weights in the network adjustment and revealed also that the observation errors of the first three levelings are up to 3 times larger than was assumed a priori.To obtain apparent uplift rates and fix velocity in kinematic adjustment, the velocity value + 2.1 mm/yr of the Ristna tide gauge on island Hiiumaa was transferred to the nearest stable benchmark by using precise levelings from tide gauge to the national height network. Average standard deviation of velocities at benchmarks was estimated to be ± 0.5 mm/yr.Based on apparent vertical velocities of the benchmarks, an interpolated land uplift surface was created using “kriging” and “minimum curvature” gridding methods. Although the two methods gave similar surfaces (RMS of differences was below ± 0.1 mm/yr), the kriging solution was used in comparison with the results of the earlier studies. The uncertainty of new velocity surface was estimated to be ± 0.5 to ± 0.7 mm/yr.The new land uplift model fits well with the Fennoscandian land uplift model NKG2005LU. The RMS of the differences was ± 0.2 mm/yr. Good agreement was also confirmed with the earlier land uplift maps of Estonia (RMS of differences ± 0.6 mm/yr). However, clear disagreement was noticed when tide gauge observations from several studies were used in comparison (RMS of differences up to ± 1.0 mm/yr). Apparently, different velocity solutions of Estonian tide gauges are systematically biased; the reasons for these biases need further investigation.