The scope of this paper is to merge in situ sea level observations with satellite altimetry observations to give a consistent view of the nontidal sea level fluctuations in the North Sea. The detided sea level variability is dominated by climate induced surges that propagate counterclockwise around the North Sea with speeds of ~25 m/s and spatial scales of 100--1000 km. The individual sea level recorders agree very well with the satellite observations off the coast with correlation coefficients of 0.9. The sampling characteristics of the satellite and the in situ observations are used to construct a regression model. Regression of TOPEX/POSEIDON sea surface height data onto coastal sea level observations gives a mean hindcast skill of 0.7 and RMS errors of 8.2 cm for the North Sea. Validation of the model against independent ERS satellite crossover observations yields a mean RMS difference of 13 cm including the 4--7 cm noise in the satellite observations. The modeling of the true sea level is most likely better than 10 cm. The performance is presently better than the Danish storm surge model, suggesting that much can be gained by including the spatial information from the satellite observations in the models.