The paper encompasses the analyses of sea level data measured at four tide gauges situated in the city of Split. Two long-term stations (operational since 1950s) comprise float-type instruments in stilling wells with analogous recording on a chart, while two modern stations were installed in 1999 and 2000, calculating sea level height from the measured sea and air pressure. Although analogous recording devices with a drum may suffer from inaccuracy in timing and drum rotation speed, clock errors and vertical pencil position, the largest errors in the data came from digitising process, which is subjected to individual work of a technician. The errors and high scatter in some periods can be noticed when analysing some daily and particularly hourly values, whereas monthly values have been already certified as high-quality ones by the Permanent Service of Mean Sea Level, nullifying high frequency data errors through the averaging. The data collected with pressure systems are supposed to be more homogeneous and accurate, but a problem may appear after several years of work as pressure sensors usually have a drift in time. This problem turned out to be accompanied with lower quality sensor being part of wider automatic measuring system, whereas the station which has been specially designed for sea level measurements showed no drifts in four years of operating. Therefore, one should be careful in choosing measuring equipment between various manufacturers, purchasing only high-precision and tested sensors. In addition, regular inspection should be done for digital sensors, calibrating them in a recommended time intervals.
