Sea level variability and trends in the Adriatic Sea ín 1993-2008 from tide gauges and satellite altimetry. Physics and Chemistry of the Earth

The scope of this paper is to give a consistent view of the low frequency sea level variability in the Adriatic Sea from both satellite altimetry and tide gauge records. We analyze 16 years of sea level observations from multi-satellite altimetry and tide gauge records in the time interval 1993–2008. First, the impact of the corrections applied to the altimetry-derived sea level variations and the consistency of the altimetric and the tide gauge sea level observations are evaluated. Both observations are then used to characterize sea level trends, interannual variability and land vertical motion in the Adriatic region. Eight tide gauges along the coast show very coherent interannual sea level variations, with an increase in sea level before 2001 and decrease afterwards. The average of the eight de-seasoned time-series agrees with the basin average of the altimeter data, with correlation coefficient 0.84 and root mean square difference 12 mm. The linear change is higher for altimetry than for tide gauges and strongly depends on the length of the time-interval, being 3.2 ± 0.3 mm/yr and 1.9 ± 0.3 mm/yr in the interval 1993–2008. The steric contribution to sea level change correlates well with the sea level suggesting that the low frequency variability is likely related to oceanic and climatic processes and mainly due to temperature and salinity variations. The decadal sea level variability is correlated in Adriatic and Eastern Mediterranean, anti-correlated in Adriatic and Ionian Sea. At a given location, the trend of the differences of sea level observations by tide gauges and co-located satellite altimetry gives the vertical land motion, if we assume that the sea level signals are truly common. We find trends statistically significant at the 90% confidence level at two locations, that indicate land uplift along the eastern coast in Rovinj (3.0 ± 1.2 mm/yr) and land subsidence in Marina di Ravenna (1.5 ± 1.1 mm/yr), while at other locations, e.g. in Trieste (1.3 ± 1.1 mm/yr) the significance is lower. The results agree in general in sign with GPS derived rates, but not in magnitude, like in Marina di Ravenna, where the strong subsidence measured by GPS is related to the local anthropogenic subsidence. The differences are partly explained by the spatial distance between the tide gauge and the co-located altimeter locations.