Low frequency Mediterranean sea level variability: The contribution of atmospheric pressure and wind

Abstract

We examine 44 yr (1958–2001) of extensively validated model data with the aim of characterizing the contribution of atmospheric pressure and wind to low frequency (the seasonal cycle and lower) Mediterranean sea level variability. The sea level data set was produced within the HIPOCAS Project framework, by means of a long-term barotropic run of the HAMSOM model, with a 1/4° × 1/6° spatial resolution. The atmospheric pressure and wind fields used to force the 44-yr ocean hindcast were produced by dynamical downscaling (1/2° × 1/2°) from the NCEP/NCAR global reanalysis, using the atmospheric limited-area model REMO. Results show that during the examined period, the effect of the mechanical atmospheric forcing has been in the sense of lowering sea level at a mean rate of − 0.6 mm/yr, mainly due to an increase of atmospheric pressure over the region. This effect concentrates in winter (− 1.3 mm/yr), while in summer the obtained trends (− 0.2 mm/yr) are not significant. The marked seasonality of the trends can be explained in terms of the seasonal trends of the North Atlantic Oscillation (NAO) and in particular of the Mediterranean Oscillation Index (MOI). The contribution of the mechanical atmospheric forcing to the observed sea level cycle is not very large in magnitude (2 cm amplitude) and is offset from the steric cycle by about 6 months, then reducing the amplitude of the annual cycle when fitting a harmonic function to tide gauge data. Finally, a modal decomposition of the detrended, deseasoned mechanical atmospheric contribution shows a basin-wide leading EOF (66% of variance explained) that implies the existence of a related flow exchange through Gibraltar. The amplitude of the leading EOF is clearly related to the MOI and less related to the NAO.