LONGTERM CLIMATE VARIABILITY OF THE ADRIATIC SEA THERMOHALINE

LONG-TERM CLIMATE VARIABILITY OF THE ADRIATIC SEA THERMOHALINE PROPERTIES USING AN ENSEMBLE OF REGIONAL OCEAN HINDCAST SIMULATIONS Natalija Dunić1, Ivica Vilibić1, Jadranka Šepić1, Florence Sevault 2, Samuel Somot 2, Robin Waldman 2, Gabriel Jorda 3 1 Institute of Oceanography and Fisheries, Split, Croatia 2 CNRM-GAME (CNRS/Meteo-France), Toulouse, France 3 IMEDEA (CSIC-UIB), Esplorles, Spain

OUTLINE • • • Motivation About THC and DWF of the Adriatic Sea Setup of the hindcast models Preliminary results Conclusions, future work

MOTIVATION Ocean models covering Mediterranean (e. g. www. medcordex. eu) Models have NOT been validated in the Adriatic Sea region their capability for reproduction of the Adriatic thermohaline processes (DWF and Bi. OS)

ADRIATIC SEA • • • Important area of dense water formations NAd. DW • Bi. OS is the dominant generator of the Adriatic variability changes in thermohaline properties influenced by the DWF in the Southern Adriatic • Strength of the DWF varies considerably between years, as do LIW intrusions SHELF CONVE C TION PALAGRUZA DEEP CONVE CT ION OTRANTO

OCEAN MEDITERRANEAN MODELS NEMOMED 8 • spatial resolution: : ~ 10 km; vertical resolution: : 43 vertical z-levels • temporal coverage: : 1961 - 2012 • atmosphere forcing: : ARPERA 50 km • river inputs as point sources CNRM – RCSM 4 • Atmosphere model: : ALADIN • Land surface scheme: : ISBA • Ocean model: : NEMOMED 8 • River routing model: : TRIP NEMOMED 12 • spatial resolution: : ~ 7 km; vertical resolution: : 75 vertical z-levels • temporal coverage: : 1980 - 2012 • atmosphere forcing: : ALDERA 12 km • river inputs as point sources + coastal runoff NEMOMED 12 – Adriatic run • Same as previous with addition of Adriatic 44 river mouth as opposed to 8

EVALUATION OF NEMOMED 8 – FIRST RESULTS - Evaluated on the long-term in-situ data: : Palagruža transect (1961 -2012) - Model overestimates temperature and salinity Model shows only a portion of observed interannual variability

RIVER FORCINGS

RESULTS – INTERANNUAL VARIABILITY OF DENSE WATER FORMATION

INTEGRATED BUOYANCY LOSS Time span - 1 st December – 31 st March Calculated following Marshall and Schott (1999)

DENSE WATER FORMATION RATES VARIABILITY

DENSE WATER FORMATION RATES VARIABILITY Vilibić et al, 2012

CONCLUSIONS AND FUTURE WORK • • Models qualitatively reproduce Adriatic thermohaline variability EMT has a significant impact on the DWF in the Adriatic • Inclusion of new river forcing in the Adriatic does not have a significant influence on dense water formation Coupled system model shows a different behaviour compared to forced simulations • • • Characterizing Adriatic dense water masses DWF influence on the Bi. OS regimes