Eelco J. Rohling
Southampton Oceanography Centre,
University of Southampton
United Kingdom

Extended abstract
Being a semi-enclosed marginal basin of relatively small volume (compared with the open ocean), the Mediterranean shows amplified and very rapid response to climate change. Consequently, climate signals are well expressed in the Mediterranean, and in the case of the glacial-interglacial contrast in stable oxygen isotope records, for example, the Mediterranean shows a signal amplitude that is twice that observed in the open ocean. The Mediterranean also shows generally elevated sedimentation rates, and reduced sediment mixing by bioturbation, relative to the open ocean, especially during the special times in the past when the bottom waters became anoxic. These attributes make the Mediterranean an excellent basin for the investigation of changes in past climate and hydrographic response, at high (decadal- to centennial-scale) temporal resolutions.

There is particular interest in the history of Mediterranean oceanographic developments because the basin occasionally shows organic-rich sediments (“sapropels”) that lack any sign of benthic life: these were deposited during periods of several thousands of years of (almost) completely anoxic bottom-water conditions. This offers a stark contrast with the modern well-oxygenated bottom-water conditions throughout the basin. We now know that such conditions recurred at times when the Northern Hemisphere insolation reached distinct maxima, controlled by periodic variations in earth’s position and orbital characteristics relative to the sun. Favourable conditions for sapropel deposition tend to develop roughly every 21,000 years. The most recent sapropel was deposited between nine and six thousand years ago. The insolation maxima were especially effective in enhancing the summer monsoons on the northern hemisphere, and the African monsoon is known to have been strongly intensified at times of sapropel depositions, discharging into the eastern Mediterranean via the Nile and likely wadi-type discharge systems along the wider N African margin. The northern borderlands of the Mediterranean also experienced much more summer humidity than today. This cannot be address in terms of the monsoons, and likely reflects precipitation originating from evaporation over the Mediterranean basin itself. Sapropel formation was most distinct in the eastern Mediterranean, and the following concentrates on that region.

Before evaluating the relationship between enhanced humidity/runoff and a change in deep-water oxygenation, a schematic view is needed of the modern deep-water ventilation in the basin (Fig. 1). Inflowing surface water is subject to high evaporation (and warming) through its pathway. Winter cooling of the high-salinity surface water in the Rhodes-Cyprus area causes it to sink to form intermediate water. The net forcing behind intermediate water formation, the “first step/stage” of deep ventilation, is predominantly the salinity gain due to net evaporation. Intermediate water spreads throughout the basin at ~150/ 200 to 600 m depth, and in the southern Adriatic Sea this high-salinity but relatively warm water mass mixes with waters that are of lower salinity, but cooler, and which originate from strong winter cooling in the N Adriatic. The mixing endproduct is a dense, relatively high salinity, relatively cool, water mass that spreads below the intermediate waters to the greatest depths of the eastern Mediterranean. This “second stage” in the deep ventilation process is predominantly related to cooling. Without the salt supplied by the intermediate water, the cooling would not suffice to create sufficiently dense water to ventilate the basin down to the bottom. Note that this is a very simplified portrayal of the deep ventilation, which in reality is driven from variety of regions and by subtle temperature and evaporation shifts, but it offers a useful concept with which to approach the dramatic changes that occurred in the past.

At times of sapropel formation, the strong humidity/runoff increase affecting the basin caused a serious reduction in the net evaporation that is so critical in the first stage of deep ventilation. Conceptual reconstructions, supported by Ocean General Circulation Models suggest that the salty intermediate water would consequently have collapsed (Fig. 2). Without its salt-supply, the deep ventilation from the Adriatic Sea could penetrate only to shallow intermediate depths, reaching about 400 m. Below that level, there was very limited or no ventilation and ongoing oxygen consumption rendered the stagnant “old” deep water virtually anoxic within a matter of centuries. Organic matter that rapidly sank to the sea floor was no longer subject to oxidation in this old deep water mass, and it consequently became preserved and buried in the sediments – a sapropel was being deposited. Here, it needs to be mentioned that productivity during these events was also enhanced relative to the present, so the organic flux was increased, which augmented its concentrations in the sediments.

The above reflects the traditional view of the conditions at times of sapropel deposition, but there have always been nagging doubts about the intensity and extent of truly anoxic conditions. Having been deposited under (virtually) anoxic conditions, which precluded benthic life that might otherwise bioturbate the sediments, many sapropels display the original (seasonal?) sedimentary lamination. These sediments can therefore be sampled in great detail, to look at variability on decadal time scales. In a recent study, we have compiled evidence from a variety of sapropels of different ages, which shows that the anoxic conditions may have been more intermittent than previously thought, not only in the northern basins (Aegean and Adriatic), but also in the main body of the eastern Mediterranean. This would suggest that there were bursts of ventilation to greater depths (as it seems in some cases even down to 2000 m), which were sufficient to allow flourishings of benthic organisms that are not low-oxygen adapted. Since a truly anoxic water column would be full of chemical elements in a reduced state, all oxygen from brief burst of ventilation into an anoxic water body would be quickly “titrated” away, chemically. Consequently, there would be no bio-available oxygen at depth. The implication of this is that the observed benthic faunas imply that the bulk of the water column (at least down to 2000 m) was never completely anoxic, but that true anoxia (as reflected in the sediments) was restricted to a thin layer of water at/above the sediment surface. Any burst of ventilation could temporarily reoxygenate that, leaving sufficient bio-available oxygen to support the observed faunas (Fig. 3).

What were the “bursts” of deep ventilation at times of sapropel deposition related to? Today, much of the cooling that drives deep ventilation is achieved by intermittent northerly outbursts of cold and dry polar/continental air masses, which are orographically channelled over the northern sectors of the eastern Mediterranean. Given the latitudinal position of the Mediterranean at the boundary of subtropical and temperate westerly influences, and given that this position has not changed over the time scales considered here, we expect these types of outbursts also to have occurred at times of sapropel formation. There may be some evidence that the frequency/ intensity of such events varies on decadal time scales (NAO?), and we consider that the evidence found in the sapropels implies that such decadal-scale variability was a feature in the past as well. Interestingly, many sapropels show a first period of several centuries to a millennium of very stable/tranquil conditions, followed by (a) period(s) of several centuries with very frequent reventilations, and a final episode of several centuries with stable/tranquil conditions. Is there perhaps a longer-term, centennial- to millennial-scale organisation in the frequency/intensity of the cold air outbursts (and, therefore, possibly in the NAO?)? This as yet remains speculative, but ongoing research should bring some useful clarification.

References
Casford, J.S.L., Rohling, E.J., Abu-Zied, R.H.., Jorissen, F.J., Leng, M., and Thomson, J. A dynamic concept for eastern Mediterranean circulation and oxygenation during sapropel formation, Palaeogeography, Palaeoclimatology, Palaeoecology, 190, 103-119, 2003.

Rohling, E.J., The dark secret of the Mediterranean - a case history in past environmental reconstruction, http://www.soes.soton.ac.uk/staff/ejr/DarkMed/dark-title.html (November 2001).

Rohling, E.J., De Rijk, S., Myers, P.G., and Haines, K., Palaeoceanography and numerical modelling: The Mediterranean Sea at times of sapropel formation, Geol. Soc. London Spec. Publ., 181, 135-149, 2000.

Rohling, E.J., Cane, T.R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K.C., Schiebel, R., Kroon, D., Jorissen, F.J., Lorre, A., and Kemp, A.E.S. African monsoon variability during the previous interglacial maximum, Earth Planet. Sci. Lett., 202, 61-75, 2002.