RECONSTRUCTING HOLOCENE RAINFALL VARIABILITY OVER THE DEAD SEA WATERSHED AND THE SOUTHERN LEVANT USING A STOCHASTIC MODELING APPROACH

The Dead Sea (DS) is the terminus of a ~43000 km² watershed in the Levant and as such records its paleohydrology and paleoclimate. In this research we applied a simple water balance model (WBM) to determine natural annual DS level changes under present-day climate and to deduce possible precipitation scenarios associated with Holocene lake level rises and drops. The current climate ranges from sub-humid Mediterranean to hyperarid and arid in the north and south parts of the watershed, respectively. The DS can be conceptually described as a simple water body receiving precipitation over the watershed and loosing water solely via evaporation from the lake surface.

The late Holocene DS level fluctuated between -390 m to -415 m. Nowadays the DS water balance is disturbed due to human interference; therefore, the WBM we developed and calibrated is based on pre-interference (1964) rain and level data. Pre-interference the Jordan River, draining the northern part of the watershed, discharged >80% of the water inflow to the DS. To estimate past Holocene rainfall regime that controlled DS levels, synthetic rain series were stochastically generated by (a) gradually shifting the annual mean and (b) adding different precipitation trends, thereby representing possible scenarios of precipitation patterns. Each scenario included 1000 series of 500 years long. These ensembles were input to the WBM and probabilities of DS level rise and drop rates for different durations were estimated. Specifically, the rises and drops of the DS during the Holocene, as reconstructed from shoreline and sedimentary records, were examined for their probability. Most of the late Holocene level fluctuations can be explained by the ±10mm/10yrs trend scenario. As expected, none of the scenarios examined can explain the modern artificial level drop, which is due to human intervention in the DS water balance. The methodology presented is based on simplified assumptions; however, it allows quantitative assessments of possible rainfall regimes over the DS watershed during the late Holocene.