Dry events in the winter in Israel and its linkage to synoptic and large-scale circulations

Dry spells during the rainy season have a far-reaching environmental impact in the Levant. A previous study by Saaroni et al. investigated prolonging dry spells (PDSs) and classified them subjectively into three types: “subtropical,” “baroclinic” and “polar.” They developed quantitative indices to identify each type. The present study analyses dry events, both individual days and spells, based on the above classification. We found that the indices identifying PDS types are effective for distinguishing among individual dry days of the three types, but not as precursors of dry days. However, when the indices keep positive for 2 or 3 days consecutively, they become effective precursors of dry spells. An alternative, automatic classification, using the K-means technique, yielded classes similar to those defined subjectively. The analysis revealed the dominance of the “baroclinic” type and the tendency of “subtropical” type to become “baroclinic.” Composite maps of cyclone tracks which were derived for days belonging to each type showed track distributions characteristic to particular types. For every type, the evolution of dry spells was studied through composite maps of 500-hPa geopotential height (GPH) anomaly. The maps were derived for 1–3 days preceding the event. Pronounced patterns were identified as precursors of specific dry spell types; positive anomaly over Canada and Greenland, combined with negative anomaly over northwest Europe turned out as precursors of “baroclinic” and “subtropical” events, whereas the North Atlantic Oscillation (NAO), in its positive phase, was found as an exclusive precursor of the “polar” type. All three types of dry days are characterized by an offshore lower-level wind, and by an absence of Cyprus Lows, due to deflection of cyclone tracks away from the Middle East. This implies that dry conditions in the Levant are caused by both thermodynamic (mesoscale) and dynamic (synoptic scale) factors.