The Chilia arm, which flows along the northern rim of Danube delta (Fig. 1), has successively built three lobes (Antipa, 1910) and it was first mapped in detail at the end of the 18th century (Fig. 2a). The depositional architecture of these lobes

was controlled by the entrenched drainage pattern formed during the last lowstand in the Black Sea, by the pre-Holocene loess relief developed within and adjacent to this lowstand drainage and by the development of Danube’s own deltaic deposits that are older than Chilia’s (Ghenea and Mihailescu, 1991, Giosan et al., 2006, Giosan et al., 2009 and Carozza et al., 2012a). The oldest Chilia lobe (Fig. 2b and c) filled the Pardina basin, which, at the time, was a shallow Bleomycin lake located at the confluence of two pre-Holocene valleys (i.e., Catlabug and Chitai) incised by minor Danube tributaries. This basin was probably bounded on all sides by loess deposits including toward the

south, where the Stipoc lacustrine strandplain overlies a submerged loess platform (Ghenea and Mihailescu, 1991). Because Galunisertib cost most of the Chilia I lobe was drained for agriculture in the 20th century, we reconstructed the original channel network (Fig. 2b) using historic topographic maps (CSADGGA, 1965) and supporting information from short and drill cores described in the region (Popp, 1961 and Liteanu and Pricajan, 1963). The original morphology of Chilia I was similar to shallow lacustrine deltas developing in other deltaic lakes (Tye and Coleman, 1989) with multiple anastomosing secondary distributaries (Fig. 2b). Bounded by well-developed natural levee deposits, the main course of the Chilia arm is centrally located within the lobe running WSW to ENE. Secondary channels bifurcate all along this course rather than preferentially at its upstream apex. This channel network pattern suggests that the Chilia I expanded rapidly as a river dominated lobe into the deepest part of the paleo-Pardina lake. Only

marginal deltaic expansion occurred northward into the remnant Catlabug and Chitai lakes and flow leakage toward the adjacent southeastern Matita-Merhei ID-8 basin appears to have been minor. Secondary channels were preferentially developed toward the south of main course into the shallower parts of this paleo-lake (Ghenea and Mihailescu, 1991). As attested by the numerous unfilled ponds (Fig. 2b), the discharge of these secondary channels must have been small. All in all, this peculiar channel pattern suggests that the Chilia loess gap located between the Bugeac Plateau and the Chilia Promontory (Fig. 2b) already existed before Chilia I lobe started to develop. A closed Chilia gap would have instead redirected the lobe expansion northward into Catlabug and Chitai lakes and/or south into the Matita-Merhei basin. The growth chronology for the Chilia I lobe has been unknown so far. Our new 6.