Model input used for 10.5194/hess-2019-151 Abstract The Nile Delta is an important agricultural area with a fast-growing population. Though traditionally irrigated with surface water, the delta increasingly relies on groundwater. However, saline groundwater extends far land inward, rendering groundwater close to the coastal zone useless for consumption or agriculture. To aid groundwater management decisions, hydrogeologists reconstructed this saline and brackish groundwater zone using variable-density groundwater models with very large dispersivities. However, this approach cannot explain the observed freshening of this zone as observed by hydrogeochemists, who hypothesize that the coastal saline zone is the effect of the Holocene transgression. Here, we investigated physical plausibility of this hypothesis by conducting a palaeo-reconstruction of groundwater salinity for the last 32 ka with a complex 3D variable-density groundwater flow model, using state-of-the-art model code that allows for parallel computation. Several scenarios with different lithologies and hypersaline groundwater provenances were simulated, of which five were selected that showed the best match with the observations. Amongst these selections, total fresh water volumes varied strongly, ranging from 1526 to 2659 km3, mainly due to uncertainties in the lithology offshore and at larger depths. This range is smaller (1511-1989 km3) when we consider the volumes of onshore fresh groundwater within 300 m depth. Regardless of the variance, in all cases the total volume of hypersaline groundwater exceeded that of sea water. We also show that during the last 32 ka, the total fresh groundwater volumes significantly declined, with a factor ranging from 1.9 to 5.4, due to the rising sea-level. Compared to a steady-state solution with present-day boundary conditions, the palaeo-reconstruction improved our validation for the saline zone (5 g/L – 35 g/L TDS). Also, under highly permeable conditions the marine transgression simulated with the palaeo-reconstruction led to a steeper fresh-salt interface compared to its steady-state equivalent, while low permeable clay layers allowed for the preservation of volumes of fresh groundwater. This shows that long-term transient simulations are needed when estimating present-day fresh-salt groundwater distribution in large deltas. The insights of this study are also applicable to other major deltaic areas, given the wide-range of lithological model scenarios used in this study and since many deltas also experienced a Holocene marine transgression.