An extreme “millennium” drought from 2007 to 2010 resulted in the lowest River Murray levels (1.75 m decline from average) in over 90 years of records at the end of the river system in South Australia. This resulted in low inflows of water to Lake Alexandrina and Lake Albert and exposed large areas of soils on the lake margins to drying. This drying caused the sediments to shrink and crack with the formation of large columnar blocks of soil. Sediment physical properties were measured on samples taken from the lake at different locations (Cook et al., 2011) and used to develop a HYDRUS2D/3D model to determine the oxygen penetration into cylindrical peds (soil blocks) with varying radii (0.05, 0.1, 0.15 m), depth of cracks (0.1, 0.2, 0.5 m) and water table depth (WTD) (0.1, 0.2, 0.5, 1.0 m). The peds were assumed to be initially saturated and lost water due to drainage (to the water table) and evaporation (potential 4 mm day-1), from both the top and sides of the ped. Oxygen (O2) penetrated the peds due to gas advection (as water was lost), and diffusion, and was lost due to oxygen consumption by organic matter (98% of the sink strength) and pyrite (2%). This drying and oxygenation process was modelled for 1000 days and the O2 concentrations recorded. O2 penetrated to the center of the ped at the sediment surface, but the radial penetration decreased with depth (Figure 1) and the concentration varied little after the first day due to near balance between diffusion and consumption. Pyrite oxidation and hence the formation of acidity in the peds was modelled using a first-order chain reaction based on oxygen concentration. This resulting in a slow increase in Fe2+ concentration with time (Figure 2). Various scenarios were tested to describe rewetting and hence transport of acidity to the lake; a rising water table with seepage from the crack and surface, inundation of the surface and crack, and water flow into the ped from the crack and seepage at the surface. These were compared to simulations of the sediment without cracks as a control, with two rewetting scenarios: rewetting due to watertable rise, and surface inundation. The mass of acidity generated from oxidation of pyrite was most sensitive to increasing ped radius, and then linearly with time. Acid generation within the peds was greater than in sediment without cracks. The scenario where the water table rose from below resulted in the most acidity transported from the soil to the lake water, while the scenario where both the surface of the ped and crack were instantaneous inundated resulted in the least acidity transported to the lake water. These results have implications for how lake rewetting is allowed to occur.