The North and South Flank deposits are located on the flanks of the Weeli Wolli anticline at Mining Area C in the central Hamersley Province. Supergene martite-goethite mineralization is hosted within the Marra Mamba Iron Formation and is developed over a strike length of more than 60 km. This multibillion metric ton resource has been drilled out on a 150- × 50- to 50- × 50-m grid, thus providing us with an unprecedented data set for analysis. This study synthesizes the drill hole data and presents a physical process model that can account for the observed distribution of mineralization. A fluid and mass flux model is proposed which envisages a three-stage process: (1) leaching of Fe from banded iron formation (BIF) in the vadose zone by reduced, acidic, meteoric-derived fluids; (2) penetration of an Fe-rich supergene-fluid plume, driven by gravity and focused by bedding-parallel permeability into the body of ambient alkaline groundwater, effecting nonredox, mimetic replacement of magnetite by hematite and of the gangue minerals (carbonate, silicate, and chert) by goethite coupled with the release of silica into the fluid phase; and (3) a change from silica leaching to silica deposition on the downdip margins of the system before the ore-fluid plume is eventually diluted and becomes indistinguishable from the surrounding body of groundwater. Despite the undoubted secondary role played by structurally enhanced permeability, the primary control on ore-fluid hydrology is gravity-driven flow along bedding planes. This central observation explains every observed feature of the three-dimensional distribution of martite-goethite mineralization, and the inherited structural architecture simply provides the context for this process to play out. This type of control is by no means obvious−the ingress of meteoric fluids during later lateritic weathering of the mineralization does not show this control and produces broadly subhorizontal, bedding-discordant zones of overprinting. The fundamental control exerted on the distribution of martite-goethite mineralization by bedding-plane permeability within BIF horizons suggests that the supergene ore-fluid plume created its own porosity via the relevant ore-forming reactions, and that these were in turn controlled by bedding. A corollary of the pseudomorphic replacement process, both the generation of hematite after magnetite and goethite after gangue phases, is that it typically introduces porosity. The mineralizing process thus creates porosity (and potentially permeability) and is likely to be self-propagating as long as there is continuous supply of ore fluid. This putative active porosity-generation process may be an important clue as to the unique conditions of martite-goethite ore formation. Indeed, it may be that the distribution of magnetite is the critical controlling feature of these ore systems, as the nonredox transformation to hematite not only releases Fe2+ to the fluid phase but concurrently introduces porosity. Further research is required to formulate a comprehensive chemical (as opposed to physical) process model for supergene martite-goethite ore formation. Based on the physical process model presented here, the development of a large-scale martite-goethite mineralizing system requires continued delivery of unleached BIF (and, perhaps ultimately, previously mineralized martite-goethite ore) into the vadose zone. The Hamersley Province has been undergoing significant uplift since at least 60 Ma. Preliminary dating of martite-goethite ores from Mining Area C indicates that they formed at about 45 Ma, at a time when the local climate was temperate and wetter than today. The combination of ongoing uplift and a wet, temperate climate is likely to be the key to the widespread formation of martite-goethite deposits in the Hamersley Province.