Uranium ore-forming systems of the - Geoscience Australia

Uranium ore-forming systems of the Lake Frome region4. Uranium depositional gradients andchemical architectureSimon van der Wielen, Allison Britt4.1 INTRODUCTIONA new method has been developed for 3D mapping of key attributes of mineral systems relevantto the uranium potential of sedimentary basins. This method has been applied in the Lake Fromeregion to map (a) the 3D geometry of selected stratigraphic units, and (b) variations in theoxidation-reduction state of the rocks (redox). The ‘chemical architecture’ of redox variations isfundamentally important in basin-related uranium systems as it may represent the signature of theflow of oxidised uranium-bearing fluids within the basin (e.g., Cuney and Kyser, 2008), and canpotentially provide exploration vectors towards mineralisation. In terms of the uranium mineralsystem, the chemical architecture constitutes parts of both the depositional gradient and fluidpathway or permeability architecture components of the system, and is important at district todeposit scales (Fig. 1.2).The method was initially applied to the Eromanga Basin (van der Wielen et al., 2009)and then adapted for the Lake Frome region, as described below.4.2 METHOD FOR MAPPING AND MODELLING THE GEOLOGICAL AND CHEMICALARCHITECTUREA large PIRSA dataset, covering the south Australian portion of the Eromanga Basin, comprisesvarious historical drilling data including location, stratigraphy, lithology and some 153,700 drilllog descriptions. Geological descriptions of rock/sediment colour and mineralogy have been usedas proxies for the oxidation-reduction state of the basin sedimentary fill. To this end, wedeveloped a list of words (Table 4.1) that were used to filter the drill log entries. The filter wordswere chosen to discriminate between lithological intervals representing oxidised conditions (e.g.,the description contained words such as ferruginous, limonite, red, or laterite) from thoserepresenting reduced conditions (e.g., lignite, pyrite, green, blue, black, chlorite).Many of the filter words were shortened so as to capture multiple versions of the word. Forexample, filtering for “ferr” captured “ferruginous”, “ferruginization”, “ferruginisation”,“ferruginized”, “ferruginised”, “ferric”, “ferro” and “ferricrete”. Similarly, “iron” captured “ironstained”, “iron staining”, “iron oxide” and “ironstone”.Once the filter word had been run against the data set, any occurrences of that word were flagged.False positive results were then eliminated. For example, the word “limonite” would be flagged bythe program as an oxidised entry, whereas the log entry may read “no more limonite”. Thus, weran a second filter against the flagged occurrences using the words “no, non, none, not, un” andthen manually checked and removed any false positives.The use of abbreviations presented a significant challenge. The following example is an entryfrom Drill Hole 229023, 94 to 105 m:Ple gy-br wkly ox vfg mx slty-chty AB wth com fg bk bi+-mt spots & bi vnlts.Vmnr VQ as abv. Tr-vmnr Ma fract fil-vnlt, 1 spk in ox qz vn fragt. Com mnr gyox-wthd py & tr cp or tarn Py?Thus we filtered for as many “shorthand” variants of the filter words as practical. Hematite, forexample, was commonly shortened to “hm”, “hem”, “he” and “ht”, and chlorite to “chl”, “cl”,“cht” or “ct”.Page 42 of 151