Simplified geological cross section of the Kloof mining property, west Wits line, showing levels 39, 41 and 43.  The youngest stratagraphic formation is the 2.4 Ga Pretoria group siliclastic sediments and diabase sills (fed by the Running Dyke?).  This overlies the 2.5 Ga Malmani Dolomite which is a principal aquifer in the region and at Libanon is unconfined, but at 4 shaft is confined.  The dolomite overlies the 2.7 Ga Ventersdorp volcanics and the contact is marked by the "Black Reef".   The Ventersdorp volcanics overlie the 2.9 Ga Central Rand Group of the Witwatersrand Supergroup and the contact is marked by the VCR.  The Central Rand contain the MR and the Carbon leader.  All of the deeper levels at 4 shaft mine the VCR, but 39 level also mines the CL of the Central Rand Group.  Shaft 7 of Kloof is just a few kilometers west of the 4 shaft and a sample was collected from the 39 level there, presumably at about the same depth as 39 level in 4 shaft.  The strata are partitioned into "compartments" by dykes of varying ages.  The 1.4 Ga syenite dikes radiating from the Pilansberg Complex (within the Bushveld Complex) to the north cut all formations (Venterspost and Kloof dykes). Water from the dolomite enters into the underlying Witwatersrand beds primarily through fissures arranged en echelon to these dike aquicludes.  Running dyke must be ca. Transvaal in age or perhaps 2.0 Ga?  The dykes associated with Danies Fault and perhaps the Glenharvie dykes are Ventersdrop in age as they are all truncated at the contact with the overlying dolomites.  Faults are often associated with dykes and vice-versa dykes will follow faults (any weakness in the crust).  Finally, the Witpoortjie Fault truncates the Wits strata and may be a 2.0 Ga fault.  If so it should come to surface ca. 10 km south of Glenharvie.  If this fault separates basement gneiss from the Wits strata, then it may act as a conduit for deeper fluids that may feed the subvertical fractures associated with the dykes.
 
 

The very first set of samples collected from Kloof were taken from 41 level 56 hanging wall drive south (K441 FW1 and 1a) and from 41 level 53 hanging wall drive north (K441FW2).  More recently, these sites have been resampled.  Danies fault zone was also sampled recently at 43 level.
 

    Samples KL441FW56HWDS come from fractures adjacent to the Glenharive Dyke.  When first encountered while cover drilling during tunnel advancement several months prior to our first sampling, the flow rates were 100,000 liters/hour.  It took a week just to put a valve on this one borehole.  Several other boreholes were drilled through cased valves and more water encountered.  A tracer test was performed using a fluorescent dye to test the connectivity of different boreholes.  The fractures believed to be responsible for the water are illustrated as dashed blue lines beyond the tunnel face.  Approximately 7 tons of chemically treated concrete were injected into these holes prior to our first sampling (blue boreholes) in '98.  Subsequent to sampling the head of the tunnel was advance another 3 meters.  No further advancement is planned.  Ventilation is generally shut off now.  Subsequent samples have come from other boreholes (red).

References:

van Heerden, A.W.C., A Lithological and Geomorphological Study of the VCR Sand Facies observed within the 4-Shaft Complex, Kloof Gold Mine.
    Unpublished Report.