THE ROSITA IN SITU URANIUM RECOVERY PROJECT – RESULTS OF GROUNDWATER RESTORATION

In situ uranium recovery (ISR) involves the circulation of naturally occurring and benign ground water, fortified with oxygen, through a uranium ore body were the uranium is oxidized and solubilized, continuing through the sandstone to the extraction wells where the uranium-bearing ground water is pumped to the surface to ion exchange for uranium removal, and then is pumped back into the injection wells to the wellfield re-fortified with oxygen, and again re-circulated through the ore body. Loaded ion exchange resin is then processing into yellowcake, which is dried, and then stored in drums for shipment to a nuclear fuel cycle facility.

The uranium recovery process does not introduce new chemical species to the ground-water system but does elevate certain species that are native to the host aquifer, so after mining is complete, ground-water restoration is conducted. The objective of ground-water restoration is to establish on a parameter-by-parameter basis the primary goal of returning all parameters to baseline conditions. Restoration to precise premining concentration on a parameter-by-parameter basis is not typically feasible. In determining whether the restoration parameters should be amended an ISR operator must address certain criteria to the satisfaction of regulatory agencies, including: the uses for which the ground water was suitable before mining against the potential future use of ground water in the mine area. In Texas, baseline quality has traditionally been established by calculating the mean value on a parameter by parameter basis from an array of baseline wells within a production area. Conversely, Texas rules provide for a statistical analysis of information from baseline groundwater samples proposed by the mining company and approved by the regulatory agency that demonstrates that the groundwater quality is representative of the mean values. ProUCL 4.00.04 (Singh et al., 2009), an EPA-sponsored statistical program is ideally suited for this purpose. ProUCL first determines if the data for a given constituent follows a normal, gamma, or logarithmic distribution and using one of these statistical distributions, a relevant UTL is calculated. If the data follow no discernible distribution a non-parametric UTL is calculated.