http://fossil.energy.gov/international/Publications/ucg_1106_llnl...

http://fossil.energy.gov/international/Publications/ucg_1106_llnl_burton.pdf has it as “Only 2 out of over 30 UCG US trials resulted in environmental contamination”.

And our very own consultants to the project have contamination as a myth:
http://ergoexergy.com/eUCG_myt.htm
“Out of 33 UCG trials conducted in the USA in the 1970s and 1980s, only two reported environmental problems: all others were environmentally benign. The Rocky Mountain 1 UCG project (1988) has demonstrated no groundwater contamination. UCG operations in Chinchilla, Australia continued for 30 months and showed no groundwater contamination and no environmental problems whatsoever. Published reports on Soviet UCG plants claim no contamination in the course of, and subsequent to, large scale UCG operations there.”
And goes on to point out the environmental benefits of UCG over digging the stuff up (not that that will hold sway – digging it up’s accepted practice, UCG isn’t so is up against tougher criteria).
http://ergoexergy.com/eUCG_env.htm

And in assessing the a priori risk for a new project, it should be borne in mind that your rule “burn pressure < hydrostatic” has arisen from the Hoe Creek failure. And other site selection critera. Had the significance of these points been appreciated earlier we might be talking 1 or even no failures out of 33 trials.

In the United Kingdom, the department of trade and industry’s view appears to be that “A UCG project site with the appropriate operational controls should present a very low risk to groundwater, but a robust assessment will be required.”, the paragraph going on to stress the need for appropriate hydrological surveys for said assessment.
http://www.dti.gov.uk/files/file19143.pdf

“I would think that Australia would be at the more stringent end of the environmental spectrum.”
As would the US and Europe (no failures therein), though admittedly the Soviets might have been a little more prepared to overlook issues.

“Do you know if the UCG burn location is upstream or downstream from the town’s water supply”
You have exposed my ignorance here, I confess to not having much idea how the town sources its water. But if it comes out of the Gordonbrook dam, that looks like surface drainage rather than subsurface aquifers. However I’d hope we can trust that local water sources were considered way back when the site was originally selected.

Actually you’ve prompted me to do some poking around.
http://www.coordinatorgeneral.qld.gov.au/library/pdf/mp_tarong_ias.pdf
gives a general stratigraphic sequence of the area and blow me, straight above the Triassic coal beds we have a Tertiary volcanic layer, variously comprising basalt, dolerite, shale, siltstone and diatomite.
How impermeable does that sound – no wonder Chinchilla & Kingaroy projects were located in this area.

I’m working myself up to thinking that 3% was a rather pessimistic assessment of this particular risk!