The UNSW study has prompted the Department of Primary Industries Water (DPIWater) to commission a review of the work, a move that has been seized upon by farmers campaigning against the much larger Shenhua Watermark coal mine proposed for a site about 20 kilometres away.

Doug Anderson, a principal groundwater engineer with the Water Research Laboratory of the University of NSW, said there were multiple problems with the groundwater modelling conducted by Whitehaven and approved by the Department of Planning and Environment. DPIWater provided advice to Planning on water management at the mine in northern NSW.

These included the drilling of just one deepwater bore – more than two years after an extension of the mine was approved – and locating it to the north end of the mine rather than a more logical site to the south, close to the main farmers' bores and between the main pit and the nearby Quipolly Creek.

Instead of focusing on the main water-bearing zone at a depth of 142 metres, it is collecting both shallow and deep groundwater between 5 metres and 168 metres. "You actually don't know what that well's monitoring," Mr Anderson said.

Whitehaven, which has rejected claims for compensation from nearby farmers for their water loss, dismissed Mr Anderson's study. Government agencies as recently as July had advised the company "they were satisfied that the water management on site was in compliance with all relevant regulations", a spokeswoman said.

"Whitehaven continues to maintain, on the basis of its own long-term monitoring, that the onset of drought conditions in 2012, and the fact that rainfall has not kept pace with groundwater extraction for irrigation by surrounding land users since then, are the primary factors implicated in declining groundwater levels in the area."

Mr Anderson estimated the mine was responsible for at least 25 per cent of the water loss.

A spokesman for DPI Water said it had done regular reviews of water management at the Werris Creek site but had decided to commission an independent report on the UNSW study.

"This is in line with normal procedures and ensures any new developments in science or best-practice water management are considered as part of the ongoing management of the site," he said.

Susie Lyle, chairwoman of the Caroona Coal Action Group which had paid for the UNSW work, said groundwater that had been flowing north to south had now reversed because of the mine – something overlooked by regulators.

"I really don't think they've done their job," Ms Lyle said. "If Whitehaven has got a problem like that and blames the drought, why is the water squirting out into the mine?"

Whitehaven had predicted water levels for nearby farmers would drop 10 centimetres, instead they have fallen by 4.5 meters to 15 metres, a sign groundwater interference could also pose a risk to the rich Liverpool Plains farmland where Shenhua plans to dig, she said.

"It's taken us, with our own resources, to say, 'hang on, there's something wrong here'," she said.

Greens mining spokesman Jeremy Buckingham said: "The risk to surface and groundwater, agricultural productivity and climate are far too great to allow these mines to go ahead.

"All the water modelling in the world does not compare to a real-life experience, and here we have a disaster only a few kilometres away from the Shenhua Watermark mine site," he said.

Mr Anderson cautioned against drawing parallels between Werris Creek and Watermark, saying the geology of the two sites is different.

Still, "there a lot of lessons that we're seeing from what's happening at Whitehaven". A detailed analysis of the groundwater declines at Quipolly Creek, for instance, might reveal approaches that could be applied to improving groundwater assessment or policy practices elsewhere in NSW, he said.

Hundreds of farmers and other groups opposed to the Shenhua mine are expected to gather at a farm near the proposed mine on Saturday.