CARSON CITY, Nev. (AP) — The state Division of Water Resources has decided to proceed with a groundwater pumping test at a planned city being built in the desert north of Las Vegas, even though an independent study suggests it could wipe out a federally protected species of fish.

Martin Mifflin, a hydrologist whose firm prepared the study for the Moapa Band of Paiutes, warned Nevada officials not to proceed with pumping at the Coyote Springs project until they determined why 60 percent of the endangered Moapa dace died in 2007 and 2008. He also said the effect of the tests won't show up for months after pumping begins.

The fish population dropped after the Coyote Springs development began pumping additional water to irrigate a golf course through existing water rights. Residential development at Coyote Springs has been delayed because of the poor economy.

Bob Williams, state supervisor for U.S. Fish and Wildlife Service, questioned the accuracy of Mifflin's study.

He said the dace population has shown a slight increase in the past two years and that hydrologists have been arguing for 10 years whether the pumping will jeopardize the federally protected fish.

He said the dace may have died because of restoration work done at the warm spring pools where the small fish live, about nine miles east of Coyote Springs.

"I am charged more than any other person in this room with protecting the Moapa dace," Williams said during a recent meeting of water officials. "We want information. Let's go forward."

State engineer Jason King said he could order an immediate halt to pumping if studies show the dace population declines.

"That's our hammer," added King, who has said he did not want to be responsible for killing off the endangered fish population.

Coyote Springs, the brainchild of developer Harvey Whittemore, is located on more than 67 square miles of desert land about 50 miles north of Las Vegas. Promoters say it could hold has many as 150,000 new homes.

The Moapa dace is found only in the nearby warm spring pools and streams at the headwaters of the Muddy River, between U.S. 93 and Interstate 15.

The entire natural habitat of the finger-length fish is confined within the 117-acre Moapa Valley National Wildlife Refuge and the neighboring Warm Springs Natural Area, a 2,000-acre tract the Southern Nevada Water Authority acquired in September 2007.

A further drop in the fish population could have wider ramifications. At risk are water rights that the state engineer tentatively awarded the Southern Nevada Water Authority, the Coyote Springs development and the Moapa Indians. More than half of that water is to be pumped into the Las Vegas area by the water authority for use by residents and businesses.

Before they could pump water from Coyote Springs wells permanently, the affected parties were required by a 2002 agreement to carry out a two-year test to pump 8,050 acre-feet of water a year from Coyote Springs wells and determine how that affects the water tables and the fish population.

The test will begin in August or September, said Jeff Johnson, a division manager for the Southern Nevada Water Authority. His agency plans to pump 6,500 acre-feet of water - about 4,000 gallons a minute - from a well at Coyote Springs. Officials say one acre-foot of water is enough to supply two average Las Vegas homes for one year.

The Southern Nevada Water Authority does not need additional water now because the economic downturn has reduced demand in Las Vegas, Johnson said.

Water taken from the well in Coyote Springs over the next two years will be pumped through a more than $21 million pipeline about 16 miles from Coyote Springs to the Muddy River. Some of it will flow downstream to Lake Mead, the Colorado River reservoir behind Hoover Dam.

Johnson said the water can be credited to Nevada for future years when the state needs it.

Nevada is allowed to draw 300,000 acre-feet of water from Lake Mead each year. That federal allocation supplies the Las Vegas area with about 90 percent of its drinking water.

Mifflin said after meeting with water officials that he was particularly disappointed with Williams' statements and his and others' unwillingness to first find a conclusive reason for the fish dying.

"This is the guy who is supposed to protect the dace and he is ignoring our evidence," Mifflin said. "Sure it is possible that the dace population died because of their improvements to the fish habitat. But until they know for sure, it is crazy to go forward."

The number of dace fell from 1,172 to 473 in 2007 and 2008. A count in February showed a population of 532.

Mifflin said his study found decreased flows to the springs that feed the dace habitat would not show up until nine or 10 months after pumping begins. He contended that pumping at full levels could continue for a year and a half before the state engineer stops it under the 2002 agreement.

The Coyote Springs development anticipates it will pump about 1,400 acre-feet of water per year for the next two years under its existing water rights.

Carl Savely, a lawyer with the Coyote Springs development, said the proposed master-planned community does not need additional water now because the poor economy has hampered lot sales.

Bill Van Liew, a hydrologist with the National Park Service, defended Mifflin's conclusions. He said the hydrologist is correct about the pumping damaging the Moapa dace population.

"I have been saying the same thing for the last 10 years," Van Liew.

Ultimately, though, he said, he must defer to Williams' view and go along with the additional pumping.

"We have had this battle for the last decade," Van Liew said. "I think the pumping definitely is going to affect the dace. What happens after that, who knows?"