Engineers and construction crews battled for months to patch up leaking underground rock strata below the Quail Creek Dike - an effort that indicates the project had severe geological problems at least two years before the dike burst New Year's Eve.

The work was outlined in an interview with Richard B. Hall, supervising engineer for the Dam Safety Section, Utah Division of Water Rights, and in a sheaf of memoranda and letters in the division's files.The Dam Safety Section is responsible for several hundred non-federal dams. Almost all are earthfill, like the Quail Creek Dike and Dam.

The dike, located about nine miles from St. George, broke on New Year's Eve. The resulting flood damaged homes, apartment complexes and farm land. Four bridges were washed out, a natural gas line was ripped from its trench and livestock was killed.

However, there were no injuries. An independent commission is examining the cause of the failure and state officials hope to get its report around the end of February.

The reservoir began filling in 1985.

Water seepage through the dike was so serious in the years before the disaster, that hundreds of grout holes were drilled through the dike and deep into the bedrock under it, Hall said.

But from time to time, the seepage would stop and then start again. Apparently water was moving through fractures in the underground bedrock, which were perpendicular to the dike.

Two main efforts were undertaken to end the seepage. Crews grouted _ or pumped concrete and other materials into holes drilled in the dike _ and a new cutoff trench was dug to block the water flow.

Hall said teams worked their way along the entire crest of the dike, drilling holes. The holes often extended 100 feet or farther below the base of the dike.

The holes were drilled closer together until there was no more "take" _ that is, until the holes were so filled with pumped-in material that they could take no more. Eventually, holes were only about three feet apart, Hall said.

"It was a major effort," costing about $1.5 million.

"In 1987 they had a concentrated leak. The thinking was that the fractures were at depth because they were breaking out downstream . . .

"Then they had a similar one in 1988, where they went in and grouted again." One hole was so saturated with water that it took 1,500 one-pound bags of grouting material to fill the hole. "But a lot of it was getting washed out," he said. "Most of it just went to kind of concrete the downstream bedrock."

He said the corrections seemed to be working. "Plus it was very carefully monitored," Hall said.

The other effort required digging a new cutoff trench in the reservoir to reinforce one built when the dike was constructed.

The first trench was 10 feet deep, dug into the ground before the dike was heaped over it. A cross section of the dike looks like a dull arrowhead, with the fill material in the trench being the base of the arrowhead and the point being the crest.

The dike's core was sandy silt, except for a one-foot layer of clay at the bottom of the trench. The core was lined on the upstream side with a layer of clay, although this vertical layer did not extend into the trench.

"But this sandy silt is impermeable by itself," Hall said.

In 1987, Hall said, because of the seepage, a second cutoff trench was dug upstream from the dike.

The new trench was filled with clay in an attempt to block the flow of water that seemed to be moving below ground along the fractures in the bedrock.

A letter from State Engineer Robert L. Morgan to the Washington County Water Conservancy District, the reservoir's owners, outlines problems with the dike after a field inspection on April 23, 1987.

"The dike has been grouted along its entire length with the highest density of grouting concentrated in areas of higher take, and a special trenching machine extended the new cutoff as deep as possible," the letter begins.

It noted that a "special trenching machine extended the new cutoff as deep as possible, although there was still evidence of cracks extending beyond the trench excavation.

"Many fractures or joints, which were oriented approximately perpendicular to the dike's axis, were discovered in the trench excavation. The trench was backfilled with clay. The eroded, exposed bedding planes in the basin near the dike were scarified and a clay blanket was placed over these areas."

The blanket was a layer of clay that was designed to seal off the bedrock, preventing water from flowing to the fractures below.

However, the letter continues, "The blanket was not protected and has shown some signs of erosion."

"The remedial measures taken have enhanced the safety of the dam," the letter said. "However, the fact remains that there is still soluble gypsum in the abutment and foundation, as well as cracks or joints in the basin below the foundation cutoff and grout curtain."

Crews continued to plug the leaks.

On Oct. 3, 1988, the division's Matt Lindon and John Mann undertook an informal inspection of the dike.

"They were somewhat alarmed to see a concentrated seepage area with sand boils immediately downstream of the location of the original grout plant," said a memo to the file from Hall, dated Oct. 5, 1988.

The file's last note before the dike burst is dated Oct. 13, 1988.

The note indicates an enginer for one of the firms that worked on the project said he finally got the latest leak sealed, but still did not feel the problem was solved.