HILO, Hawai'i — Using new techniques, scientists plan to more accurately map the path of lava from Kilauea in an effort to anticipate a curious eruption that isn't quite like any the experts have seen.

The lava flow emerging from a crack near Kupaianaha built itself a huge channel raised above the surrounding terrain, a kind of mile-long elevated lava flume system 120 feet above ground level when the eruption began.

Scientists say this kind of perched lava channel has never been seen before, and experts can't predict what will happen next.

The stakes are high.

The fissure is 11 miles from Pahoa Village, and has been pumping out an average of 500,000 cubic meters per day since the July 21 fissure eruption began east of Pu'u 'O'o. The channel has been directing that lava toward the northeast, in the general direction of Pahoa and rural subdivisions that surround the town.

The fissure and channel system have extended flow fingers as far as 3 1/2 miles to the northeast, but each time the flow has stalled well short of any populated areas. Pahoa Village is 700 feet above sea level, and the stop-and-go flow activity has been centered at the 1,800- to 2,200-foot elevation level, and eight to nine miles from the town.

Last week, the flows advanced downhill to just below the 1,700-foot level in the Wao Kele O Puna rain forest, but it appeared possible the flow would stall again because the lava tube feeding it seemed to be blocked.

This Kilauea eruption began more than 24 years ago, and most of its lava flows have moved southeast down the south face of the rift zone often continuing toward the ocean.

The lava has covered about 29,000 acres and burned 189 homes and other buildings including Kalapana Village, but the vast majority of the flows have moved through unpopulated areas. No buildings have burned since 2002.

NEW TOOLS AVAILABLE

The current eruption is different because it has been channeled generally to the north and northeast. Residents and civil defense officials have been monitoring where the lava would have to cross to reach Pahoa and Highway 130. The goal is to predict the lava's likely path if the flow becomes more focused and begins moving in a single direction.

One tool scientists have now that was not available when Kalapana was destroyed in 1990 is geographic information software that allows scientists to develop digital elevation models to calculate likely paths that lava may follow.

Maps generated from the data show how fluids such as water would drain across the topography, and two tests by Hawai'i scientists suggest that data is a reasonably good predictor of how lava flows over the same terrain.

Jim Kauahikaua, scientist-in-charge at the U.S. Geological Survey's Hawaiian Volcano Observatory, said the first test used the data to predict how lava would flow from a Mauna Loa eruption, and then compared that with how 'a'a lava had actually moved toward Hilo on Mauna Loa's Northeast Rift Zone during the 1984 eruption.

The result was encouraging, with the flows following the predicted paths more than 60 percent of the time, Kauahikaua said.

A second test was tried at Kilauea volcano during what was known as the "Mother's Day" flow in 2002 down the west side of the Kilauea flow field.

The U.S. Park Service wanted to create fire breaks to protect critical native forest from fires started by the lava, and wanted the best information available on where the lava would likely go to put the fire breaks in the right places.

USGS used the data to compute flow paths in that area "and again, it was a very good indicator of where the flows were going," Kauahikaua said. "In general, these flows as they crossed the pali especially, they were right on the money in that same direction, and our computed flow paths were all pretty close together in there."

LAVA'S NOT LIKE WATER

Predicting the likely flow paths in the Puna area between Kupaianaha and Pahoa is more difficult because the slopes there tend to be gradual, making it less obvious which way lava will move. USGS now hopes to use much more detailed Light Detection and Ranging, or LiDAR, data collected on the Wao Kele O Puna rain forest to recalculate likely flow paths for the area.

Another problem is that lava doesn't move quite like water. A lava flow may be moving downhill and abruptly stop and harden because of an interruption at the magma source. The cooled lava may then present an obstacle to future flows, causing the next flow to veer off in another direction.

Perhaps the most striking feature of the fissure eruption is the impressive elevated channel that was created by the flow, and has now become a flumelike system for transporting lava northeast from the fissure near Kupaianaha toward Pahoa.

The channel was created when 'a'a cooled and slowed at the end of a lava steam, creating a blockage that caused the stream to back up. The steam then overflowed onto its banks in layers that quickly cooled, and the cooling lava on the banks built up the sides of the channel.

This process has been repeated over and over. Each time the flow out of the mile-long channel is obstructed, molten lava backs up, spills over the sides and hardens, scientists said. Each overflow builds the sides of the structure higher.

TO DIVERT OR NOT?

The channel is now perched about 40 yards high above the original level of the original terrain, and is still routing lava to the northeast.

Several times since August, the sides of the channel have broken and allowed lava to spill out, but the channel always repairs itself, scientists said. Hardened chunks of cooled lava that drop back into the lava stream are carried along until they pile up at the breaches in the channel and plug them up.

However, since the channel is above the surrounding terrain and is directing lava toward a population center, some have wondered if it might be possible to deliberately breach the channel and divert the flow if it threatens the village or highway.

Kauahikaua said that what it would take to breach the channel, and where the ideal place would be to make the breach to steer lava away from populated areas, are questions for civil defense officials and engineers.

Civil Defense Administrator Troy Kindred sees several problems, and said the possibility of diverting the lava is not being actively pursued. Among other things, consideration would need to be given to the cultural implications for Hawaiians, he said.

"There's a lot more things that influence this decision than what might just appear on the surface," he said. "There's a cultural context, and there's a technical context, and whether or not it will work."

Another issue is if the authorities were successful in diverting the flow and then it damaged homes, the authorities bear responsibility for the damage.

Big Island Managing Director Dixie Kaetsu agreed that "diversion is not something that is actively being considered as feasible."

Reach Kevin Dayton at kdayton@honoluluadvertiser.com.