Tool may help focus well-pad reclamation efforts
New federal research has reconfirmed the challenges of revegetating oil and gas pads on the high and arid Colorado Plateau, while leading to development of a tool for better assessing regional recovery trends in order to help guide reclamation efforts in the future.
The sweeping analysis of more than 1,800 wells by U.S. Geological Survey researchers found that most pads where wells had been plugged at least nine years have more bare ground and less vegetation than surrounding undisturbed areas. The majority of pads had a 15 percent to 45 percent increase in bare ground when compared to nearby areas.
More bare ground can lead to more soil erosion and dust emissions, less forage for wildlife and livestock, and increased chances of infestation by weeds.
Of note, however, is that pads in northwest Colorado’s sagebrush-dominated Piceance Basin are generally recovering better than elsewhere in the region.
“One of the surprising things is the sagebrush shrublands seem to be doing better in terms of having more vegetation cover, less bare ground, when compared to the other vegetation types,” said researcher Michael Duniway.
But he noted that the satellite images researchers relied on to evaluate vegetation cover are too coarse to determine what kind of revegetation is occurring — whether it’s sagebrush or something else.
Duniway is one of the co-authors of the study, published in the journal Science of the Total Environment. He works out of a USGS office in Moab, Utah, along with fellow soil scientist Travis Nauman, the study’s lead author.
Duniway said the study’s original goal was to try to identify landscape reclamation patterns based on factors such as climate and the landscape itself, to help guide future development and reclamation success. The researchers were motivated by the tens of thousands of wells that already have been drilled in the region, the amount of oil and gas reserves the USGS has estimated remain in places like the Piceance Basin, and the fact that it’s such a challenging environment for reclaiming and restoring landscapes.
“It’s very dry. Soils can often be thin or failing, generally making it hard to have success in restoration/reclamation activities,” Duniway said.
He said researchers weren’t overly surprised by the amount of bare soil identified by the study, based on prior research pointing to the difficulty of restoring lands on the plateau.
A new tool
The study focused on pads where wells had been plugged and the pads deemed abandoned between 1997 and 2005. It included 481 sites in Utah, 399 in Colorado and 986 in New Mexico, with Arizona being excluded due to issues including low site numbers, Duniway said. He said probably more than half of the Colorado sites studied are in northwest Colorado’s Piceance Basin.
It was in the process of trying to identify reclamation patterns that the researchers focused on technological tools that can help with the job. In doing so they developed a new approach to do the job at a broad scale, something potentially useful around the world. They call it DART, which stands for disturbance automated reference tool. It makes use of satellite imagery, digital soil mapping, predictive ecological modeling and field assessments to evaluate vegetation recovery.
Duniway described digital soil mapping, or predictive soil mapping, as a statistically based approach for predicting soils across a landscape. Researchers were able to combine such mapping with a new cloud computing platform along with machine learning, which involves analyzing algorithms with high-powered computers, to study reclamation at a broader scale than otherwise would have been possible, he said.
Researchers noted that well pads aren’t recovering as well in grasslands, canyon regions, blackbrush shrublands and shale badlands as in some other ecotypes. While poor recovery in an area such as badlands, with their poor soil, wasn’t unexpected, the grasslands result was unexpected because grasslands in general recolonize faster than woody species after severe disturbance, the study noted.
Among other trends, it also found that warmer areas with summer-dominated precipitation also haven’t recovered as well. Such areas are found more on the southern part of the plateau. Duniway said reclamation just could be harder more generally there, but perhaps some new reclamation methods might have more success there given the conditions.
He said the higher revegetation levels seen in sagebrush shrubland may reflect increased precipitation and more fertile, finer soils that lend themselves better to restoration.
However, Richard Alward of Grand Junction, who does consulting work for the oil and gas industry, echoes Duniway’s caution against jumping to conclusions from the study about the type of revegetation occurring on pads in sagebrush country or elsewhere. Alward and his wife and fellow plant ecologist Tamera Minnick, a Colorado Mesa University professor, have done highly focused research on a small number of well pads in Rio Blanco County, concluding that not all revegetation is equal.
“One of the things we noted in our study is that the species that are there matter,” Alward said.
Revegetation can mean weeds, or something palatable for mule deer to eat, he noted. Alward and Minnick’s research found that where revegetation involves rabbitbrush rather than sagebrush, there’s less recovery of healthy soil important to supporting regrowth of a range of plants.
FOCUSING ON AREAS OF NEED
But while land managers shouldn’t be satisfied just because land is green rather than bare, Alward said, the USGS research is important in its conclusions and in providing a new assessment tool.
“Their results kind of confirmed what we found, that well pads haven’t recovered very well, and they were able to look at a much larger area and draw similar conclusions,” Alward said. “There’s a problem with getting well pads to recover.”
He said the new USGS tool offers “some real potential to be able to pick out problem areas, maybe see patterns in what sort of systems our reclamation efforts are working in and the systems in which we’re having more trouble.”
The tool should help land managers prioritize and focus on areas of particular need, Alward said.
He pointed to other research finding that 7.5 million acres have been disturbed by oil and gas development in the central United States and Canada in a little over a decade. That’s roughly the size of Garfield, Mesa and Rio Blanco counties combined, he said.
“There’s a lot of disturbance and you’re going to need some big tool like what these guys have brought to bear to try to address it,” Alward said.
Said Duniway, “We are really hopeful that this new approach and associated tools can provide the needed understanding to guide land use decision processes in these landscapes now and into future.”
He said researchers have put in a funding request to develop web-based and mobile apps to let land management agencies, consultants and oil and gas producers do similar evaluations.
They also plan to look into associated questions such as how revegetation levels might relate to dust emissions and resulting air pollution in the region.
Meanwhile, other researchers such as Alward and Minnick will continue to look at how reclamation looks at a ground level. Their work involves some high technology as well. Alward said they’re planning field work this summer involving the use of drones, which will allow for the same kind of imaging work as the USGS researchers did with satellites, but with a much higher resolution. Whereas the USGS researchers sought to get images with at least eight pixels per pad, “we’re looking at getting that many pixels for each plant,” Alward said.