New technology in data collection benefits reclamation specialists in Wyoming

Douglas – University of Wyoming Graduate Student Michael Curran is working toward more efficient surveys of reclamation sites disturbed by mineral extraction.

“If we can be in and out of the field quickly for monitoring, we can then invest more time into doing the groundwork,” Curran explained.

Collecting data for his master’s degree work, he realized that vegetation surveys vary greatly between companies and field technicians.

“Monitoring reports always look different. Between years, monitoring reports change style and the timing of monitoring changes between years. When there are two different technician groups or reclamation contracting companies, they are usually doing something different,” Curran remarked.

Results can also be subject to bias, age, fatigue or expectations.

“I want to capture the spatial heterogeneity over a well pad, so I have an unbiased estimate of what my actual, true populations are at that pad,” Curran stated.

Random samples

To obtain accurate records, Curran is working on a quasi-random sampling method, to ensure an accurate capture of the big picture at any given site, including areas of strong vegetation, areas of bare ground and everything in between.

To begin, Curran uses a computer program, drawing a polygon over the surface of the disturbance site.

“The first time I used this equipment, I arbitrarily selected 16 points and added a buffer so that no point could be within 15 meters of any other point,” he commented.

In other words, the computer randomly selected points within the polygon, rejecting any points that fell within 15 meters of another point, until 16 dots had been placed onto the map.

“The first time, it took me 20 minutes to get to all of the 16 points,” said Curran.

Using a GPS tool, he traced his path from point to point, taking a photo at each spot to ensure his accuracy. He then compared his own photos to Google Earth to further support his data.

Image analysis

Looking at the pictures from the site, Curran noted, “We can see that we have patches of bare ground, we have some dense vegetation areas, and we have some that look moderately vegetated.”

When he returned to the lab, Curran uploaded his photos into his computer program to analyze the vegetation populations.

“It is essentially a digital Daubermire square,” he stated.

By using pre-generated buttons within the computer program, Curran identified the range of objects within his photographs.

“For the sake of example, I might mark grasses, forbs, shrubs, cacti, litter, soil, rocks, invasives and unknowns,” Curran explained.

As he further develops his program, those labels will become more specific, including exact species of vegetation.

“At the end of the day, I group those 16 pictures together. I analyze them and, depending on how dense the vegetation is, it takes me 35 seconds to one minute and 20 seconds of analysis per picture,” he continued. “In under 40 minutes, both time on the pad and analyzing, I had a report generated and ready to use.”

Sufficient data

The second time Curran used his process to record data, he was able to capture 35 images in roughly 15 minutes at the well pad site, followed by approximately 35 minutes to complete reporting.

“In that case, I had 35 sample frames with 30 sample points per image, so I had 1,050 points for that well pad,” he noted, comparing his numbers to a traditional survey that would result in one sample frame and 50 sample points.

By obtaining a greater number of sample points, Curran believes that his method provides more statistical power when calculating reclamation data.

“Also, there is a permanent record. With the images, if something doesn’t make sense, I can go back and reanalyze the data, or I can have someone else do it,” he noted.

Curran’s method also allows for consistency in analysis.

Data integrity

“We can train 25 field techs to take a picture the same way, and we can have one person analyze, taking the bias of different vegetation analysis skills,” he explained.

The results provide evidence that technicians were at the correct site, that sample points were generated without bias and that data is collected efficiently and cost-effectively.

“The three most important words in any stakeholder’s vocabulary are, ‘make informed decisions,’” Curran mentioned.

He hopes that his monitoring method will allow reclamation specialists to be more accurate in their reports.

“We have to construct our plans to hit a moving target. I think the only way we can really be confident in hitting our moving target is to have a strong monitoring plan in place,” stated Curran.

Michael Curran spoke in Douglas on July 22 at the Douglas Reclamation Plan Workshop, hosted by the University of Wyoming Reclamation and Restoration Center and Extension.

Natasha Wheeler is editor of the Wyoming Livestock Roundup and can be contacted at natasha@wylr.net.