Sources/Usage: Public Domain. View Media Details

The overall mission of the Rangeland Condition Monitoring Assessment and Projection (RCMAP), a project based at the USGS Earth Resources Observation and Science (EROS) Center, is to provide crucial information to people managing Western ecosystems for recreation, livestock forage, erosion control, fire, and wildlife habitat—and that includes understanding habitat needs for the sage-grouse.

The project released its latest fact sheet on January 26, 2023, and for the first time, RCMAP is including tree canopy cover as one of the components it features. While rangeland managers are typically focused on grasses and shrubs, tree cover is also significant to the sage-grouse, a sagebrush-dependent species whose population was as large as 16 million a century ago and is estimated at only 200,000 to 500,000 today.

The story behind the addition of tree canopy to RCMAP’s product suite demonstrates the project’s responsiveness to rangeland managers’ needs, its constant refinement of the data it provides, and future directions for the project.  

Scientists and agencies like the Bureau of Land Management (BLM) have been using Landsat data processed at EROS to help study sage-grouse habitat for decades. Currently, RCMAP helps identify, pixel by pixel, where sagebrush occurs and, more importantly, where it is retreating over time, reducing habitat for the sage-grouse and hundreds of other species dependent on the shrub.

For rangelands across the West, RCMAP pinpoints what percentage—or fraction—of each 30-meter by 30-meter Landsat pixel is contributed by a series of components. Shrub, sagebrush, bare ground, herbaceous, annual herbaceous, and litter were included in the first RCMAP data release in 2021. Since then, partly in response to user feedback, the project has added perennial herbaceous, non-sagebrush shrub, and now tree canopy cover to its annual time-series estimate, available on the Multi-Resolution Land Characteristic Consortium’s website for 1985-2021.

Pinyon-Juniper Encroachment

But why tree canopy? RCMAP project manager Matthew Rigge notes that there are plenty of trees on BLM-managed land, as well as rangeland in areas under U.S. Forest Service jurisdiction, and the project’s maps include all federal, state, and private land. “We don’t care—Landsat sees it all,” he said.

However, the impetus for tree canopy’s inclusion remains protecting sage-grouse habitat. And that means protecting the sagebrush ecosystem, an increasingly challenging prospect across the West.

While the popular image from movies is of dry, open landscapes filled with sagebrush, the shrub also thrives in higher elevation areas where there’s more moisture—and more trees. In particular, pinyon-juniper woodland, which encompasses a number of species such as Western juniper, pinyon pine, and three-needle pine, has been expanding downslope for a variety of reasons. “It’s invading some of the very best quality sagebrush habitat that exists,” Rigge said. Estimates are that pinyon-juniper coverage has grown by 125% to 650% since European settlement in the 1800s, and 90% of that growth has been into sagebrush shrublands.

This expansion absorbs sage-grouse territory and is detrimental to the species in other ways, said Cameron Aldridge, research ecologist and branch chief at the USGS Fort Collins Science Center. The conifers provide “big, tall perches. These trees provide raptors and species that prey on sage-grouse places to hunt from, which results in lower sage-grouse survival,” he said.

Cross-Agency Effort

In response to decreasing sagebrush in general and sage-grouse habitat in particular, the Bureau of Land Management and the Intermountain West Joint Venture collaborated to form the Partnering to Conserve Sagebrush Rangelands initiative, which includes hundreds of federal, state, tribal, and local agencies as well as organizations and private landowners.

Where pinyon-junipers invade sagebrush habitat, a variety of methods can be used to curtail conifer incursions. Cutting trees and burning are typical measures; “chaining” is another option. “They literally take two giant bulldozers and get a ginormous ship’s anchor chain and just go through these pinyon-juniper forests and knock everything over,” Rigge said. “And if you really want to do it well, you turn 90 degrees and then repeat the process going the other direction.”

All options have some drawbacks. Burning is more effective but can make treated areas more vulnerable to invasive annual grasses. Cutting and chaining have to be repeated because the conifer seeds remain.

The measures have been successful in holding the annual 1.6% pinyon-juniper expansion at bay, with one Oregon study seeing a 12% increase in sage-grouse population in response to conifer treatments.

All of these methods can benefit from the careful analysis and detailed data provided by RCMAP to determine the best locations for action. For example, studies suggest that it’s easier to cull pinyon-juniper in places where the young saplings are just entering into sagebrush territory, so observing the year-over-year change where the fractional component of tree cover is expanding gives range managers an idea where to focus their efforts.

Sources/Usage: Public Domain. View Media Details

Pinyon Jays

However, rangeland ecologists also are watchful to ensure that measures meant to protect the sage-grouse don’t end up causing erosion, creating opportunities for invasive species, or reducing habitat for other species such as the pinyon jay, which thrives in pinyon-juniper woodland. The blue-headed bird is more abundant than the sage-grouse but still regarded as vulnerable because the species’ population has been declining at similar rates to sage-grouse over time.

Recent research indicates that pinyon jays thrive in the ecotone, or area where two ecosystems meet, where treatment is most effective to protect the sage-grouse: areas where pinyon-juniper woodland is just beginning to impinge on sagebrush. A recent study co-authored by Aldridge that modeled the habitat relationships of pinyon jays found that jays are most abundant in habitats that contain both high pinyon-juniper cover and an abundance of sagebrush. This work suggests restricting conifer treatments to places where pinyon jays and several other birds are least abundant. “We could shift where we chop trees, still have benefits for sage-grouse, yet reduce the losses to pinyon jays and other species in the pinyon-juniper ecosystems,” he said.

That might mean more work for land managers, and a perfect solution to balance the needs of the species involved might be difficult. “There are always tradeoffs in land management,” Rigge said. “Neither pinyon jay nor sage-grouse are doing particularly great.”

But sage-grouse and their habitat are threatened by much more than just pinyon-juniper encroachment.

Cheatgrass

In drier areas, cheatgrass is stealing territory from sage-grouse habitat, covering more than 100 million acres of U.S. rangeland. The exotic annual invasive species, imported in the late 1800s from Eurasia, outcompetes native plants because it greens up much earlier in the season, as early as February if there is a short span of nice weather. Cheatgrass’ lifespan is brief; in as little as two weeks, it dries out and leaves behind a thick thatch of litter that acts as a mulch, preventing native forbs—non-grass flowering plants—and grasses that make up the understory of sagebrush shrubs from robust growth.

The cheatgrass litter also is extremely flammable. When it burns, it takes everything with it, right down to the roots. And while cheatgrass and other exotic annuals are quick to return, sagebrush is not. “The nearest seed source for sagebrush is sometimes miles and miles away. Sagebrush will often not come back to those locations without human intervention,” Rigge said. Climate change adds even greater uncertainty to the mix as already dry landscapes face more years of drought and sagebrush competes with cheatgrass for moisture.  

Stuck between cheatgrass invasion in dry ecosystems and pinyon-juniper expansion in areas with more rainfall, “sagebrush is being squeezed from both ends,” Rigge said.

RCMAP Critical for Evaluation

In order to manage habitat effectively for sage-grouse, pinyon jays, and the myriad other species potentially affected by both the loss of and treatment of sagebrush shrubland, rangeland ecologists need to know where to expend time and money in the most effective way. RCMAP is an essential tool to make that assessment, whether the cause of a potential disturbance is from human activity, exotic annual grass invasion, overgrazing, pinyon-juniper expansion, or fire.

The project answers pertinent questions, including identifying which sagebrush acres are priority areas for conservation for sage-grouse in the first place. The species has specific percentages of sagebrush cover that are optimal for breeding and nesting, for example, and RCMAP’s fractional component analysis offers that level of specificity per pixel.

And RCMAP helps land managers determine where tough choices are necessary to decide where not to spend money, too. “How long does it take for sagebrush to come back to the same amount cover that it once was prior to that disturbance? In dry places, that can take 80-100 years” Aldridge said. “Thirty years from now, are those places even going to have the appropriate climatic conditions that allow sagebrush to persist there? RCMAP products can help us understand these questions.”

How It Works

RCMAP is continuously improving its ability to provide the answer to these and other questions. But in order to evaluate rangeland with the level of detail required, RCMAP needed fieldwork and a view closer than Landsat’s 30 meters.

Rigge was a member of the team that spread out to gather field data to compare to high quality WorldView imagery—2-meter resolution data with eight spectral bands—at 330 sites across the West between 2013 and 2018. “We made predictions at that high-resolution scale using field observations to train models across these high-resolution images, then used those predictions from the high-resolution scale to train our Landsat models,” he said.

The training model allowed scientists to re-evaluate existing Landsat imagery back through 1985. “Landsat data give us that time machine and allow us to go back and get these repeat observations through time, and just paint a lot richer picture of the spatial and temporal variance,” Rigge said. At this point, there have been four major iterations of the dataset, and Rigge estimates error rates were reduced 10% to 15% with each iteration. Every advance in machine learning or addition of training data improves results.

User feedback shows the effort has been worthwhile. “Where we map sagebrush is exactly the places where they're finding sage-grouse. So that's good news. We're mapping sagebrush in the right places,” Rigge said.

The “bad” news, from Rigge’s perspective, is that RCMAP has become so good at leveraging its existing training data, he didn’t get to go out in the field again to collect training data for adding tree canopy for the latest data release. A training pool of tree cover types was developed for RCMAP’s existing high-resolution images using artificial intelligence, then hand-edited and scaled from 2 meters to 30 meters. After that, Rigge said, “it was fairly plug and play to add this new component to our process and generate time-series of tree cover in addition to the rest of the components.”

What’s Next for RCMAP

Rigge’s goals include further increasing RCMAP’s accuracy and product development speed through various machine learning classifiers and artificial intelligence. In addition, there are always interesting science questions to investigate, he said, questions that get “put off for a rainy day that never comes” in the continuous race for improvement.

Landsat Next, with a planned launch date in 2030, also offers tantalizing opportunities, with three satellites creating a complete image of Earth every six days and offering 26 spectral bands, 15 more than are currently available. The six-day repeat images will be directly beneficial for cheatgrass observations, since the exotic annual invasive’s growing season is so short.

And every edge in understanding cheatgrass helps RCMAP—and the rangeland managers who depend on the project’s ground and now treetop data to preserve sage-grouse habitat.