Estimating Abiotic Thresholds for Sagebrush Condition Class in the Western United States☆,☆☆,☆☆☆
Introduction
Sagebrush (Artemisia spp.) ecosystems of the western United States are imperiled (Chambers and Wisdom, 2009, US Fish and Wildlife Service, 2013). Threats to the ecosystems include wildfire, climate change, development, invasion of non-native annual grasses, and expansion of conifers (Chambers et al. 2017). The threats compromise the ecosystems’ abilities to provide services like clean water and air, wildlife habitat, forage for grazing, recreational opportunities, and biodiversity (Rose et al. 2015). The amount of area sagebrush currently occupies is little more than half its historical range (Chambers et al., 2017, Davies and Bates, 2019). Euro-American migration into the western United States, and the accompanying increase in disturbances and invasion of non-native grasses, coincided with sagebrush range reduction (US Fish and Wildlife Service, 2013, Chambers et al., 2017). Disturbances (e.g., land-use change, fire, overgrazing), often multiple compounding disturbances, caused sagebrush ecosystems to transition from extended periods of relatively stable conditions where native shrub and perennial grass species dominated to ecologically degraded conditions where non-native annual grasses invaded and now dominate. To identify transitional locations, we defined criteria for three classes of sagebrush condition, developed a dataset that reflected the classes, integrated the dataset with relevant independent variables into a decision-tree model, and used the resulting model algorithms to develop spatially explicit maps of sagebrush condition class. For purposes of this study, we named the three classes sagebrush recovery, tipping point (representing ecological degradation), and stable.
Most restoration efforts in sagebrush ecosystems have been minimally effective (Blomberg et al., 2012, Svejcar et al., 2017). Therefore, recovery to a sagebrush-dominated system after a disturbance can be expensive and take many years, if recovery ever occurs (Svejcar et al. 2017), although recent studies have preliminarily shown enhanced success of sagebrush restoration (Davies et al., 2018, Germino et al., 2018, Davies and Bates, 2019). Sagebrush ecosystems vary in their abilities to resist non-native annual grass invasion and recover from disturbance. Systems with low resistance and resilience were manifested in large geographical areas of non-native annual grass stands that have increased fire frequencies and threaten adjacent healthy rangeland systems. Chambers et al. (2007) found that specific factors influenced how vulnerable a sagebrush ecosystem was to cheatgrass (Bromus tectorum L) invasion, the most ubiquitous non-native annual grass in the study area. Climate, disturbance regime, the competitive abilities of the resident species, and traits of the invader were all influential factors of invasibility. Invasibility increased when resources were unused by native vegetation, such as after a fire (Rau et al., 2014, Roundy et al., 2018). Invasion also occurred when resource availability was inconsistent (Rau et al. 2014), which led to periods when resource supply exceeded the resident species’ ability to use it while invasive species’ propagule pressure existed (Davis et al. 2000). This phenomenon could have occurred in low to mid elevations of the sagebrush steppe (Chambers et al. 2014) where precipitation (ppt) exhibited high temporal variability (Bradley and Mustard 2005). In areas of relatively high perennial vegetation productivity, greater resource utilization by perennial vegetation reduced ecosystems’ invasibility (Chambers et al. 2014). How efficient an invading plant was at using resources when they become available could have determined a plant’s invasion success (Bansal et al. 2014). Greater resilience levels have been positively associated with higher ppt, greater soil resources, and higher plant productivity and linked to higher levels of resistance (Chambers et al. 2014).
The goals of this study were to identify abiotic variables that most influenced the prediction of three classes of sagebrush condition and to establish the most common environmental thresholds that characterized each class. We clarified and defined some of the abiotic characteristics and associated thresholds that influenced sagebrush ecosystems’ resilience to disturbance, invasibility to non-native annual grasses, and stability. We parameterized decision-tree software in two ways so that it generated 1) a predictive model based on a tree structure and 2) a descriptive model based on rulesets. These two models allowed us to achieve the following objectives:
- 1)
Develop a spatially explicit predictive map of three classes of sagebrush condition.
- 2)
Develop a spatially explicit ruleset map that shows where every ruleset occurred.
- 3)
Identify abiotic variables that most strongly drive development of the sagebrush condition class model.
- 4)
Establish thresholds of the most commonly used abiotic variables that delineate each sagebrush class.
Section snippets
Study Area
We focused our study on arid and semiarid sagebrush ecosystems in the western United States where annual grass invasion was likely, sagebrush was native, and wildlife and livestock graze. The study excluded areas higher than 2 250 m elevation because cheatgrass was much less likely to invade at elevations above ≈2 000 m in the northern Great Basin (Boyte et al. 2015) and because these areas were more resistant to cheatgrass invasion and more resilient to disturbances than areas at lower
Model Development
The predictive and descriptive models used independent variables to drive model development, and although the same variables were available to both models, the predictive model used all available variables, whereas the descriptive model used only some (Table 2). The two models also used variables at different frequencies. These phenomena primarily occurred because we developed the predictive model by applying few constraints to the decision trees that predicted classes, and we developed the
Discussion
Mapping areas of stability and areas of change using machine-learning algorithms allowed both the identification of dominant abiotic variables that drive ecosystem dynamics and the variables’ important thresholds. Both the predictive and descriptive models for this study were developed using multiple drivers that represented climatic, topographic, and soils data because these variables were considered among the most important drivers of invasibility of sagebrush ecosystems (Roundy et al. 2018).
Implications
The spatially explicit maps revealed that most of the study area has been altered from its native state and has never recovered. These tipping points are likely to persist in the future as natural recovery in sagebrush ecosystems is a long-term process and restoration efforts can be marginally effective. Oftentimes subsequent disturbances interrupt natural recovery and restoration projects. Some areas have recovered or have been restored as evidenced by sagebrush recovery pixels. Stable pixels
Acknowledgment
We thank Neal Pastick for his astute insights on the modeling process and his input to an earlier version of the manuscript.
References (42)
- et al.
The effects of precipitation and soil type on three invasive annual grasses in the western United States
Journal of Arid Environments
(2014) - et al.
Cheatgrass percent cover change: comparing recent estimates to climate change−driven predictions in the northern Great Basin
Rangeland Ecology & Management
(2016) - et al.
Validating a time series of annual grass percent cover in the sagebrush ecosystem
Rangeland Ecology & Management
(2019) - et al.
Identifying land cover variability distinct from land cover change: cheatgrass in the Great Basin
Remote Sensing of Environment
(2005) - et al.
Using resilience and resistance concepts to manage persistent threats to sagebrush ecosystems and greater sage-grouse
Rangeland Ecology & Management
(2017) - et al.
Longer-term evaluation of sagebrush restoration after juniper control and herbaceous vegetation trade-offs
Rangeland Ecology & Management
(2019) - et al.
Soil resources influence vegetation and response to fire and fire-surrogate treatments in sagebrush-steppe ecosystems
Rangeland Ecology & Management
(2014) - et al.
Using remote sensing to quantify ecosystem site potential community structure and deviation in the Great Basin, United States
Ecological Indicators
(2019) - et al.
Ecosystem performance monitoring of rangelands by integrating modeling and remote sensing
Rangeland Ecology & Management
(2012) - et al.
Spatiotemporal patterns of cheatgrass invasion in Colorado Plateau National Parks
Landscape Ecology
(2019)
Characteristics of climate and landscape disturbance influence the dynamics of greater sage-grouse populations
Ecosphere
The integration of geophysical and enhanced Moderate Resolution Imaging Spectroradiometer Normalized Difference Vegetation Index data into a rule-based, piecewise regression-tree model to estimate cheatgrass beginning of spring growth
International Journal of Digital Earth
A time series of herbaceous annual cover in the sagebrush ecosystem. U.S. Geological Survey data release
Estimating environmental thresholds for three classes of sagebrush condition in the western United States (2001 -2015). U.S. Geological Survey data release
Regional analysis of the impacts of climate change on cheatgrass invasion shows potential risk and opportunity
Global Change Biology
Cheatgrass (Bromus tectorum) distribution in the intermountain Western United States and its relationship to fire frequency, seasonality, and ignitions
Biological Invasions
Emerging technologies and cultural shifts advancing drylands research and management
Frontiers in Ecology and the Environment
Winter cold tolerance and the geographic range separation of Bromus tectorum and Bromus rubens, two severe invasive species in North America
Global Change Biology
Resilience to stress and disturbance, and resistance to Bromus tectorum L. invasion in cold desert shrublands of western North America
Ecosystems
What makes Great Basin sagebrush ecosystems invasible by Bromus tectorum?
Ecological Monographs
Priority research and management issues for the imperiled Great Basin of the western United States
Restoration Ecology
Cited by (3)
Agricultural Land Degradation in the United States of America
2023, Handbook of Environmental ChemistryTrends analysis of rangeland condition monitoring assessment and projection (RCMAP) fractional component time series (1985–2020)
2022, GIScience and Remote Sensing
- ☆
Current address: Yingxin Gu, I. M. Systems Group, Inc., at NOAA NESDIS Center for Satellite Applications and Research, College Park, MD 20740, USA.
- ☆☆
US Geological Survey Land Change Science and National Land Imaging programs and the Bureau of Land Management funded the study. The funders had no role in the study or preparation of this manuscript.
- ☆☆☆
Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US government.