Greater Sage-Grouse Population Monitoring Framework: Trends Analysis Information Sheet
Land and wildlife managers require accurate estimates of sensitive species’ trends to help guide conservation decisions that maintain biodiversity and promote healthy ecosystems. Accurately assessing greater sage-grouse (Centrocercus urophasianus; sage-grouse) population trends can be difficult because (1) missing lek counts or incomplete repeat counts, (2) variation in counts of sage-grouse from observer error, (3) administrative units do not match biological/population boundaries that capture drivers of regional and population-level performance, and (4) sage-grouse populations naturally vary across years leading to population cycles that operate irrespective of long-term trends. The purpose of this analysis was to estimate trends of sage-grouse populations across their range in the United States in a manner that addressed each of these obstacles.
Accounting for Imperfect Lek Counts
A substantial challenge facing trend estimation for sage-grouse populations are the incomplete timeseries of count data associated with many leks. Previous studies have dealt with this by excluding populations with irregular data. However, a shortcoming to this approach is that inferences can be restricted in both time and space. Therefore, we implemented a modeling approach that allowed for information to be shared across aggregations of spatially structured (similar size and movement) leks. The observation component of that model was aimed to address issues of imperfect detection which are well documented for sage-grouse populations.
Assessing Trends at Relevant Spatial Scales
This modeling approach included three spatial scales that are biologically relevant to sage-grouse: (1) the lek scale, which is the smallest scale that reflects the scale at which managers monitor sage-grouse populations, (2) the Neighborhood Cluster scale, which captures movement among leks and represents trends that may be driven by local conditions (Figure 1a), and (3) the Climate Cluster scale, which captures trends that are driven by broad scale climatic conditions (Figure 1b). Estimates of population change at lek and Neighborhood Cluster scales were made available during years of missing data by sampling from probability distributions that were based on (1) hierarchically nested spatial relationships and (2) within population variability in rate of change parameters estimated from the available data.
Assessing Trends at Relevant Temporal Scales
Inter-annual variation in abundance is common among sage-grouse populations and stems from cyclic climatic conditions that have been linked to availability of resources. This relationship results in a pattern of relatively short increases and decreases in population size that operate irrespective of long-term trend (Figure 2a). Calculating trends from short-term gains and losses can lead to erroneous conclusions about population performance. Recent studies have demonstrated that sage-grouse oscillations occur at intervals of 6–12 years on average. As such, studies of sage-grouse population trend should consider datasets spanning multiple decades and should consider start and stop years that represent one or more complete oscillations. Therefore, we restricted inferences to complete oscillations, and we examined population trends using multiple time periods (Figure 2b), each containing a different number of complete oscillations over the past 60 years (1960–2023). We used population abundance nadirs (low points), versus apexes (high points), to define start-stop temporal scales of inference. Variability was lower among nadirs compared to apexes producing more consistent estimates of population trends and was considered more relevant for managers since populations are at greater risk of extirpation when abundance is low.
Results and Implications
There was an average 2.9% annual decline in sage-grouse abundance across the range (Figure 3). We found an 80.1% decrease in abundance over the long time period (55 years) and a 42.5% decline in the short time period (19 years). There was variation in trends by climate cluster (Figure 4). The Bi-State Climate Cluster (CC-A) had the lowest recent rate of decline (1.2%), while the Washington Climate Cluster (CC-B) had the greatest decline (5.5%). As an extension of our analysis, we projected abundance into the future for each lek and Neighborhood Cluster using three temporal scales that reflected two- (short), four- (medium), and six-periods (long) of oscillation. We then calculated the proportion of the possible abundance distribution that was less than two male sage-grouse (minimum number to represent a lek) for the last prediction year of each temporal scale. Although this value is not true extirpation (zero or one bird), we refer to it as extirpation to align with state definitions of lek inactivity. Thus, this proportion of the distribution represented the probability of extirpation for each lek and Neighborhood Cluster at a nadir, approximately at short, medium, and long temporal scales into the future. Extirpation of leks within a Neighborhood Cluster was thought to reflect a loss of a meta-population because of reduced demographic rates.
This framework allows managers to assess and compare trends across different time periods and spatial extents. Because of the use of nadir-to-nadir (trough-to-trough) comparisons, managers can accurately compare trends over the short time period to trends over the long time period to determine if populations are declining due to chronic stressors or due to recent disturbance. Also, comparing trends across different spatial scales can help elucidate the cause of sage-grouse declines and help guide management actions. Trends at the Climate Cluster scale can be used to assess regional performance of sage-grouse populations and point to declines driven by climatic factors. Trends within neighborhood clusters can reveal information regarding local disturbances that may be mitigated with management actions. The extirpation probabilities from the analysis can help managers prioritize declining populations by pointing to areas where populations are at risk of being lost.
Future co-production
We continue working with all collaborators to improve sage-grouse management tools. Each year, a new standardized database is developed to include newly digitized historical data and improve data quality using critical quality control methods. These data are incorporated into the hierarchical population modeling framework to produce results for use in annual decision making. Because sage-grouse populations oscillate at 6–12-year intervals, population trends will change more slowly depending on the timing of the interval.
Data restrictions
State wildlife agencies collect and manage lek databases. Because sage-grouse are a species of conservation concern and sensitive to activities during breeding, these data are available only after acquiring data-sharing agreements with individual states.
Funders
U.S. Geological Survey (Ecosystem Mission Area, Land Management Research Program and Species Management Research Program, Wyoming Landscape Conservation Initiative) and U.S. Bureau of Land Management.
Partners
State Wildlife Agencies (California Department of Fish and Wildlife; Colorado Parks and Wildlife; Idaho Department of Fish and Game; Montana Fish, Wildlife & Parks; Nevada Department of Wildlife; North Dakota Game and Fish Department; Oregon Department of Fish and Wildlife; South Dakota Department of Game, Fish and Parks; Utah Division of Wildlife Resources; Wyoming Game and Fish Department; Washington Department of Fish and Wildlife), Colorado State University, BLM, US Fish and Wildlife Service, US Forest Service, researchers who provided field data to evaluate results.
Greater Sage-Grouse Population Monitoring Framework
Data Harmonization for Greater Sage-Grouse Populations
Estimating trends for greater sage-grouse populations within highly stochastic environments
Hierarchical Units of Greater Sage-Grouse Populations Informing Wildlife Management
Land and wildlife managers require accurate estimates of sensitive species’ trends to help guide conservation decisions that maintain biodiversity and promote healthy ecosystems. Accurately assessing greater sage-grouse (Centrocercus urophasianus; sage-grouse) population trends can be difficult because (1) missing lek counts or incomplete repeat counts, (2) variation in counts of sage-grouse from observer error, (3) administrative units do not match biological/population boundaries that capture drivers of regional and population-level performance, and (4) sage-grouse populations naturally vary across years leading to population cycles that operate irrespective of long-term trends. The purpose of this analysis was to estimate trends of sage-grouse populations across their range in the United States in a manner that addressed each of these obstacles.
Accounting for Imperfect Lek Counts
A substantial challenge facing trend estimation for sage-grouse populations are the incomplete timeseries of count data associated with many leks. Previous studies have dealt with this by excluding populations with irregular data. However, a shortcoming to this approach is that inferences can be restricted in both time and space. Therefore, we implemented a modeling approach that allowed for information to be shared across aggregations of spatially structured (similar size and movement) leks. The observation component of that model was aimed to address issues of imperfect detection which are well documented for sage-grouse populations.
Assessing Trends at Relevant Spatial Scales
This modeling approach included three spatial scales that are biologically relevant to sage-grouse: (1) the lek scale, which is the smallest scale that reflects the scale at which managers monitor sage-grouse populations, (2) the Neighborhood Cluster scale, which captures movement among leks and represents trends that may be driven by local conditions (Figure 1a), and (3) the Climate Cluster scale, which captures trends that are driven by broad scale climatic conditions (Figure 1b). Estimates of population change at lek and Neighborhood Cluster scales were made available during years of missing data by sampling from probability distributions that were based on (1) hierarchically nested spatial relationships and (2) within population variability in rate of change parameters estimated from the available data.
Assessing Trends at Relevant Temporal Scales
Inter-annual variation in abundance is common among sage-grouse populations and stems from cyclic climatic conditions that have been linked to availability of resources. This relationship results in a pattern of relatively short increases and decreases in population size that operate irrespective of long-term trend (Figure 2a). Calculating trends from short-term gains and losses can lead to erroneous conclusions about population performance. Recent studies have demonstrated that sage-grouse oscillations occur at intervals of 6–12 years on average. As such, studies of sage-grouse population trend should consider datasets spanning multiple decades and should consider start and stop years that represent one or more complete oscillations. Therefore, we restricted inferences to complete oscillations, and we examined population trends using multiple time periods (Figure 2b), each containing a different number of complete oscillations over the past 60 years (1960–2023). We used population abundance nadirs (low points), versus apexes (high points), to define start-stop temporal scales of inference. Variability was lower among nadirs compared to apexes producing more consistent estimates of population trends and was considered more relevant for managers since populations are at greater risk of extirpation when abundance is low.
Results and Implications
There was an average 2.9% annual decline in sage-grouse abundance across the range (Figure 3). We found an 80.1% decrease in abundance over the long time period (55 years) and a 42.5% decline in the short time period (19 years). There was variation in trends by climate cluster (Figure 4). The Bi-State Climate Cluster (CC-A) had the lowest recent rate of decline (1.2%), while the Washington Climate Cluster (CC-B) had the greatest decline (5.5%). As an extension of our analysis, we projected abundance into the future for each lek and Neighborhood Cluster using three temporal scales that reflected two- (short), four- (medium), and six-periods (long) of oscillation. We then calculated the proportion of the possible abundance distribution that was less than two male sage-grouse (minimum number to represent a lek) for the last prediction year of each temporal scale. Although this value is not true extirpation (zero or one bird), we refer to it as extirpation to align with state definitions of lek inactivity. Thus, this proportion of the distribution represented the probability of extirpation for each lek and Neighborhood Cluster at a nadir, approximately at short, medium, and long temporal scales into the future. Extirpation of leks within a Neighborhood Cluster was thought to reflect a loss of a meta-population because of reduced demographic rates.
This framework allows managers to assess and compare trends across different time periods and spatial extents. Because of the use of nadir-to-nadir (trough-to-trough) comparisons, managers can accurately compare trends over the short time period to trends over the long time period to determine if populations are declining due to chronic stressors or due to recent disturbance. Also, comparing trends across different spatial scales can help elucidate the cause of sage-grouse declines and help guide management actions. Trends at the Climate Cluster scale can be used to assess regional performance of sage-grouse populations and point to declines driven by climatic factors. Trends within neighborhood clusters can reveal information regarding local disturbances that may be mitigated with management actions. The extirpation probabilities from the analysis can help managers prioritize declining populations by pointing to areas where populations are at risk of being lost.
Future co-production
We continue working with all collaborators to improve sage-grouse management tools. Each year, a new standardized database is developed to include newly digitized historical data and improve data quality using critical quality control methods. These data are incorporated into the hierarchical population modeling framework to produce results for use in annual decision making. Because sage-grouse populations oscillate at 6–12-year intervals, population trends will change more slowly depending on the timing of the interval.
Data restrictions
State wildlife agencies collect and manage lek databases. Because sage-grouse are a species of conservation concern and sensitive to activities during breeding, these data are available only after acquiring data-sharing agreements with individual states.
Funders
U.S. Geological Survey (Ecosystem Mission Area, Land Management Research Program and Species Management Research Program, Wyoming Landscape Conservation Initiative) and U.S. Bureau of Land Management.
Partners
State Wildlife Agencies (California Department of Fish and Wildlife; Colorado Parks and Wildlife; Idaho Department of Fish and Game; Montana Fish, Wildlife & Parks; Nevada Department of Wildlife; North Dakota Game and Fish Department; Oregon Department of Fish and Wildlife; South Dakota Department of Game, Fish and Parks; Utah Division of Wildlife Resources; Wyoming Game and Fish Department; Washington Department of Fish and Wildlife), Colorado State University, BLM, US Fish and Wildlife Service, US Forest Service, researchers who provided field data to evaluate results.