The Starkey Project

Starkey Study Areas.

Starkey Project research occurs at the Starkey Experimental Forest and Range in northeast Oregon and replicate sites in the Western United States. Research focuses on ecology and management of elk, mule deer, cattle, and associated human activities, land uses, and disturbance regimes common to public and private lands. Topics are of high societal interest with major ecological, social, and economic implications that have influenced national and regional policies of public land management. Research is led and coordinated by the Starkey Ungulate Ecology Team.

Diverse partner collaborations in Starkey Project research and management applications have resulted in widespread acceptance and use of results to address a broad spectrum of national issues in natural resource management. More than 80 studies and 300 scientific publications have been completed as part of applied research for management.  

Scientists have relied on stakeholder engagement to identify topics of high relevance to management of forests and rangelands, and to implement the scientific process from start to finish.

Major changes in policy and management from use of Starkey Project findings include the following:

• Hunting research has dramatically improved harvest designs for elk by state wildlife agencies throughout the United States and Canada and enhanced diversity of hunting opportunities.
• Timber harvest and thermal cover studies have settled national controversies on these issues, resolving litigation and saving the U.S. Forest Service millions of dollars in land use planning.
• Motorized traffic and off-road recreation research was used by the U.S. Forest Service in developing national roadless policy, national off-highway vehicle policy, and travel management plans across the United States.
• Elk-deer-cattle grazing interactions studies resulted in new methods of allotment management planning for cattle grazing on U.S. Forest Service rangeland allotments in the Western United States.
• Elk nutrition and habitat use models developed and validated at regional scales have guided new silvicultural and road management strategies in Oregon and Washington, with approaches adopted in other ecoregions of the Western United States.
• Fire and fuels reduction research provided key information about management designs to benefit ungulate habitat use and performance, while reducing fire risk in forest ecosystems.
• Riparian and stream restoration research documented the benefits of sound restoration methods to recover threatened salmonids and evaluates the compatibility of new cattle grazing practices with salmonid recovery—while testing economic sustainability—for use in national forest allotments.  

Starkey Project research represents one of the longest running, continuous research efforts by U.S. Forest Service Research and Development (R&D), now spanning 30 years. Research was initiated in 1987 by then Pacific Northwest (PNW) Research Station research wildlife biologist Jack Ward Thomas, who became chief of the U.S. Forest Service in 1993. The legacy of Starkey Project research from Jack Thomas continues today with dedicated sharing of a permanent work force that includes 12 scientists, five field staff, and five data analysts among the PNW Research Station, Oregon Department of Fish and Wildlife, National Council for Air and Stream Improvement, and Oregon State University. Leveraging of staff as well as significant funding contributions from these partners continue to ensure the long-term success of this unique and invaluable research effort.

The long-term nature of Starkey Project research has generated a substantial data legacy; the project now manages one of the largest long-term datasets on animal locations, spatial data, animal performance, weather, human locations, and traffic that have been collected at landscape scales. These data provide diverse, compelling opportunities to test a broad range of hypotheses and address a variety of ecological theories by interested scientists and graduate students. Data documentation and archiving of foundational datasets began in 2018 to facilitate use by interested scientists and students worldwide. These data will reside on the U.S. Forest Service R&D Data Archive website. Additional long-term data from Starkey Project research will continue to be documented and archived on this website in the next several years.   

Purpose and Scope

The purpose of the Starkey Project is threefold:

1. Measure the habitat, behavioral, nutritional, energetic, and population responses of elk, mule deer, and cattle to intensively managed forests and rangelands at landscape scales at which management occurs.
2. Gain knowledge of the ecological effects and roles of elk, mule deer, and cattle as they contribute to ecological processes and patterns in managed forests and rangelands as part of disturbance regimes of herbivory and other disturbance processes.
3. Integrate socioeconomic systems studies with foundational ecological research.

The project’s scope includes the following research on disturbance regimes and agents common to forests and rangelands of the Western United States:

• Silviculture and Timber Management
• Fire and Fuels
• Elk-Deer-Cattle Interactions
• Ungulate Herbivory
• Hunting
• Trail-based Recreation
• Roads and Traffic
• Insect Pests
• Non-native Plants
• Riparian and Upland Forest Restoration

Methods

Experimental design and methods are supported by these unique technologies and research infrastructure at Starkey EFR:

Elk wearing tracking collars as part of the Starkey Project. Photo courtesy of Les Naylor.

• The world’s largest ungulate research enclosure (25,000 acres), allowing cause-effect landscape studies to be conducted that are essential to credible use of results in management.
• An automated telemetry system, generating more than 24 million deer, elk, cattle, and recreationist locations and representing the largest telemetry dataset on ungulates ever collected to evaluate land use activities.
• An automated system of traffic counters on roads and trails to evaluate effects of traffic and human activities.
• One of the largest, long-term datasets of environmental variables collected for long-term research.
• More than 30 years of hunting research that uses hunters as research treatments to address major questions about hunting season designs to meet hunting opportunity and harvest goals.
• A state-of-the art winter feeding and animal handling facility for animal care and evaluation of animal condition for research.
• Use of tractable (“tame”) elk, mule deer, and cattle for nutrition and herbivory studies to address critical management questions difficult to address with wild ungulates.

Scientists use Starkey EFR infrastructure to measure biologically significant responses and effects of ungulates at landscape scales with new methods to conduct the following:

• Estimate and model ungulate nutrition.
• Build and validate landscape assessment models for ungulates.
• Evaluate effects of land use activity on ungulates.
• Assess animal and population performance.
• Link animal performance to environmental change over long periods.
• Model animal movements, energetics, and resource selection.
• Document ungulate effects on ecosystem processes.
• Associate hunter behavior with hunting success.
• Evaluate real-time interactions of humans with animals.

Research designs and methods incorporate experimental approaches that document cause-effect processes in a variety of studies:

Game-proof fence. USDA Forest Service photo by Michael Wisdom.

• Controlled, manipulative experiments with defined treatments and paired controls.
• Before-after-control-impact designs for long-term monitoring and documentation of effects.
• Observational studies with independent data validation using spatial replication across ecoregions or biomes.
• Model building and validation with empirical data used in meta-analyses that synthesize landscape patterns across large areas.
• Simulation modeling with empirical and hypothesized relations to evaluate potential effects of multiple, interacting environmental factors in time and space that are difficult or impossible to study because of long time periods or complexity of spatial factors.
• Models that link data about nutrition and habitat use with animal condition and performance at landscape scales for long-term analyses of environmental change effects on populations.

Research findings are relevant to dry, mixed conifer forests and semi-arid rangelands in the Western United States. This inference space includes all public and private lands supporting these forest and rangeland types. Ungulate nutrition and habitat use models have been built and validated for uses across the Pacific Northwest, including ecoregions in Oregon, Washington, and Idaho. These modeling approaches have been replicated by other scientists in other areas of Idaho and Montana. Inference space of Starkey Project findings continues to expand across the Western United States with continued modeling that uses data from diverse landscapes, allowing spatial and temporal replication across an increasingly broad geographic area.

Key Findings

Silviculture and timber management

Elk nutrition trials with tractable (tame) elk. USDA Forest Service photo by Michael Wisdom.

• Commercial timber harvest increases forage biomass and quality for ungulates in forest stands during spring-early summer, with substantial ungulate use in these areas during this period when accumulation of body fat of lactating females is critical to successful production of young.
• Landscape mosaics of harvested and unharvested forest areas further benefit ungulates because unharvested stands maintain higher quality forage during late summer-fall when senescence has occurred in harvested units, and ungulates switch to use of unharvested areas.
• Elk do not require “thermal cover,” defined as forest stands with more than 70 percent overstory canopy cover, to mitigate or meet thermoregulatory requirements during summer or winter. Elk use thermal cover during late-summer fall, however, possibly in response to delayed forage senescence under these dense forest canopies or as security areas for refuge from hunters, other human disturbances, or predators.

Fire and fuels

Prescribed burn after mechanical fuels reduction. USDA Forest Service photo by Marty Vavra.

• Understory fuels reduction (non-commercial thinning and prescribed burning) to reduce fire risk provides benefits to ungulates in a similar manner as timber harvest and unharvested mosaics: fuels-treated forest areas produce higher forage biomass and quality during spring-early summer, and untreated areas do likewise during late summer-fall. Ungulate use follows this pattern, resulting in better nutritional choices for lactating females with high maintenance demands spring through fall.
• Ungulate herbivory reduces understory shrub development following disturbance “resets” such as fuels reduction and prescribed burning. Fuels-treated forest stands have lower density and height of nutritious shrubs when exposed to cattle, elk, and mule deer herbivory than do protected stands. Below-ground herbivory by pocket gophers further magnifies ungulate effects. 

Elk-deer-cattle interactions

• Cattle, elk, and mule deer compete for limited forage during late summer-fall when diets converge on shrubs following senescence of grasses and forbs. Elk have the most flexible diet and cattle are well adapted to a grass diet during late summer-fall senescence, but mule deer remain focused on shrubs. Thus even small increases in shrub intake by elk and cattle during late summer-fall may negatively affect mule deer nutrition and increase exploitative competition with elk and cattle.
• Elk avoid cattle and mule deer avoid elk, suggesting interference competition. Cattle rotation from one pasture to another documented these avoidance effects, with elk moving away when cattle arrive in a pasture, and further avoidance and compressed space use by mule deer largely explained by landscape choices of cattle and elk. However, elk may benefit from increased foraging efficiency and regrowth of grasses following cattle grazing, but these benefits do not extend to mule deer because of their avoidance of elk and heavier reliance on a shrub diet.
• Spatial partitioning and potential competition among cattle, elk, and mule deer can be accounted for with forage allocation models developed and published from Starkey Project research, and used in demonstration range allotments. Forage allocation concepts from Starkey have been adopted in public range allotments to improve allocation among ungulates and modify cattle stocking rates to meet ecological and natural resource goals.

Trail-based recreation

Horseback riders participating in a recreation study. USDA Forest Service photo.

• Elk avoid all-terrain vehicles (ATVs), mountain bikes, hikers, and horseback riders, all common recreational activities on public lands. Avoidance is strongest to ATV riding, intermediate to mountain biking, and weakest to hiking and horseback riding. Avoidance includes population shifts from both trails and recreationists; increased movement rates, energetic costs, and flight responses; and decreased foraging times, all of which reduce can reproductive potential.

Hunting

• Hunting seasons that result in few or no mature bull elk available for fall breeding causes a delay and asynchrony in conception dates of female elk, affecting productivity. Harvest designs that allow survival of mature males result in efficient breeding and increased population productivity, and these designs are now commonly used by state wildlife agencies in North America based on this Starkey Project research.
• The indirect effects of hunting on female elk not targeted for harvest can be biologically significant. Female elk increase their movement rates and energy expenditures during hunting, resulting in lower body fat, which can affect productivity. Hunting season designs now consider the indirect effects of multiple hunting seasons on non-targeted animals and population objectives.
• Archery hunting for either sex of elk can affect female productivity when hunter densities are high and season lengths are long (e.g., five weeks). The cumulative effects of extended archery hunts can reduce pregnancy rates for females in poor nutritional condition, and hunting season designs now consider these potential outcoes to avoid undesired indirect effects.
• Landscape strategies used by hunters affect hunting success. Successful archers for male elk hunt far from motorized access, effectively using remote areas in more complex terrain. Unsuccessful archers hunt closer to open roads and use more gentle terrain. Rifle deer hunters stay close to motorized access regardless of success. Successful rifle elk hunters, in contrast to archers and rifle deer hunters, effectively use roads closed to motorized access. Results provide a direct link of hunting season designs by state wildlife agencies with motorized access management by federal agencies as they jointly affect harvest objectives and desired hunting experiences.
• Hunters have a wide range of preferences for motorized access, road closures, hunting experiences, and hunting season designs. State wildlife agencies consider these preferences to design hunting seasons, and federal land management agencies can likewise consider them in managing motorized access.

     Starkey Experimental Forest and Range - Hunt Map
     How-to Guide: Using Avenza Maps* to Download Starkey Hunt Map
 

Roads and traffic

Motorized vehicles passing a traffic counter. Photo courtesy of Dane Johnson.

• Elk avoid roads open to motorized traffic, and this relationship is best captured in distance-to-roads analyses. Previous evaluations of road effects on elk, which relied on metrics of open road densities, did not accurately capture the spatial effects of open road avoidance by elk.
• Elk avoidance of roads increases with rates of traffic. Mule deer avoid elk and select areas increasingly close to roads as traffic rates increase, suggesting interference competition by elk that results in spatial compression of areas used by mule deer. Reduced areas of use by elk and mule deer in relation to roads reduces carrying capacity.

Ungulate nutritional ecology

• Summer nutrition is a limiting factor for elk productivity in the West, given the high metabolic demands of lactating female elk to provide sufficient resources for calf survival and growth. Summer nutritional demands of female elk vary by geographic area with variation in potential vegetation types and forest structural conditions, which affect the capacity of landscapes to respond to nutritional enhancement via use of silviculture and fire.
• Landscape limitations in summer nutrition for female elk affect accretion of body fat, pregnancy rates, timing of pregnancy, and calf growth and survival. Landscapes with higher levels of summer nutrition result in higher pregnancy rates and body fat of female elk, which are key to population productivity. 

Landscape wildlife models

• Elk consistently select areas of higher summer nutrition, resulting in better animal performance in more nutritionally rich landscapes. Summer nutrition can be enhanced through active silviculture (timber regeneration harvest and commercial thinning) and use of understory prescribed fire.  It is likely that wildfire generally enhances summer nutrition as well but is understudied.
• Non-nutrition factors of human disturbance, vegetation, and topography substantially affect summer elk use of landscapes and dictate the degree to which areas of better nutrition are used. These landscape relationships were captured in nutrition and habitat-use models for management applications in Oregon and Washington that have been adopted as formal planning tools by the U.S. Forest Service and U.S. Bureau of Land Management. (LINK)
• Meta-analyses of contemporary landscape data available from multiple populations of elk and mule deer provide appropriate inference space for modeling elk and deer nutrition and use across large ecoregions or biomes of the Western United States. Improved satellite and global positioning system telemetry systems provide the volume and quality of ungulate locations needed for meta-analyses and strengthening of inference space for management.
• Cattle select habitats based on seasonal changes in forage phenology and weather. In early summer, open forests and grasslands are selected in response to higher forage biomass and quality. In late summer-early fall, cattle increasingly select riparian areas and dense forests as forage senesces in open habitats, and weather becomes drier and hotter.

Riparian and upland forest restoration

Meadow Creek riparian restoration area. USDA Forest Service photo.

• Habitats and populations of threatened salmonids recover with intensive stream inputs of large woody debris. Intensive placement of large woody debris substantially increases pool areas used for summer juvenile rearing, and gravel areas used for adult spawning.
• Elk herbivory can severely depress survival and growth of riparian shrubs, impeding restoration of canopy cover needed to increase stream shading and reduce stream temperatures to ensure survival of juvenile salmonids during summer. Improving distribution of elk in upland areas through silvicultural improvements in nutrition and reduced human disturbance from motorized access would help mitigate elk effects on riparian shrub recovery.
• Native bee diversity in riparian areas is high, and potential diet overlap between bees and ungulates is also high, suggesting that ungulate herbivory may affect floral resources used by bees. Accounting for native pollinators in grazing riparian objectives may be a future consideration for federal riparian standards.
• Compatibility of new cattle grazing practices with salmonid recovery is constrained by management direction for minimizing streambank alteration, further constraining economic sustainability of grazing in riparian pastures. Innovations in cattle management to maintain salmonid recovery goals and economic sustainability will be challenging but options are under study.