Source: UNIVERSITY OF NEVADA submitted to
LONG-TERM TRENDS IN SAGE-GROUSE DEMOGRAPHY AND HABITATS ON THE SHELDON -HART MOUNTAIN NATIONAL WILDLIFE REFUGE COMPLEX AND ADJACENT BLM LANDS: AN OPPORTUNITY TO ASSESS IMPACTS OF FERAL HORSES AND REGIONAL HABITAT CONNECTIVITY
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
NEW
Funding Source
HATCH
Reporting Frequency
Annual
Accession No.
1000141
Grant No.
(N/A)
Project No.
NEV05284
Proposal No.
(N/A)
Multistate No.
(N/A)
Program Code
(N/A)
Project Start Date
Aug 19, 2013
Project End Date
Jun 30, 2018
Grant Year
(N/A)
Project Director
Sedinger, J.
Recipient Organization
UNIVERSITY OF NEVADA
(N/A)
RENO,NV 89557
Performing Department
Natural Resources & Environmental Sciences
Non Technical Summary
Sage-grouse were determined by U.S. Fish and Wildlife Service (USFWS) to be warranted for listing under the Endangered Species Act but listing was precluded by other higher-priority species. USFWS has been ordered by a federal district court judge to make a final determination by 2015. Numerous factors, including energy development, loss of habitat to agriculture (especially in Washington and southern Idaho), wildfire, diseases and livestock grazing, have been implicated in the decline. Potential impacts of livestock grazing are, however, confounded with the grazing impacts of feral horses, especially in the western part of the sage-grouse distribution. Ranchers have complained that they are unfairly blamed for, what are in fact, impacts caused by horses. Additionally, advocates for feral horses have resisted appropriate management of the abundance of horses on public lands. The Bureau of Land Management now spends about $75M annually on management of feral horses on their lands. It is therefore important to better understand the relative impacts of feral horses versus livestock on vegetation and the associated impacts on sage-grouse populations. The project proposed here takes advantage of a unique combination of historical data and current distributions of feral horses and livestock to separate the effects of horses and livestock on Great Basin vegetation and sage-grouse. Dr. Mike Gregg, USFWS, radio-tagged 904 sage-grouse on the Hart Mountain and Sheldon National Wildlife Refuges (NWRs), from 1988 through 2005. Dr. Gregg also measured key vegetation variables (grasses, forbs and other understory vegetation, shrub cover, etc.) that are important to sage-grouse. Dr. Gregg's data predate removal of livestock from Hart Mountain in the early 1990s. Dr. Gregg also collected data following removal of livestock on both refuges during the 1990s and 2000s. The current proposal repeats Dr. Gregg's field work for four field seasons on Hart Mountain and Sheldon NWRs, and on the Massacre-Vya Population Management Unit, adjacent to Sheldon over four field seasons. Feral horses have built up to large numbers (about 800) over the past decade on Sheldon NWR, while Hart supports only small numbers of horses that are having minimal impact on vegetation (G. Collins, U. S. Fish and Wildlife Service unpublished data). Repeating Dr. Gregg's work, therefore, provides the opportunity to study vegetation and sage-grouse in the presence of livestock only (Hart historical data), immediately post removal of livestock (Hart historical data), in the absence of horses or livestock (Hart new data), in the presence of horses only (Sheldon new data) and in the presence of both horses and livestock (Massacre-Vya new data). Dr. Gregg's historical data and new data will be analyzed together to understand the effects of different grazing (horses versus livestock) regimes on vegetation and key sage-grouse population parameters, including nest success, chick survival, hen survival, and habitat use. The proposed study represents a unique opportunity to understand different outcomes of grazing by livestock versus horses and will inform management of grazing on public lands.
Animal Health Component
0%
Research Effort Categories
Basic
30%
Applied
70%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1360820107070%
1210799107030%
Knowledge Area
136 - Conservation of Biological Diversity; 121 - Management of Range Resources;

Subject Of Investigation
0799 - Rangelands and grasslands, general; 0820 - Wild birds;

Field Of Science
1070 - Ecology;
Goals / Objectives
The main objective of this research is to separate effects on Great Basin vegetation and sage-grouse of feral horses from those of livestock. Specific objectives to meet the main objective are: Measure understory and shrub vegetation in relation to use by horses or livestock. Estimate breeding propensity, nest success, chick survival, adult survival and habitat use by sage-grouse in relation to vegetation variables and habitat use by horses or livestock. Estimate population trendsof sage-grouse in relation to vegetation variables and use by horses or livestock.
Project Methods
Field Methods Four person field crews will conduct field studies from March 1 through July during three breeding seasons on each Refuge unit. A three person crew, supervised by the Sheldon unit crew leader, will conduct the work on the Vya-Massacre PMUs. We will conduct an abbreviated field season (April 1-July) during the fourth year of the project to take advantage of radio-tagged females remaining from the earlier years of the project. Each refuge crew will be supervised by a Ph.D. student supervised by Jim Sedinger at the University of Nevada Reno. Crews will operate out of remote field camps near study areas to minimize travel and living costs during each field season. Sedinger and students have maintained a similar camp in Eureka County, Nevada for nine field seasons, while conducting the Falcon-Gondor transmission line study. The goal will be to capture 60 female sage-grouse using standard nightlighting techniques (Giessen et al. 1982, Wakkinen et al. 1992) at each of the three sites during each spring of the study. Each female will receive a size 14 aluminum leg band and a 22 g necklace-style ATS transmitter. Transmitters have a life of 383-766 days (Blomberg et al. 2010) and will have a mortality mode allowing us to determine when females die. Each female will be checked twice weekly until movements become localized, at which point the female's location will be approached sufficiently closely to determine the nest site. Nests will be checked twice weekly until their fates are determined. After nest fates are determined, shrub cover, and cover and height of understory vegetation, will be estimated at each nest and a corresponding random point using the same methods employed by Gregg et al. (1994). Broods will be monitored weekly using spotlighting (Blomberg et al. 2010). Females will be monitored monthly by aircraft during the nonbreeding season to allow estimation of monthly survival. We will also apply backpack-mounted, 30 g PTT_100 Argos/GPS tags to 10 sage-grouse from each of the Hart and Sheldon units (20 total) to provide more detail about movements of individuals from these two populations. Analyses and Hypothesis Tests We will use analysis of variance to evaluate variation in vegetation variables between Sheldon in the current period versus historic data from Sheldon, both periods form Hart Mountain, and the current period for Vya-Massacre. We will analyze nest survival data using modern maximum likelihood methods (Rotella et al. 2004, Kolada et al. 2009) implemented in Program Mark (White and Burnham 1999), because these methods are less biased than classical approaches (e.g., apparent nest success) and they allow explicit modeling of proposed biological hypotheses. For example, vegetation variables can be incorporated directly into models of nest survival to assess their importance to nest survival. We will incorporate landscape level measures of fire history and shrub cover into models of nest survival (Moynahan et al. 2007, Kolada et al. 2009). The hypothesis about spatial temporal variation in nest survival will be evaluated by comparing nest survival on Sheldon and Vya-Massacre during the current period with that on Sheldon in the past and Hart during both periods. The extensive data available from Dr. Gregg, combined with the nests to be studied in this proposal, will make these among the most powerful assessments of factors affecting nest success of sage-grouse ever conducted. We will analyze monthly female survival using known-fate survival models implemented in Program Mark. We will assess effects of landscape level vegetation, fire history, and location (Hart versus Sheldon versus Vya-Massacre) by time period (historical versus current) effects on monthly survival. We will include vegetation variables (e.g., shrub cover) as covariates in survival analyses. We will evaluate seasonal variation in survival because three different analyses of survival indicate lowest survival during early fall, followed by spring (Sedinger et al. 2011, Blomberg et al. 2010, Farinha unpublished). We will analyze brood survival using methods developed by Lukacs et al. (2004), which allow for variation in detection and lack of independence among brood mates. We anticipate that sample size will restrict this analysis to comparison of new data from Sheldon and Vya-Massacre (current grazing) with historic Sheldon combined with both periods for Hart Mountain. We will calculate seasonal home ranges and habitat use for the nesting, and early, and late brood-rearing periods and associate home rages with landscape level habitat variables (e.g. Atamian et al. 2010). We will also assess effects of grazing treatments on habitat use, home range size, movements from nesting to brood-rearing areas. Finally, we will use data from both GPS collared sage-grouse and aerial survey to assess connectivity among the Hart, Sheldon, Massacre-Vya and Beatys Butte areas.

Progress 10/01/16 to 09/30/17

Outputs
Target Audience:Nevada Department of Wildlife, Bureau of Land Management, U. S. Fish and Wildlife Service, livestock producers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Three graduate students are using this work as part of their Ph.D. degeree. Four undergraduates, recent graduates are using this work to gain additional experience in conducting field experience. How have the results been disseminated to communities of interest?We gave six presentations at professional conferences. We met with staff of the Hart Sheldon National Wildlife Refuge staff to explain our results. We presnted a webinar through the Great Basin Landscape Conservation Consortium and we met with stakeholders and funders to present results. What do you plan to do during the next reporting period to accomplish the goals?We plan to continue field work as in 2017.

Impacts
What was accomplished under these goals? We followed > 100 additonal radio-tagged female Sage-grouse, monitored their nesting behviaor and the survivla of resulting chicks. We also sampled vegetation at nests, brood sites and random points. Most significantly, we completed a spatial model of use of the landscape by feral horses and livestock which will serve as inputs to subsequent analyses.

Publications


    Progress 10/01/15 to 09/30/16

    Outputs
    Target Audience:Nevada Department of Wildlife U. S. Fish and Wildlife Service Bureau of Land Management Citizens of Nevada and the Great Basin Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Two Ph.D students continue to work on the project. Additionally, we employed 10 undergradautes or recent graduates who received training and gained field experience. How have the results been disseminated to communities of interest?Annual progress report has been distributed to cooperators including U. S. Fish and Wildlife Servie, Bureau of Land Management and Nevada Department of Wildlife. We met with staff of the Sheldon hart Refuge complex and will meet with Nevada Department of Wildlife, U. S. Fish and Wildlife Service personnel on March 8. What do you plan to do during the next reporting period to accomplish the goals?We are beginning to develop products from the first four years of research but we have aquired funding for four more years of work.

    Impacts
    What was accomplished under these goals? We measured vegetation at a total of 2629 locations combined over the three years spanning 2013 to 2016 (Fig. 5-1, Fig. 5-2, Fig. 5-3, Fig. 5-4). We measured vegetation at 544 nest locations, 867 brood locations, 523 spring random locations, and 695 summer random locations (Table 5-1). We did not record any spring random locations in 2013 at any of the study sites because Mike Gregg was only able to train us in vegetation sampling after the time period for spring vegetation sampling was past. Similar to previous work (Hagen et al. 2007), we did observe significant differences between nests and random points in canopy cover of sagebrush at all height categories. Grass cover differed between nests and random locations which corresponds to previous work (Hagen et al. 2007), but differs from what Gregg et al. (1994) observed for unsuccessful nests at Hart Mountain NAR where grass cover was similar for random locations and unsuccessful nests. Coefficient values indicate that cattle use negatively influenced grass cover, grass height, and tall sagebrush, while horse use negatively influenced tall sagebrush. At brood locations compared to summer random locations, we observed that the differences in vegetation characteristics varied for all measurements except summer precipitation. Cover of key forbs, other forbs total grass, low sagebrush, and tall sagebrush were greatest at brood locations. Random locations had greater grass height, medium sagebrush and non-sagebrush shrubs. Greater high sagebrush at brood locations is similar to what was found by Hagen et al. (2007) where all sagebrush cover was greater at brood locations. Coefficient values indicate horse use was negatively related with other forbs and positively related to non-sagebrush canopy cover, while cattle use was negatively related to grass cover, grass height, and low sagebrush cover. Using a multistate framework with live and dead recoveries and separating individuals by state allowed us to account for greater variation in our individuals and in our sampling techniques. Our results showed that the age of our transmitters does have an effect on our ability to recover mortalities. Acknowledging this effect will allow us to better correct for collar failure and improve our analyses in the future. Our estimates would be biased by unequal detection rates for the three different states, if we did not account for collar failure. Nesting birds, once found on a nest, were much easier to relocate than non-breeding birds. The type of detection effort, by people on the ground or by aircraft, also had very different detection probabilities, and was important for correcting survival estimates. Transition probabilities between states differed by year, which is consistent with our observations in the field. We observed birds beginning to nest much earlier in the season in 2015, than in previous years. Female survival estimates for this study varied by year, and were lowest in the fall and spring seasons, which is consistent with the findings of previous studies (Blomberg et al. 2013b; Moynahan et al. 2006). The first year of the study, 2013, had the lowest survival rates, however, the sample size was also smaller, so this could be biased. Birds on nests or with broods did not show significantly different rates of survival compared to concurrent non-breeders. Precipitation, as a spatial covariate had the greatest impact on fall female survival. The effect was very negative in 2015, which had the highest annual precipitation, but positive in 2013. This may indicate that there is a threshold amount of precipitation, above which reduced survival occurs in sage-grouse. Spending time in wetter areas, such as meadows may come at a higher cost for adults in some years, possibly due to increased predation or hunting pressure. There may also be a trade-off for females, in which their survival is lowered, but brood survival is increased. We will be investigating this possibility further.

    Publications


      Progress 10/01/14 to 09/30/15

      Outputs
      Target Audience:State and Federal Natural Resources Agencies including Nevada Department of Wildlife, U. S. Fish and Wildlife Service and Bureau of Land Management. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Three graduate students (2 Ph.D. and 1 M.S.) are being trained as part of this project. Addtionall, we employ 10 field technicians annually. These technicians are typically recdnt graduates and thier employment provides valuable training How have the results been disseminated to communities of interest?We meet with all cooperators annually. We have distributed a detailed progress report to all funders and cooperators. We have given several presentaitons at regional professional meetings. What do you plan to do during the next reporting period to accomplish the goals?We will continue field work and begin final analyses of data.

      Impacts
      What was accomplished under these goals? We estimated all demographic rates identified above. We are in the process of identifying habitats used by sage-graouse during key seasons: nesting and brood-rearing. We have completed analyses of vegetation dats for sites used for nests and broods. We ahve developed surfaces of use byu horses and livestock based on our sampling of feces. WE have not yet tied sage-grouse populaiton dynamics to horse and livestock densities. Detailed highlights of our work are below. Adult survival:Female survival estimates for this study varied by year, which is consistent with the findings of previous studies. Our analysis has provided moderate support for a cost of reproduction for female Greater sage-grouse. This cost in terms ofsurvival may be exacerbated by local precipitation. Survival estimates for 2015 were considerably lower than either of the previous years of the study, particularly for non-breeding birds during the spring and summer. Our calculated annual survivals indicate a very low probability of hens surviving through the entire year of 2015. This will likely have major implications for abundance and recruitment in our 2016 field season. Nest success:We were able to monitor 446 nesting attempts during the 2013, 2014, and 2015 field seasons. Daily nest survival increased with nest age (β=0.032, SE=0.009). We also observed evidence for an interaction (β=0.134, SE=0.056) between grass height (β=-0.084, SE=0.070) and total sagebrush cover (β= -200.138, SE=0.067). Nest success was highest at nests with both high sagebrush cover and tall grass, but increased with decreasing shrub cover for nests with low grass height. There was evidence for initiation date having a positive influence on nest survival with the cumulative model weight of 1. Survival increased as nests were initiated later in the season (β= 0.218, 0.072). We did not observe strong evidence that nest success varied among study areas nor did we see strong evidence for a year effect. All inference was drawn from a model that that allowed daily nest survival to vary with nest age, an interaction between grass height and sage cover, and initiation date. According to this model, the probability of a nest surviving the entire 37 day nesting period was 0.22 (SE 0.01). Brood habitat use and chick survival:From 2013 to 2015, we collected 610 locations from 278 sage grouse during the months of July and August for the North West Nevada dataset. Of these 610 locations, 105 represented females that had broods with them. For the central Nevada dataset, there were 239 total locations from 125 females, with 87 of the locations being brooding females. Females avoided areas with steep slopes (β= -0.178, SE=0.060). We also observed that females tended to avoid roads (β=0.205, SE=0.041). Most of the models that we fit were highly generalizable from the dataset used to train our models to the dataset collected in Eureka county. We monitored 183 hens with broods between 2013 and 2015. The best model in terms of AICc allowed chick survival to vary with the age of the chick, age at capture, and predicted late summer habitat values as additive effects. The average age of each chick when captured was 2.83 days old, however, all values reported are predicted for an initial age of 0. We observed chick survival increasing with age (β= 3.33, SE= 2.07) and decreasing with lower values of habitat quality predicted from the late summer habitat model (β=2.11, SE=0.84). Vegetation:We measured vegetation at a total of 1993 locations combined over the three years spanning 2013 to 2015. We measured vegetation at 384 nest locations, 717 brood locations, 360 spring random locations, and 532 summer random locations. We did not record any spring random locations in 2013 at any of the study sites because Mike Gregg was only able to train us in vegetation sampling after the time period for spring vegetation sampling was past. As correlations of seasonal maximum temperature to seasonal minimum temperature and average other forbs to average total forbs were greater than 75% for nest, brood and their respective random points, we removed seasonal minimum temperature and average total forbs from the MANOVAs. The nest MANOVA indicated that vegetation measurement means differ by all covariates except winter precipitation and an interaction of location type spring against random by year. We observed that grass coverdiffered between nest and random locations (df = 1, F = 4.03, p = 0.04). All but one measure of shrub cover differed between location type, only high sagebrush did not differ between nest and random locations. The brood MANOVA indicated that vegetation measurement means differed by all covariates (Table 5-4). When considered as univariates over an interaction of study area and location type, vegetation characteristics at brood locations differed from random locations for key forbs (df = 2, F = 3.04, p = 0.05), grass height (df = 2, F = 17.03, p = 0.00), and high sagebrush (df = 2, F = 10.16, p = 0.00). The total distance covered by transects in 2015 was 98.128 km over both sites. This resulted in a total of 10,204 observations of feces for 2015 and an overall number of observations of 20,717 for combined horse and cattle feces in all age classes. Horses We observed 4,987 observations of horse feces of which 471 were estimated to be less than one year old in 2015. There were 1,714 observations of horse feces less than one year old from 2013 to 2015 over 276 transects. We created density surfaces for feces less than two weeks old and for feces less than one year old. Spatial distribution of horse use based on two week or less observations did not change from year to year but densities across areas did change. The highest densities of horses occurred in 2013and ranged from 0 horses to 133.935 horses/day/0.09ha over two weeks and the lowest densities occurred in 2015and ranged from 0 to 1.66 horses/day/0.09ha over a two week period. Density of feral horses estimated from the less than or equal to one year observations did not vary among years and showed greater densities in horse management areas and had an inverse relationship with water. In 2015, we observed 5,134 cattle feces of which 608 were estimated to be less than one year old and 160 observations of feces less than two weeks old. Between 2013 and 2015 there was a total of 2,043 observations of cattle feces less than one year old, although eight of these were incidental observations from Sheldon NWR and not included in analyses. Greatest densities of cattle occurred in 2014 ranging 0.03 to 8.59 cattle/day/0.09ha over two weeks. Lowest densities occurred in 2015 ranging from 0.01 - 2.07 cattle/day/0.09ha over two weeks. For the two week observations, cattle occurred in greatest densities at low topographic index values and densities within horse management areas were lower than outside horse management areas. For the less than one year observations, locations with the greatest cattle densities were at the extremes of the range of topographic position index over a basin-wide area, were less likely to be used as distance to water increased and horse management areas, like the ≤ two week models, had lower densities. The densities we observed in the less than or equal to one year dataset were very similar between 2014and 2015.

      Publications

      • Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Jaster, L., P. Street, T. Behnke, and J. Sedinger. 2016. Using Fecal Surveys to Estimate Feral Horse and Cattle Densities in Northern Nevada. Nevada Chapter of the Wildlife Society Annual Meeting, (Abstract published), Reno, NV
      • Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Behnke, T., P. Street, L. Jaster, and J. Sedinger. 2016. Reproductive costs for female Greater-sage Grouse in Northern Nevada and Southern Oregon. Nevada Chapter of the Wildlife Society Annual Meeting, (Abstract published), Reno, NV.
      • Type: Conference Papers and Presentations Status: Published Year Published: 2016 Citation: Street, P.a., T. L. Behnke, L. A. Jaster, and J. S. Sedinger. 2016. Late Summer Habitat As A Limiting Factor Of Fitness For Greater Sage-Grouse. Nevada Chapter of the Wildlife Society Annual Meeting, (Abstract published), Reno, NV.
      • Type: Conference Papers and Presentations Status: Published Year Published: 2015 Citation: Street, P., and J. Sedinger. 2015. Greater Sage-grouse brood survival and habitat in the great basin: an opportunity to assess impacts of feral horses and livestock. Western Section of the Wildlife Society Annual Conference (Abstract published), Santa Rosa, CA.


      Progress 10/01/13 to 09/30/14

      Outputs
      Target Audience: Target audienceincludes sate and federal managmenet agencies and key NGOsconcnerned with amangement of the Great Basin. These include: Nevada Department of Wildlife, Oregon Department of Fish and Wildlife, Bureau of Land Management, US Fish and Wildlife Service, Greater hart Sheldon Conservation Fund, Nevada Chukar Foundation. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Two PhD students are being trained on this project. Additionally, 10 field tecyhn iciasn wee employed, providing valuable field experience for students and recent graduates. How have the results been disseminated to communities of interest? A detailed progress report has been distributed. We met with important agency stakeholders to review results and plans for 2015. What do you plan to do during the next reporting period to accomplish the goals? We paln another full field season in 2015. Additionally, the long-term historical data set (1988-2005) has been entered into a computer and will be available for analysis following the 2015 field season.

      Impacts
      What was accomplished under these goals? At the conclusion of the 2014 season we had marked a total of 843 sage-grouse: 504 males and 339 females since the projects inception. Models of female survival that allowed for a full month by area interaction failed to converge. A model that allowed survival to vary among months was the most parsimonious . Models that allowed for variation among areas or between juveniles and adults were not supported. Monthly female survival ranged from 0.81 for the interval between August and September 2013 to 0.99 during the winter months. Monthly survival was lowest in months during the nesting season and fall. Monthly survival was highest during summer and winter months (June-July, December-February) and lowest during the spring and fall (April-May, August-October), although survival during fall 2014 was higher than during fall 2013. Blomberg et al. (2013), observed a similar pattern in Eureka County Nevada. We observed survival rates during winter months (December 2013 - March 2014) consistent with other studies (Moynahan et al. 2006, Blomberg et al. 2013), however Anthony and Willis (2009) observed lower female survival associated with extreme winter weather on Hart Mountain during the winter of 1990-1991. Adult survival was almost identical among study areas and there was no support for models including variation among areas. We were able to monitor 293 nesting attempts during the 2013 and 2014 field seasons. Our model building strategy resulted in 2,803 total models. Of these models, 14 had an AICc weight greater than 0.01. Daily nest survival increased with nest age (β=0.06, SE=0.009). We also observed evidence for an interaction (β=0.20, SE=0.07) between grass height (β=0.29, SE=0.10) and total sagebrush cover (β= -0.05, SE=0.085). Nest success was highest at nests with both high sagebrush cover and tall grass, but increased with decreasing shrub cover for nests with low grass height. There was slight evidence for initiation date having a negative influence on nest survival with the cumulative model weight of 0.50. Survival decreased with nest that are initiated later in the season (β= -0.20, 0.10). We did not observe strong evidence of nest success varying among study areas with a cumulative weight of 0.18 nor did we see strong evidence for a year effect with a cumulative weight of 0.12. All inference was drawn from a model that that allowed daily nest survival to vary with nest age, an interaction between grass height and sage cover, and initiation date. According to this model, the probability of a nest surviving the entire 37 day nesting period was 0.21 (SE 0.01). In 2013 we measured vegetation at 183 random points, 63 at Hart, 60 at Sheldon and 60 at Massacre. There were 84 nest vegetation points. In 2014 we measured 177 (Hart: 60, Sheldon: 57, Massacre: 60) random vegetation points in the spring. There were 160 nest vegetation points measured. In 2013 we observed that only key forbs and total sage canopy were significantly different between nest and random vegetation. In 2014, total sage canopy cover was significantly different between nest locations and random points but key forbs were not significantly higher at nests as they were in 2013. In 2013 we measured vegetation at 183 random points during brood rearing, 63 at Hart, 60 at Sheldon and 60 at Massacre. There were 65 brood locations that were sampled for vegetation. In 2014 we measured vegetation at 183 random locations during brood rearing (Hart: 61, Sheldon: 62, Massacre: 60). There were 332 brood locations across all sites. In 2013 only key forb cover at brood locations differed significantly from random points with higher average cover at brood sites. We observed that total forb cover, key forb cover, other forb cover and total grass cover differed significantly between brood and random locations. All four cover variables were greater at broods than at random locations.

      Publications


        Progress 08/19/13 to 09/30/13

        Outputs
        Target Audience: Management agencies (USFWS, BLM, Nevada Department of Wildlife), State of Nevada, Conservtion Organizations. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Two PhD stuents are being trained as part of this project. Ten field technicaincs will also gain valuable experience as part of this research. How have the results been disseminated to communities of interest? Report has been distributed to all interested parties. What do you plan to do during the next reporting period to accomplish the goals? Continue protocols from eyar 1.

        Impacts
        What was accomplished under these goals? We captured 332 sage-grouse during the 2013 field season: 180 males and 152 females. We fitted 107 females with VHF transmitters and 8 with GPS backpacks during spring trapping (49 females at Hart Mountain, 38 at Sheldon NWR and 28 at Massacre). An additional 37 hens were added on the Massacre site during July and August. We also spent 19 nights attempting to capture additional hens at Sheldon NWR in August-September but were unsuccessful. We captured 78 males at Hart Mountain, 26 males at Sheldon NWR and 76 males at Massacre. We identified 86 nests initiated by radio-marked hens, 32 on Hart Mountain, 34 on Sheldon NWR, and 20 at Massacre. Clutch sizes (x± SE) were similar among sites (Hart: 7.45 ± 0.24, Sheldon 6.94 ± 0.26, Massacre 7.25 ± 0.30). Of those 86 nests, 14 hatched at Hart, 11 hatched at Sheldon, and 5 hatched at Massacre. We used 84 nests for analysis; the two nests not used were nest bowls abandoned before a single egg was laid. The best supported model for nest survival included covariates for site, hen age and tall grass. Other competitive models included covariates for total shrub cover, non-sage shrub cover and a quadratic trend in daily survival. We estimated nest success at Hart Mountain = 38% (95% CI: 20 - 56), while nest success at Sheldon and Massacre were similar at approximately 11% (95% CI: 3 - 25) and 10% (95% CI: 2 - 25), respectively. While nests success at Hart was three times greater than nest success as Sheldon or Massacre, the 95% confidence intervals overlapped. The 95% CIs for the betas for hen age and tall grass included zero indicating that they also were not significantly different. We observed 29 broods during 2013 and we monitored 27 of them. The best supported model of chick survival included a negative trend with brood age and covariates for low sage canopy cover, total grass percent cover, frequency of Lomatium spp., Antennaria spp., and Erigeron spp. The second best supported model only differed by having a quadratic trend in brood age. The first and second best models received more than 99% of the model weights (55.3% and 44.5% respectively). Brood survival did not differ among sites in any of the top models. Overall brood survival over the seven weeks from hatch to fledging was 24.6% (95% CI: 8.0- 45.7). Weekly chick survival decreased through the brood period. Brood survival was negatively affected as canopy cover of low sage and frequency of Lomatium spp. increased, and was positively affected by higher percent cover of total grass and higher frequencies of Erigeron. The effect of Antennaria spp. was not significant as the 95% confidence intervals included zero. We calculated adult survival using the known fates of 103 hens that were captured during the spring trapping periods. Hens captured during the late summer trapping period were not included nor were hens whose radios failed or they left the study area. Monthly survival was most influenced by hen age and mass at capture, as those covariates were included in all of the top models. Yearly survival of sage-grouse hens was estimated at 55.6% (Lower 95% CI: 45.3, Upper 95%CI: 64.8) when calculated using the top model. Survival of sage-grouse hens was lower for adults than juveniles and was negatively impacted by mass at capture in the top model. We observed that only in the ninth best model did site impact survival of adult and juvenile sage-grouse hens. We compared random vegetation measurements to vegetation measurements at nest sites. We measured vegetation at 183 random points, 63 at Hart, 60 at Sheldon and 60 at Massacre. There were 84 nest vegetation points. Of the six measurements of canopy cover only total sage and total shrubs differed between random and nest points Table 5. Both measurements were greater at nest sites than at random sites. Of the nine categories of herbaceous cover only key forbs and total forbs were significantly different between random and nest points. Both had greater cover at nest points. There were 16 key forb groups for which we measured frequency. Of those 16 we found that only three (Lomatium spp., Agoseris/Micoseris spp., Astragalus spp.) were significantly different at nest points compared to random points. Again all were more frequent at nest points. When we compared brood use points to random points none of the six types of canopy cover differed significantly. Six measurements of herbaceous per cent cover were significant with greater percent cover of those measurements at brood points. Of the 16 key forbs, five of them were significantly more common at brood than at random points. Lomatium spp., Agoseris/Micoseris spp., Astragalus spp., Trifolium spp. and Aster spp. occurred in greater frequency at brood points than at random points. We recorded 297 feral horse and 38 cattle fecal observations along 62.9 km of transects. We sampled 40 transects at Hart Mountain, totaling 21.3 km, and 24 transects at Massacre, representing 12.3 km. We observed 27 feral horse feces and 38 cattle feces at Massacre. Sample sizes at Massacre required pooling data from different strata by feces type for models of detection function to converge. At Sheldon we sampled 36 transects, that covered 21.6 km. We observed 270 feral horse feces at Sheldon and were able to analyze each stratum separately. The best model for feral horse feces at Massacre gave a density of 439.8 fecal piles per km2 (CI: 162.2- 1192.5). This translates into 91.6 (CI: 33.8 – 304.4) horse use days km-2 per two week period. The best model for cattle feces at Massacre yielded a density of 414.9 fecal piles per km2 (CI: 117.8 - 1461.3). This resulted in 36.1 (CI: 10.2 – 103.7) cow use days km-2 over a two week period. At Sheldon the best models resulted in a large difference in estimated density of feral horse feces per square kilometer in the 0.4 – 1.6 km stratum compared to the other strata (Table 11). Horse use days varied from 63 to 3270 use days km-2 per two week period among strata.

        Publications