SEARCH

SEARCH BY CITATION

Keywords:

  • aerial survey;
  • energy development;
  • grassland;
  • hierarchical distance sampling;
  • landscape characteristics;
  • natural gas;
  • oil;
  • roads;
  • shrubland;
  • Tympanuchus pallidicinctus

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

As with many other grassland birds, lesser prairie-chickens (Tympanuchus pallidicinctus) have experienced population declines in the Southern Great Plains. Currently they are proposed for federal protection under the Endangered Species Act. In addition to a history of land-uses that have resulted in habitat loss, lesser prairie-chickens now face a new potential disturbance from energy development. We estimated lek density in the occupied lesser prairie-chicken range of Texas, USA, and modeled anthropogenic and vegetative landscape features associated with lek density. We used an aerial line-transect survey method to count lesser prairie-chicken leks in spring 2010 and 2011 and surveyed 208 randomly selected 51.84-km2 blocks. We divided each survey block into 12.96-km2 quadrats and summarized landscape variables within each quadrat. We then used hierarchical distance-sampling models to examine the relationship between lek density and anthropogenic and vegetative landscape features and predict how lek density may change in response to changes on the landscape, such as an increase in energy development. Our best models indicated lek density was related to percent grassland, region (i.e., the northeast or southwest region of the Texas Panhandle), total percentage of grassland and shrubland, paved road density, and active oil and gas well density. Predicted lek density peaked at 0.39 leks/12.96 km2 (SE = 0.09) and 2.05 leks/12.96 km2 (SE = 0.56) in the northeast and southwest region of the Texas Panhandle, respectively, which corresponds to approximately 88% and 44% grassland in the northeast and southwest region. Lek density increased with an increase in total percentage of grassland and shrubland and was greatest in areas with lower densities of paved roads and lower densities of active oil and gas wells. We used the 2 most competitive models to predict lek abundance and estimated 236 leks (CV = 0.138, 95% CI = 177–306 leks) for our sampling area. Our results suggest that managing landscapes to maintain a greater percentage of grassland and shrubland on the landscape with a greater ratio of grasses to shrubs in the northeast Panhandle should promote greater lek density. Furthermore, increases in paved road and active oil and gas well densities may reduce lek density. This information will be useful for future conservation planning efforts for land protection, policy decisions, and decision analyses. © 2013 The Wildlife Society

The occupied range of lesser prairie-chickens (Tympanuchus pallidicinctus) has been reduced by >90%, a decline attributed to direct habitat loss from conversion of native grassland to cropland, livestock overgrazing, and invasion of woody plants, and indirect habitat loss from disturbance by energy development (Taylor and Guthery 1980, Applegate and Riley 1998, Hagen et al. 2004). As a result, the lesser prairie-chicken is currently proposed for protection as a threatened species under the Endangered Species Act (U.S. Fish and Wildlife Service [USFWS] 2012). Lesser prairie-chicken populations have faced steady declines during the past 100 years in Texas, USA (Jackson and DeArment 1963, Crawford and Bolen 1976, Sullivan et al. 2000), and Texas Parks and Wildlife Department (TPWD) estimated a minimum of 6,000 birds from mostly road-based surveys located in high quality lesser prairie-chicken habitat (Davis et al. 2008).

Texas currently produces the most wind-generated electricity in the United States (i.e., 22% of the nation's total; American Wind Energy Association 2012) and 5 Competitive Renewable Energy Zones (CREZ) were designated in west Texas to encourage further wind energy development (Electric Reliability Council of Texas [ERCOT] 2006). Transmission lines are being constructed to deliver wind-generated electricity from the CREZs to customers in urban centers (ERCOT 2006). The 2 CREZs in the Texas Panhandle overlap approximately 27% (3,288 km2) of the known occupied range of lesser prairie-chickens in Texas. Furthermore, the amount of active oil and gas wells in the occupied range has increased >80% during the past decade (Railroad Commission of Texas 2012). However, little is known about the relationship of lesser prairie-chicken leks in Texas relative to existing landscape features such as roads, transmission lines, and oil and gas wells or how lesser prairie-chickens may respond to a change in these landscape features.

Several recent studies have examined impacts of energy development on prairie grouse species (Tympanachus and Centrocercus spp.) inhabiting rangelands with a high potential for wind, geothermal, and natural gas energy development (Hagen 2010, Jarnevich and Laubhan 2011, Naugle et al. 2011). Many of these studies demonstrate avoidance of anthropogenic structures and human disturbance that leads to habitat loss and fragmentation (Holloran 2005, Pitman et al. 2005, Walker et al. 2007, Doherty et al. 2008, Pruett et al. 2009). For example, Naugle et al. (2011) reviewed 7 studies to examine the impact of energy development on greater sage-grouse (C. urophasianus). Each study reported negative responses by greater sage-grouse to energy development. Responses included a decrease in lek attendance within or near gas fields and an avoidance of development by nesting hens. Hagen (2010) conducted a meta-analysis of published and unpublished reports pertaining to prairie grouse and the impacts of energy development. He reported a general displacement of grouse by anthropogenic features and reduced demographic rates (e.g., nest success) from energy development. In contrast, Winder et al. (2013) documented an increase in adult survival for greater prairie-chicken hens after construction of a wind energy facility in Kansas and the authors attributed the increased survival to reduced predation risk.

Spatially explicit models allow researchers to associate landscape features with animal occurrence, abundance, or density (Hedley and Buckland 2004, Royle et al. 2004). Identifying suitable habitat and predicting species occurrence or density is especially useful when balancing energy development and the needs of species of conservation concern, such as lesser prairie-chickens (Jarnevich and Laubhan 2011). Models that incorporate presence-only data from opportunistic sampling have been used for prairie grouse but are susceptible to problems associated with incomplete detectability of individuals and non-random samples, which can result in misleading relationships between species occurrence and environmental characteristics (Elith et al. 2011, Royle et al. 2012, Welsh et al. 2013). Hierarchical distance-sampling models provide a tractable way to account for incomplete detection and relate landscape features to density (Royle et al. 2004, Sillett et al. 2012, Blank 2013).

Given the plan for energy development in west Texas where declining lesser prairie-chicken populations occur (ERCOT 2006), managers need a better understanding of lesser prairie-chicken distribution in relation to changes caused by energy development. This information will be useful for future conservation planning efforts for land protection and policy decisions and to guide energy development decisions. In addition, McRoberts et al. (2011) identified a need for more effective monitoring of lesser prairie-chicken populations given their conservation status. Therefore, our objectives were to estimate lek density in the lesser prairie-chicken occupied range of Texas and examine the relationship of lek density with anthropogenic and vegetative landscape characteristics.

STUDY AREA

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

The occupied range of lesser prairie-chickens in Texas, as delineated by Davis et al. (2008), lies mostly in the northeast and southwest regions of the Texas Panhandle, with a few leks thought to be scattered throughout the central portion of the Panhandle. Lesser prairie-chickens occupied 34 counties in Texas before 1980, but their range has declined to only 12 counties in the Panhandle (Davis et al. 2008). Our sampling area encompassed these 12 counties or 86.9% of the occupied lesser prairie-chicken range in Texas (i.e., we excluded portions that were not lesser prairie-chicken habitat such as riparian woodlands and cotton fields). Our sampling area intersected 2 of the CREZs in Texas.

The northeast region of the study area was comprised of a mixed-grass prairie dominated by sand sagebrush (Artemisia filifolia) and little bluestem (Schizachyrium scoparium). The southwest region of the study area was a short-grass prairie dominated by shinnery oak (Quercus havardii) and little bluestem with some mesquite (Prosopis glandulosa). Cotton, winter wheat, and grain sorghum were the main crops grown in both regions (United States Department of Agriculture [USDA] 2008). The climate was mostly dry and the majority of precipitation occurred during the fall and spring (PRISM Climate Group 2011). The southwest region of the Panhandle received an average of 40–51 cm of precipitation yearly and the northeast region received an average of 50–61 cm of precipitation yearly (PRISM Climate Group 2011).

METHODS

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

We used a stratified random sampling design to estimate lesser prairie-chicken lek density (Thompson et al. 1998). We defined 4 strata based on combinations of vegetation types believed to influence lesser prairie-chicken density (e.g., grassland, shrubland, agriculture, and a mosaic of the 3 land cover types; Crawford and Bolen 1976, Taylor and Guthery 1980, Hagen et al. 2004) and areas with the potential for energy development within lesser prairie-chicken-occupied habitat (Timmer et al. 2013). We delineated vegetation types based on the USDA Texas cropland data layer (USDA 2008), which classified grassland as patches comprised of >80% grass and shrubland as patches comprised of shrubs <5 m tall and ≥20% of the total vegetation. We divided the sampling area into 329, 7.2-km × 7.2-km survey blocks; at this size, we could complete 1 survey block per morning. Of the 329 survey blocks covering the sampling area, we excluded 44 blocks as potential lesser prairie-chicken habitat because they were mostly urban, open water, cotton fields, or woodland.

We used ArcGIS 9.3 (Environmental Systems Research Institute, Inc., Redlands, CA) to create a set of grid cells (7.2 km × 7.2 km) over the extent of the occupied lesser prairie-chicken range in Texas. We re-classified the Texas cropland data layer (USDA 2008) into 8 categories (i.e., cotton, grains, other crops, grassland or idle pasture, shrubland, woodland, open water, and barren or developed areas) and calculated the area of grassland, shrubland, and grain in each survey block. We combined vegetation area for each grid cell with a CREZ designation (i.e., cell was either fully or partially in a CREZ or not at all). We assigned survey blocks to 1 of 4 strata and randomly selected blocks from each stratum (strata defined below). We divided our sampling area into 2 regions for the 2 field seasons. During spring 2010, we surveyed blocks in the northeast and central regions of the Panhandle (hereafter, northeast region) and during spring 2011, we surveyed blocks in the southwest and west-central regions (hereafter, southwest region).

The first stratum was composed of survey blocks that were within a CREZ and ≥50% grassland (i.e., native grassland, Conservation Reserve Program land [CRP], or idle cropland). The second stratum was composed of survey blocks that were also ≥50% grassland, but not within a CREZ. The third stratum was composed of survey blocks with >50% shrubland. The fourth stratum was composed of survey blocks with a ≥75% combination of grassland, shrubland, and grain field (this mosaic was comprised of 30–50% grassland, ≤50% shrubland, and >0% grain field). The specifications for this stratum were meant to include potential lesser prairie-chicken habitat while excluding non-habitat, such as urban areas, water bodies, cotton fields, and woodland regions (e.g., riparian cottonwood [Populus deltoides] galleries). None of the blocks with vegetative characteristics of the third stratum were within a CREZ and only 1 block in the fourth stratum was located within a CREZ.

We allocated samples to each stratum using the following formula

  • display math

where ui is the number of survey blocks allocated to each stratum i, U is the total number of survey blocks (n = 180) allocated for the 2-year study and gi is the weighting factor for each stratum i. We calculated the weighting factor as

  • display math

where ri was the rank for each stratum i. We ranked the strata from 1 to 4 with 4 representing the highest priority stratum (i.e., survey blocks within a CREZ and ≥50% grassland). Because our research was part of a larger project designed to examine lek density in areas subject to wind energy development, we prioritized the strata based on the potential for wind energy development to impact lek distribution.

Based on the weighting factors, we randomly selected 72 survey blocks in the first stratum, 54 in the second stratum, 36 in the third stratum, and 18 in the fourth stratum. We were able to survey more blocks than originally planned in the second year, but there were no additional blocks in stratum 1 in the southwest region. Therefore, we randomly selected additional samples in stratum 2. With the additional blocks, we surveyed 76 survey blocks in the first stratum, 73 in the second stratum, 39 in the third stratum, and 20 in the fourth stratum (Timmer et al. 2013).

We used ArcGIS 9.3 to generate a flight path for each survey block and measure the nearest distance from each lek detection to a transect (Hiby and Krishna 2001). We oriented transects north-south with 400-m spacing between them following the survey protocol used by McRoberts et al. (2011). The observer's global positioning system (GPS) unit recorded a track log of each flight path to provide the actual transect lengths that were surveyed. We set the track logs to record points at least every 2 seconds.

We conducted our surveys from an R-22 helicopter (Robinson Helicopter Co., Torrance, CA), which seated an observer and the pilot, who also served as an observer. To train technicians, we also conducted flights early in each field season from an R-44 helicopter (Robinson Helicopter Co.). We conducted flights between early March and late May 2010–2011 and surveyed from sunrise until approximately 2.5 hours post-sunrise. We surveyed at a target altitude of 15 m above ground level and a target speed of 60 km/hr (McRoberts et al. 2011). We did not include portions of transects that were surveyed outside the set survey protocol (e.g., when the pilot increased the helicopter's altitude to avoid houses or feedlots) in the final transect lengths. When we detected lesser prairie-chickens, the pilot deviated from transect and flew over the center of the group of birds or the center of the location from where birds flushed. We used a GPS unit to record the exact location of detected lesser prairie-chickens (Marques et al. 2006). We classified a detection as a lek if ≥1 displaying male was detected.

Data Analysis

We selected 10 vegetative and anthropogenic variables that could influence lek density based on previous literature and our research objectives (Crawford and Bolen 1976, Taylor and Guthery 1980, Fuhlendorf et al. 2002, Hagen et al. 2004, Pitman et al. 2005; Table 1). We divided each survey block into 4, 12.96-km2 quadrats and summarized landscape variables for each quadrat. We developed 3 a priori model sets (Table 2). Our vegetation model set included percent grassland (i.e., native grassland, CRP, or idle cropland), percent shrubland (i.e., shrubs <5 m tall), total percentage of grassland and shrubland, percent grain field (e.g., corn, winter wheat, or grain sorghum), and edge density of all patches (km/km2; USDA 2008). We included a quadratic term for percent grassland and percent shrubland because previous literature has suggested that optimum lesser prairie-chicken habitat consists of native grassland interspersed with shrubland (Copelin 1963, Taylor and Guthery 1980, Applegate and Riley 1998). We also included an interaction between percent grassland and region and between percent shrubland and region. The northeast and southwest regions are composed of different vegetation types (i.e., shinnery oak-dominated type in the southwest region and sand sagebrush with bunchgrasses in the northeast region; Lyons et al. 2009) and the distinct lesser prairie-chicken populations (Corman 2011) may respond differently to the landscape in these 2 regions. Our road model set included paved road density (km/km2), unpaved road density (km/km2), and all road density (km/km2; U.S. Environmental Protection Agency 1998, Texas Department of Transportation 2011). Our energy model set included density of transmission lines ≥69 kv (km/km2; Platts 2011) and active oil and gas well density (wells/km2; Railroad Commission of Texas 2011). We performed a correlation analysis in program R (R Development Core Team 2011) for the landscape variables (Timmer 2012). We did not include variable(s) in the same model that had a pair-wise correlation ≥0.50 to avoid potential problems with multicollinearity (Zar 1999, Ribic and Sample 2001, Graham 2003).

Table 1. Descriptions of the landscape variables we included in hierarchical distance-sampling models of lesser prairie-chicken lek density in Texas, USA, 2010–2011
VariableaDescriptionMeanSDMin.Max.
  • a

    Each variable was calculated for a 12.96-km2 quadrat based on the United States Department of Agriculture (USDA) Texas cropland data layer (USDA 2008).

RegionIndicator variable for the 2 regions of the lesser prairie-chicken range in Texas where northeast region = 0 and southwest region = 1.    
GrassPercent of the quadrat composed of grassland patches (native grassland, Conservation Reserve Program, or idle cropland comprising >80% of the total vegetation) including a quadratic term.0.610.2990.001.00
ShrubPercent of the quadrat composed of shrubland patches (shrubs <5 m tall comprising ≥20% of the total vegetation) including a quadratic term.0.250.2910.001.00
Grass-shrubTotal percentage of grassland and shrubland.0.860.1730.211.00
GrainPercent of the quadrat composed of grain field patches (e.g., winter wheat, corn, or grain sorghum).0.090.1360.000.73
EdgeEdge density for all landcover patches (km/km2).11.306.0860.1027.65
PavedPaved road density (km/km2).0.140.1830.000.84
UnpavedUnpaved road density (km/km2).0.290.2780.001.37
RoadsPaved and unpaved road density (km/km2).0.440.3360.001.68
Trans linesTransmission line (>69 kv) density (km/km2).0.070.1640.001.74
WellActive oil and gas well density (wells/km2).1.252.8340.0023.69
Table 2. Three sets of hierarchical distance-sampling models predicting lesser prairie-chicken lek density in Texas, USA, 2010–2011. For each candidate model, we give −2 × log-likelihood (−2LL), number of parameters (K), Akaike's Information Criterion (AIC), difference in AIC compared to lowest AIC of the model set (Δi), AIC weight (wi), predicted lek abundance (N), and coefficient of variation for abundance (CV)
Modela−2LLKAICΔiwiNCV
  • a

    Grass, shrub, grass-shrub, and grain represent the percentage of the quadrat composed of grassland, shrubland, grassland and shrubland combined, and grain fields, respectively. Region indicates northeast or southwest Texas. Edge, paved, unpaved, roads, well, and trans lines represent the densities of edges, paved roads, unpaved roads, all roads, active oil and gas wells, and transmission lines, respectively.

Vegetation model set
Grass + region + grass × region + grass2 + grass-shrub876.1157890.1150.0000.662237.60.138
Grass + region + grass × region + grass2 + grass-shrub + edge875.8898891.8891.7730.273237.60.135
Shrub + region + shrub × region + shrub2 + grass-shrub881.3767895.3765.2600.048237.30.139
Shrub + region + shrub × region + shrub2 + grass-shrub + edge881.3428897.3427.2270.018237.30.142
Grass + region + grass × region + grass2 + grain900.1307914.13024.014<0.001240.50.139
Grass + region + grass × region + grass2 + grain + edge898.8988914.89824.782<0.001240.50.136
Shrub + region + shrub × region + shrub2 + grain909.8727923.87233.756<0.001237.80.137
Grass-shrub + edge917.5434925.54335.427<0.001254.70.134
Grass-shrub919.6113925.61135.496<0.001253.90.136
Shrub + region + shrub × region + shrub2 + grain + edge909.8028925.80235.686<0.001237.60.140
Grass + region + grass × region + grass2919.5506931.55041.434<0.001242.70.138
Shrub + region + shrub × region + shrub2 + edge917.7907931.79041.675<0.001235.10.134
Shrub + region + shrub × region + shrub2920.8636932.86342.748<0.001240.40.134
Grass + region + grass × region + grass2 + edge918.9357932.93542.820<0.001242.80.141
Grain929.7603935.76045.645<0.001252.20.140
Grain + edge929.3204937.32047.204<0.001252.50.132
Edge940.9163946.91656.801<0.001249.70.137
Road model set
Paved930.4473936.4470.0000.640245.00.139
Paved + unpaved929.9224937.9221.4750.306245.80.137
Roads935.5433941.5435.0960.050249.40.140
Unpaved940.3993946.3999.9530.004250.20.138
Energy model set
Well928.3973934.3970.0000.544250.40.138
Well + trans lines926.7704934.7700.3720.452248.80.138
Trans lines938.3443944.3449.9460.004248.50.132

We analyzed our data using the distsamp function of package unmarked (Fiske and Chandler 2011) in program R, which implements the multinomial-Poisson mixture model (hierarchical distance sampling; Royle et al. 2004). We binned our distance data into 7 intervals (i.e., 0–35 m, 35–50 m, 50–70 m, 70–90 m, 90–120 m, 120–150 m, 150–179 m) according to recommendations by Buckland et al. (2001) and used the half-normal model to describe the detection function (Royle et al. 2004, Sillett et al. 2012, Timmer et al. 2013). We standardized all predictor variables (Fiske and Chandler 2011) and used the 3 a priori model sets (vegetative variables, road variables, and energy infrastructure variables) to model the lek density relationships (Table 2). For all model sets, we included models for each individual variable and the variables combined (Table 2). However, for the vegetation model set, we did not allow percent grassland or percent shrubland to appear together in the same model to reduce the complexity and avoid multicollinearity among the variables. For the model including all road density, we did not include either paved or unpaved road density to avoid multicollinearity.

We determined competitive models as a model with ΔAIC ≤ 2 and excluded models with uninformative parameters (Arnold 2010). We considered the best model(s) from each model set and combined those models (i.e., combined all covariates from the best models) in a final model set along with a null model (Table 3). We selected competitive models from this final model set and evaluated goodness-of-fit of the best models using a Freeman–Tukey chi-squared procedure with 1,000 bootstrap replicates (parboot; Fiske and Chandler 2011). We model averaged among our most competitive models to account for model selection uncertainty (Burnham and Anderson 2002). We model averaged predicted lek density for each 12.96-km2 quadrat covering the lesser prairie-chicken range in Texas and created a map. We summed the model-averaged predicted lek densities within each quadrat to estimate the total number of leks for our sampling area and used the parametric bootstrap procedure with 1,000 bootstrap replicates to estimate uncertainty for estimates of total lek abundance and average density (parboot; Fiske and Chandler 2011).

Table 3. Best overall hierarchical distance-sampling models predicting lesser prairie-chicken lek density in Texas, USA, 2010–2011. For each candidate model, we give −2 × log-likelihood (−2LL), number of parameters (K), Akaike's Information Criterion (AIC), difference in AIC compared to lowest AIC of the model set (Δi), AIC weight (wi), predicted lek abundance (N), and coefficient of variation for abundance (CV)
Modela−2LLKAICΔiwiNCV
  • a

    Grass represents the percentage of the quadrat composed of grassland. Region indicates northeast or southwest Texas. Grass-shrub represents the percentage of the quadrat composed of grassland and shrubland combined. Well and paved represent the densities of active oil and gas wells and paved roads, respectively.

Grass + region + grass × region + grass2 + grass-shrub + well + paved853.7499871.7490.0000.559235.50.136
Grass + region + grass × region + grass2 + grass-shrub + well856.2238872.2230.4740.441237.00.140
Grass + region + grass × region + grass2 + grass-shrub + paved870.1768886.17614.427<0.001235.40.141
Grass + region + grass × region + grass2 + grass-shrub876.1157890.11518.367<0.001237.60.138
Well + paved920.6974928.69756.948<0.001245.60.135
Well928.3973934.39762.649<0.001250.40.138
Paved930.4473936.44764.698<0.001245.00.139
Null940.9262944.92673.177<0.001249.70.143

RESULTS

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

During spring 2010 and 2011, we inventoried 208, 51.84-km2 survey blocks across the lesser prairie-chicken range in Texas (as delineated by Davis et al. 2008). We surveyed 88.6% of the sampling area (10,782.7 of 12,167.1 km2) and detected 96 leks during our surveys. We estimated 109 leks (CV = 0.191, 95% CI = 72–151 leks) in the northeast region and 127 leks (CV = 0.157, 95% CI = 92–170 leks) in the southwest region for a total of 236 leks (CV = 0.138, 95% CI = 177–306 leks) in our sampling area. Estimated lek density ranged from <0.001 leks/12.96 km2 (SE < 0.001) to 0.38 leks/12.96 km2 (SE = 0.09) in the northeast region (mean = 0.16 leks/12.96 km2, CV = 0.156) and from <0.001 leks/12.96 km2 (SE < 0.001) to 1.77 leks/12.96 km2 (SE = 0.47) in the southwest region (mean = 0.27 leks/12.96 km2, CV = 0.160). Mean lek density was 0.21 leks/12.96 km2 (CV = 0.126, 95% CI = 0.16–0.27 leks/12.96 km2) in Texas.

For our vegetation model set, we did not include percentage of grassland and shrubland in the same model or total percentage of grassland and shrubland and percent grain field in the same model because these variables were correlated (r = −0.828, P = < 0.001 and r = −0.854, P = < 0.001, respectively). We found 2 competitive models from our vegetation model set: percent grassland + region + percent grassland × region + percent grassland2 + total percentage of grassland and shrubland (AIC weight [wi] = 0.662) and a similar model that also included edge density (ΔAIC = 1.773, wi = 0.273; Table 2). The model containing edge density was ≤2 ΔAIC units of the top-ranked model and the parameter estimate for edge density did not differ from 0 (β = −0.093, SE = 0.194, P = 0.633); therefore, it was probably an uninformative parameter (Arnold 2010). Lek density increased with an increase in the total percentage of grassland and shrubland (β = 1.172, SE = 0.214, P ≤ 0.001).

We found 2 competitive models from the road model set: paved road density (wi = 0.640) and paved road density + unpaved road density (ΔAIC = 1.475, wi = 0.306; Table 2). Unpaved road density was an uninformative parameter because the model containing it was ≤2 ΔAIC units of the top-ranked model but the parameter estimate did not differ from 0 (β = −0.075, SE = 0.105, P = 0.474). Paved road density was inversely related to lek density (β = − 0.387, SE = 0.129, P = 0.003).

We found 2 competitive models from the energy model set: active oil and gas well density (wi = 0.544) and active oil and gas well density + transmission line density (ΔAIC = 0.372, wi = 0.452; Table 2). Although the model that included transmission line density was ≤2 ΔAIC units of the top-ranked model, the parameter estimate for transmission line density did not differ from 0 (β = −0.156, SE = 0.132, P = 0.238), indicating that the model was likely spurious (Arnold 2010). The best model indicated an inverse relationship between lek density and active oil and gas well density (β = −0.718, SE = 0.279, P = 0.010).

We included the covariates from the top competitive models from the 3 sets of models to fit the final model set (Table 3). Our best model included percent grassland × region + percent grassland2 + total percentage of grassland and shrubland + active oil and gas well density + paved road density (wi = 0.559; Table 3). The second best model was similar but did not include paved road density (ΔAIC = 0.474, wi = 0.441; Table 3). The goodness-of-fit test indicated good model fit for both models (P = 0.512 and P = 0.498, respectively).

We model averaged estimates from these models to produce a map of predicted lek density across the study area (Fig. 1). The quadratic relationship between lek density and percent grassland indicated lek density peaked at 0.39 leks/12.96 km2 (SE = 0.09) and 2.05 leks/12.96 km2 (SE = 0.56) in the northeast and southwest region, respectively when approximately 88% and 44% of a quadrat was composed of grassland patches in the northeast and southwest region, respectively (i.e., paved road density and active oil and gas well density held constant at 0 and total proportion of grassland and shrubland held constant at 100%; Fig. 2). The actual range of landscape composed of grassland was 7.1–99.5% and 0–99.5% in the northeast and southwest region, respectively. The predictor variables used in the final models exhibited relationships with lek density similar to the ones described above (Table 4). A positive relationship between lek density and total proportion of grassland and shrubland was indicated (Fig. 3). Both paved road density (Fig. 4) and active oil and gas well density (Fig. 5) exhibited inverse relationships with lek density.

image

Figure 1. Predicted lesser prairie-chicken (LEPC) lek density for 12.96-km2 quadrats covering the occupied lesser prairie-chicken range in Texas, 2010–2011, based on the 2 most competitive hierarchical distance-sampling models. We classified white areas inside the occupied range as non-lesser prairie-chicken habitat and did not include them in the sampling area.

Download figure to PowerPoint

image

Figure 2. Predicted lesser prairie-chicken lek density for the northeast and southwest regions of the Texas Panhandle in response to the percent of the landscape composed of grassland patches, 2010–2011, based on the 2 most competitive hierarchical distance-sampling models. We held paved road density constant at 0, oil and gas well density constant at 0, and total percentage of grassland and shrubland at 100%.

Download figure to PowerPoint

Table 4. Parameter estimates (β), standard errors (SE), and P-values for the 2 most competitive hierarchical distance-sampling models predicting lesser prairie-chicken lek density in Texas, USA, 2010–2011
VariableaModel 1bModel 2c
βSEPβSEP
  • a

    Grass represents the percentage of the quadrat composed of grassland. Region indicates northeast or southwest Texas. Grass-shrub represents the percentage of the quadrat composed of grassland and shrubland combined. Well and paved represent the densities of active oil and gas wells and paved roads, respectively. All variable values were standardized (i.e., [observation–mean]/SD).

  • b

    Our most competitive model of lek density was grass + region + grass × region + grass2 + grass-shrub + well + paved.

  • c

    Our second most competitive model of lek density was grass + region + grass × region + grass2 + grass-shrub + well.

Intercept−5.5530.378<0.001−5.5480.378<0.001
Grass1.1690.4190.0051.1400.4190.006
Region2.0120.407<0.0012.0150.406<0.001
Grass2−0.6290.137<0.001−0.6190.138<0.001
Grass-shrub1.1890.209<0.0011.2500.209<0.001
Well−0.9340.3210.004−0.9810.3190.002
Paved−0.1960.1300.131   
Grass × region−1.8520.489<0.001−1.8230.490<0.001
image

Figure 3. Predicted lesser prairie-chicken lek density for the northeast and southwest regions of the Texas Panhandle in response to total percentage of grassland and shrubland, 2010–2011, based on the 2 most competitive hierarchical distance-sampling models. We held paved road density constant at 0, active oil and gas well density constant at 0, and percentage grassland constant at 50%.

Download figure to PowerPoint

image

Figure 4. Predicted lesser prairie-chicken lek density for the northeast and southwest regions of the Texas Panhandle in response to paved road density, 2010–2011, based on the 2 most competitive hierarchical distance-sampling models. We held active oil and gas well density constant at 0, percentage of grassland constant at 66%, and total percentage of grassland and shrubland at 100%.

Download figure to PowerPoint

image

Figure 5. Predicted lesser prairie-chicken lek density for the northeast and southwest regions of the Texas Panhandle in response to active oil and gas well density, 2010–2011, based on the 2 most competitive hierarchical distance-sampling models. We held paved road density constant at 0, percentage of grassland constant at 66%, and total percentage of grassland and shrubland at 100%.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

Though both grassland and shrubland patches (i.e., contiguous areas of native grassland and low-growing shrubs, respectively) are recognized as integral components of lesser prairie-chicken habitat, the proportion of each type needed on the landscape is an important consideration. Researchers have debated the optimal proportion of grassland and shrubland patches in lesser prairie-chicken habitat because of sparse documentation of their historical habitat and seasonal variation in habitat use (Taylor and Guthery 1980, Silvy 2006). For example, one habitat guideline recommends landscapes (i.e., >100 km2) of approximately 80% native grassland and 20% shrubland to support lesser prairie-chicken populations (Bidwell 2003). Declining lesser prairie-chicken populations in New Mexico, Oklahoma, and Texas were associated with landscapes containing less shrubland cover compared to populations that were not declining (Woodward et al. 2001), whereas percent grassland or percent grassland and CRP were important predictors of lesser prairie-chicken lek occurrence in Kansas (Jarnevich and Laubhan 2011). Our models indicate the optimal mix of grassland and shrubland patches for lesser prairie-chicken lek density varies by region within Texas. However, habitat structure and composition at smaller scales than our landscape metrics are likely important as well (i.e., vegetation structure and composition within a patch; Hagen et al. 2004).

Given that the northeast and southwest regions of the Texas Panhandle represent 2 distinct populations of lesser prairie-chickens (Corman 2011) and the vegetation composition and structure differ between the regions (Lyons et al. 2009), it is not surprising that lek density varies according to the landscape in these 2 regions (Fig. 2). Lek density is 1 index of prairie grouse population trends (Cannon and Knopf 1981), but other studies have reached a similar conclusion for different lesser prairie-chicken population parameters. Lyons et al. (2009) documented greater survival rates for lesser prairie-chickens in the northeast region of the Texas Panhandle compared to the southwest region and attributed the difference in survival rate to vegetation differences for nesting and brood-rearing. Lyons et al. (2009) concluded that the shinnery oak monoculture in the southwest region may not contain the insect density and residual cover of the sand sagebrush grassland in the northeast region. In contrast, the survival rates reported by Grisham (2012) for the southwest region are among the highest in the literature for hens during the breeding season. Grisham (2012) partly credited the high survival rate to shinnery oak-dominated areas, which provide an important winter food source (i.e., acorns), as well as escape and nesting cover. Patten et al. (2005) examined adult survival for lesser prairie-chickens in southeastern New Mexico and northwestern Oklahoma, which are similar in vegetation composition to the southwest and northeast region of the Texas Panhandle, respectively. Adult survivorship did not differ between their 2 study sites.

Our models predicted higher lek densities in the southwest region compared to the northeast region. In the northeast region, lek density peaked when approximately 88% of the landscape was composed of grassland patches and lek density peaked at approximately 44% grassland in the southwest region (Fig. 2). Further, because lek density increased with an increase in the total percentage of grassland and shrubland (Fig. 3), our models suggested that lek density would be maximized if the remaining proportion of the landscape was composed of shrubland patches (i.e., approximately 12% and 56% shrubland in the northeast and southwest region, respectively). Therefore, any land use activity that alters or reduces the total percent of grassland and shrubland on the landscape (e.g., cultivation) will likely reduce lek density.

Our models indicated anthropogenic disturbances (i.e., paved roads and oil and gas wells) negatively affected lek density (Figs. 4 and 5). Anthropogenic features, such as roads, can fragment contiguous rangeland and result in habitat loss due to avoidance of these features by lesser prairie-chickens (Pruett et al. 2009, Hagen et al. 2011). Other studies have observed a similar relationship between paved roads and prairie grouse. Lesser prairie-chicken nests in Kansas were located farther than expected from paved and high-traffic graveled roads even though otherwise-suitable habitat surrounded those features (Pitman et al. 2005). Pruett et al. (2009) concluded that highways do not appear to impede lesser prairie-chicken movement, but noise and traffic associated with highways may render surrounding habitat unsuitable. Niche modeling of lesser prairie-chicken leks in Kansas showed an increase in lek habitat suitability with increasing distance from highways (Jarnevich and Laubhan 2011) and a separate study in Kansas documented an avoidance of paved roads by radio-marked hens (Hagen et al. 2011). An avoidance of high road densities at a 5-km scale was a significant predictor of greater prairie-chicken (T. cupido) lek locations in Kansas (Gregory et al. 2011), and Hagen (2010) found that prairie grouse displacement by anthropogenic features in several studies was greatest for transmission lines and roads.

In contrast, we did not identify a significant relationship between lek density and transmission line density, which is contrary to recent data. For instance, Hagen et al. (2011) documented avoidance of transmission lines by lesser prairie-chickens in Kansas and lesser prairie-chicken lek suitability increased with increasing distance from transmission lines (Jarnevich and Laubhan 2011). However, in 77.9% of our survey blocks, transmission line density was zero (<1.7 km/km2 in the others) but is expected to increase with wind energy development.

Lesser prairie-chickens will use oil or gas pads as lek sites if the associated activity or traffic is minimal (Jamison et al. 2002), but they also exhibit avoidance of oil and gas activity. In a southwestern Kansas study, lesser prairie-chicken nests were located farther than expected from oil and gas wellheads, although this distance was only significant in 1 study area, possibly because of differences in topography and size or noise levels of pump jacks (Pitman et al. 2005). Distance from oil or gas wells was the most influential anthropogenic feature affecting lek occurrence (for lek locations recorded after 1995) in Kansas and oil or gas well density was the most influential feature affecting lek occurrence at the largest scale (i.e., 3,000 m; Jarnevich and Laubhan 2011).

Research has not documented direct effects on lesser prairie-chicken productivity from oil and gas activity (Hagen et al. 2011), but evidence suggests that productivity of greater sage-grouse, a closely related species, may be directly and indirectly affected (e.g., Lyon and Anderson 2003, Holloran 2005, Walker et al. 2007). In a natural gas field development region in western Wyoming, greater sage-grouse hens nested farther from leks within 3 km of a natural gas well pad or road than more distant leks (Lyon and Anderson 2003). The authors attributed this behavior to an avoidance of vehicular traffic associated with the gas wells. In a separate study in northwestern Wyoming, Holloran (2005) found that nesting sage-grouse hens avoided areas with a high density of active natural gas wells, the number of males displaying at leks decreased with increasing gas field-related disturbances around leks, and leks surrounded by gas field development had low juvenile male recruitment and high displacement of adult males. Walker et al. (2007) observed a greater decline in male sage-grouse lek attendance and a decline in number of active leks in areas near coal-bed natural gas development. Doherty et al. (2008) found that sage-grouse hens avoided coal-bed natural gas development surrounded by otherwise-suitable winter habitat.

The hierarchical modeling technique we used is different than the techniques used in previous studies examining lek occurrence and landscape features. We established a formal sampling design that provided spatial coverage of our sampling area and used probabilistic sampling for the occupied lesser prairie-chicken range in Texas. We accounted for incomplete detection of leks by modeling a detection function and were thus, able to extrapolate relationships between lek density and our predictive variables to the entire lesser prairie-chicken range in Texas (Royle et al. 2004). In contrast, previous modeling efforts to predict greater and lesser prairie-chicken lek occurrence and describe relationships with landscape characteristics have not been based on formal statistical designs, which can introduce potential biases (e.g., Gregory et al. 2011, Jarnevich and Laubhan 2011). However, conclusions from our study are limited to lek density within the lesser prairie-chicken range in Texas. A similar modeling effort across the lesser prairie-chicken range could inform regional habitat-priority maps, which are currently lacking (Hagen 2010). This would account for inherent variability throughout the lesser prairie-chicken range due to variation in anthropogenic activity, grazing intensity, fire frequency, soil types, local weather, and a suite of other factors and thus, improve local management efforts within a regional framework. Further, modeling demographic parameters such as nest success with change in habitat composition or anthropogenic features over time could improve our ability to adaptively manage lesser prairie-chicken populations in a dynamic landscape.

MANAGEMENT IMPLICATIONS

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

Balancing grassland and shrubland on the landscape will promote greater lesser prairie-chicken lek densities in Texas. Our results suggested that in the northeast region, lek density will be maximized on landscapes composed of 88% grassland and 12% shrubland, whereas in the southwest region, greatest lek density will likely occur on landscapes composed of 44% grassland and 56% shrubland. This can be achieved through habitat management techniques, such as prescribed burns or herbicide treatments to control trees or shrubs and promote native grass recruitment (Applegate and Riley 1998, Patten et al. 2005). We observed a negative relationship between lek density and oil and gas well density in the Texas Panhandle, which is consistent with observed negative impacts of anthropogenic development and activities on lesser prairie-chickens and other prairie grouse. Therefore, increasing oil and gas development in the occupied range could be a significant factor in the long-term conservation of lesser prairie-chicken populations in Texas. Given that most of our lek detections and greatest predicted lek density estimates occurred in Gray, Hemphill, and Lipscomb counties in the northeast Panhandle, and in Bailey, Cochran, and Yoakum counties in the southwest Panhandle (Fig. 1), careful consideration of oil and gas well placement and the construction of additional paved roads in these areas is warranted to reduce potential negative impacts on lesser prairie-chickens.

ACKNOWLEDGMENTS

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED

We thank J. Bonner and R. Martin from TPWD for organizing and leading TPWD's contribution to the aerial surveys in 2010 and 2011, respectively. We also thank TPWD for sharing the transmission line and oil and gas data for our modeling efforts. We thank J. Ashling, J. Leal, J. R. Leal, and K. Pyle for assisting with the aerial surveys. We also could not have completed our aerial surveys successfully or safely without the skill and guidance of our pilots, A. Wheatly, D. Wooten, A. Lange, M. Huggins, K. Lange, R. Norris, and T. Webb. This project was funded by a U.S. Department of Energy grant and additional financial contribution from TPWD and Texas Tech University. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the U.S. Fish and Wildlife Service. The use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. This is Texas Tech University, College of Agricultural Science and Natural Resources technical publication T-9-1232.

LITERATURE CITED

  1. Top of page
  2. ABSTRACT
  3. STUDY AREA
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. MANAGEMENT IMPLICATIONS
  8. ACKNOWLEDGMENTS
  9. LITERATURE CITED
  • American Wind Energy Association. 2012. U. S. Wind Industry: fourth quarter 2011 market report. American Wind Energy Association, Washington, D.C., USA.
  • Applegate, R. D., and T. Z. Riley. 1998. Lesser prairie-chicken management. Rangelands 20:1315.
  • Arnold, T. W. 2010. Uninformative parameters and model selection using Akaike's Information Criterion. Journal of Wildlife Management 74:11751178.
  • Bidwell, T. G. 2003. Ecology and management of the lesser prairie-chicken in Oklahoma. Oklahoma Cooperative Extension Service, E-970. Oklahoma State University, Stillwater, USA.
  • Blank, P. J. 2013. Northern bobwhite response to Conservation Reserve Program habitat and landscape attributes. Journal of Wildlife Management 77:6874.
  • Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001. Introduction to distance sampling: estimating abundance of biological populations. Oxford University Press, Oxford, United Kingdom.
  • Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second edition. Springer, New York, New York, USA.
  • Cannon, R. W., and F. L. Knopf. 1981. Lek numbers as a trend index to prairie grouse populations. Journal of Wildlife Management 45:776778.
  • Copelin, F. F. 1963. The lesser prairie chicken in Oklahoma. Technical Bulletin No. 6. Oklahoma Wildlife Conservation Department, Oklahoma City, USA.
  • Corman, K. S. 2011. Conservation and landscape genetics of Texas lesser prairie-chicken: population structure and differentiation, genetic variability, and effective size. Thesis, Texas A&M University-Kingsville Kingsville, USA.
  • Crawford, J. A., and E. G. Bolen 1976. Effects of land use on lesser prairie chickens in Texas. Journal of Wildlife Management 40:96104.
  • Davis, D. M., R. E. Horton, E. A. Odell, R. D. Rogers, and H. A. Whitlaw. 2008. Lesser prairie-chicken conservation initiative. Unpublished Report. Lesser prairie chicken interstate working group. Colorado Division of Wildlife, Fort Collins, USA.
  • Doherty, K. E., D. E. Naugle, B. L. Walker, and J. M. Graham. 2008. Greater sage-grouse winter habitat selection and energy development. Journal of Wildlife Management 72:187195.
  • Electric Reliability Council of Texas. 2006. Analysis of transmission alternatives for competitive renewable energy zones in Texas. Electric Reliability Council of Texas, Austin, USA.
  • Elith, J., S. J. Phillips, T. Hastie, M. Dudik, Y. E. Chee, and C. J. Yates. 2011. A statistical explanation of MaxEnt for ecologists. Diversity and Distributions 17:4357.
  • Fiske, I. J., and R. B. Chandler. 2011. Unmarked: an R package for fitting hierarchical models of wildlife occurrence and abundance. Journal of Statistical Software 43:123.
  • Fuhlendorf, S. D., A. J. Woodward, D. M. Leslie, Jr., and J. S. Shackford. 2002. Multi-scale effects of habitat loss and fragmentation on lesser prairie-chicken populations of the US southern Great Plains. Landscape Ecology 17:617628.
  • Graham, M. H. 2003. Confronting multicollinearity in ecological multiple regression. Ecology 84:28092815.
  • Gregory, A. J., L. B. McNew, T. J. Prebyl, B. K. Sandercock, and S. M. Wisely. 2011. Hierarchical modeling of lek habitats of greater prairie-chickens. Pages 2132 in B. K. Sandercock, K. Martin, and G. Segelbacher, editors. Ecology, conservation, and management of grouse. Studies in Avian Biology. University of California Press, Berkeley, USA.
  • Grisham, B. A. 2012. The ecology of lesser prairie-chickens in shinnery oak-grassland communities in New Mexico and Texas with implications toward habitat management and future climate change. Dissertation, Texas Tech University, Lubbock, USA.
  • Hagen, C. A., B. E. Jamison, K. M. Giesen, and T. Z. Riley. 2004. Guidelines for managing lesser prairie-chicken populations and their habitats. Wildlife Society Bulletin 32:6982.
  • Hagen, C. A. 2010. Impacts of energy development on prairie grouse ecology: a research synthesis. Transactions of the North American Wildlife and Natural Resources Conference 75:96103.
  • Hagen, C. A., J. C. Pitman, T. M. Loughin, B. K. Sandercock, R. J. Robel, and R. D. Applegate. 2011. Impacts of anthropogenic features on habitat use by lesser prairie-chickens. Pages 6375 in B. K. Sandercock, K. Martin, and G. Segelbacher, editors. Ecology, conservation, and management of grouse. Studies in Avian Biology. University of California Press, Berkeley, USA.
  • Hedley, S. L., and S. T. Buckland. 2004. Spatial models for line transect sampling. Journal of Agricultural, Biological, and Environmental Statistics 9:181199.
  • Hiby, L., and M. B. Krishna. 2001. Line transect sampling from a curving path. Biometrics 57:727731.
  • Holloran, M. J. 2005. Greater sage-grouse (Centrocercusurophasianus) population response to natural gas field development in western Wyoming. Dissertation, University of Wyoming, Laramie, USA.
  • Jackson, A. S., and R. DeArment. 1963. The lesser prairie chicken in the Texas Panhandle. Journal of Wildlife Management 27:733737.
  • Jamison, B. E., J. A. Dechant, D. H. Johnson, L. D. Igl, C. M. Goldade, and B. R. Euliss. 2002. Effects of management practices on grassland birds: lesser prairie-chicken. Northern Prairie Wildlife Research Center, Jamestown, North Dakota, USA.
  • Jarnevich, C. S., and M. K. Laubhan. 2011. Balancing energy development and conservation: a method utilizing species distribution models. Environmental Management 47:926936.
  • Lyon, A. G., and S. H. Anderson. 2003. Potential gas development impacts on sage grouse nest initiation and movement. Wildlife Society Bulletin 31:486491.
  • Lyons, E. K., B. A. Collier, N. J. Silvy, R. R. Lopez, B. E. Toole, R. S. Jones, and S. J. DeMaso. 2009. Breeding and non-breeding survival of lesser prairie-chickens Tympanuchus pallidicinctus in Texas, USA. Wildlife Biology 15:8996.
  • Marques, T. A., M. Andersen, S. Christensen-Dalsgaard, S. Belikov, A. Boltunov, Ø. Wiig, S. T. Buckland, and J. Aars. 2006. The use of global positioning systems to record distances in a helicopter line-transect survey. Wildlife Society Bulletin 34:759763.
  • McRoberts, J. T., M. J. Butler, W. B. Ballard, H. A. Whitlaw, D. A. Haukos, and M. C. Wallace. 2011. Detectability of lesser prairie-chicken leks: a comparison of surveys from aircraft. Journal of Wildlife Management 75:771778.
  • Naugle, D. E., K. E. Doherty, B. L. Walker, M. J. Holloran, and H. E. Copeland. 2011. Energy development and greater sage-grouse. Pages 489503 in S. T. Knicks, and J. W. Connelly, editors. Greater sage-grouse: ecology and conservation of a landscape species and its habitats. Studies in Avian Biology. University of California-Press, Berkeley, USA.
  • Patten, M. A., D. H. Wolfe, E. Shochat, and S. K. Sherrod. 2005. Effects of microhabitat and microclimate selection on adult survivorship of the lesser prairie-chicken. Journal of Wildlife Management 69:12701278.
  • Pitman, J. C., C. A. Hagen, R. J. Robel, T. M. Loughin, and R. D. Applegate. 2005. Location and success of lesser prairie-chicken nests in relation to vegetation and human disturbance. Journal of Wildlife Management 69:12591269.
  • Platts. 2011. Electric transmission lines. Platts, Westminster, Colorado, USA.
  • PRISM Climate Group. 2011. PRISM homepage. <http://www.prism.oregonstate.edu>. Accessed September 4, 2011.
  • Pruett, C. L., M. A. Patten, and D. H. Wolfe. 2009. Avoidance behavior by prairie grouse: implications for development of wind energy. Conservation Biology 23:12531259.
  • Railroad Commission of Texas. 2011. Oil and gas well records. <http://www.rrc.state.tx.us/data/wells/wellrecords.php>. Accessed October 1, 2011.
  • Railroad Commission of Texas. 2012. Data and statistics. Online research queries. <http://www.rrc.state.tx.us/data/index.php>. Accessed December 28, 2012.
  • R Development Core Team. 2011. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: <http://www.R-project.org>. Accessed August 22, 2011.
  • Ribic, C. A., and D. W. Sample. 2001. Associations of grassland birds with landscape factors in southern Wisconsin. American Midland Naturalist 146:105121.
  • Royle, J. A., D. K. Dawson, and S. Bates. 2004. Modeling abundance effects in distance sampling. Ecology 85:15911597.
  • Royle, J. A., R. B. Chandler, C. Yackulic, and J. D. Nichols. 2012. Likelihood analysis of species occurrence probability from presence-only data for modelling species distributions. Methods in Ecology and Evolution 3:545554.
  • Sillett, T. S., R. B. Chandler, J. A. Royle, M. Kéry, and S. A. Morrison. 2012. Hierarchical distance-sampling models to estimate population size and habitat-specific abundance of an island endemic. Ecological Applications 22:19972006.
  • Silvy, N. J. 2006. In my opinion: shinnery oak is not a requirement for lesser prairie chicken habitat. Pages 138143 in J. W. Cain III, and P. R. Krausman, editors. Managing wildlife in the Southwest: new challenges for the 21st Century. Southwest Section of The Wildlife Society, Tucson, Arizona, USA.
  • Sullivan, R. M., J. P. Hughes, and J. E. Lionberger. 2000. Review of the historical and present status of the lesser prairie-chicken (Tympanuchus pallidicinctus) in Texas. The Prairie Naturalist 32:177188.
  • Taylor, M. A., and F. S. Guthery. 1980. Status, ecology, and management of the lesser prairie chicken. General Technical Report RM-77. U.S. Department of Agriculture, Forest Service Rocky Mountain Forest, and Range Experimental Station, Fort Collins, Colorado, USA.
  • Texas Department of Transportation. 2011. TxDOT_Roadways. Transportation Planning and Program Division, Data Analysis Branch, Texas Department of Transportation, Austin, USA.
  • Thompson, W. L., G. C. White, and C. Gowan. 1998. Monitoring vertebrate populations. Academic Press, San Diego, California, USA.
  • Timmer, J. M. 2012. Abundance and density of lesser prairie-chickens and leks in Texas, Thesis, Texas Tech University, Lubbock, USA.
  • Timmer, J. M., M. J. Butler, W. B. Ballard, C. W. Boal, and H. A. Whitlaw. 2013. Abundance and density of lesser prairie-chickens and leks in Texas. Wildlife Society Bulletin 37:in press. DOI:10.1002/wsb.304
  • U.S. Department of Agriculture. 2008. USDA, National Agricultural Statistics Service, 2008 Texas cropland data layer. 2008 edition. U. S. Department of Agriculture, National Agricultural Statistics Service, Marketing and Information Services Office, Washington, D. C., USA.
  • U.S. Environmental Protection Agency. 1998. DOT/FHA major roads for CONUS, Alaska, and Hawaii in BASINS. Office of Water, Washington, D.C., USA.
  • U.S. Fish and Wildlife Service. 2012. Candidate status review. Federal Register 77:7382873888.
  • Walker, B. L., D. E. Naugle, and K. E. Doherty. 2007. Greater sage-grouse population response to energy development and habitat loss. Journal of Wildlife Management 71:26442654.
  • Welsh, A. H., D. B. Lindenmayer, and C. F. Donnelly. 2013. Fitting and interpreting occupancy models. PLoS ONE 8(1) e52015. DOI: 10.1371/journal.pone.0052015
  • Winder, V. L., L. B. McNew, A. J. Gregory, L. M. Hunt, and S. M. Wisely. 2013. Effects of wind energy development on survival of female greater prairie-chickens. Journal of Applied Ecology 50:in press. DOI:101111/1365-2664.12184
  • Woodward, A. J., S. D. Fuhlendorf, D. M. Leslie, Jr., and J. Shackford. 2001. Influence of landscape composition and change on lesser prairie-chicken (Tympanuchus pallidicinctus) populations. American Midland Naturalist 145:261274.
  • Zar, J. H. 1999. Biostatistical analysis. Second edition. Prentice Hall, Upper Saddle River, New Jersey, USA.