Annual population changes of most grouse, including the imperiled Attwater's prairie-chicken Tympanuchus cupido attwateri, are driven by annual reproductive success. Previous research identified poor survival of chicks as a primary bottleneck for recovery of this species. We evaluated the relative importance of 26 factors in 5 categories (weather and topography, habitat, plant phenology, time and site, hen characteristics) on Attwater's prairie-chicken brood survival to 2 wk posthatch (the period when chick mortality is highest) and on the number of chicks per brood at 6 wk posthatch (when chicks are capable of independent survival). Factors with most support for predicting brood survival to 2 wk included invertebrate dry mass, ordinal date, an index to maximum photosynthetic activity of vegetation from multispectral imagery, and proportion of brood locations within areas treated to suppress red imported fire ants Solenopsis invicta. Broods were most likely to survive if they hatched between early and late May and were located within areas 1) that were treated to suppress red imported fire ants, 2) where vegetation produced intermediate values for the maximum photosynthetic activity index, and 3) that supported high invertebrate biomass. The number of chicks per brood surviving to 6 wk posthatch was best predicted by a nonlinear relationship with a drought index during the first 2 wk posthatch, and it was maximized when average values of the drought index indicated moderately depleted soil moisture, but not severe drought. Our finding that the average drought index during the first 2 wk after hatch had more support for predicting the number of chicks per brood at 6 wk than did the average drought index for the entire 6 wk emphasizes the importance of the first 2 wk for Attwater's prairie-chickens. This comprehensive analysis of factors affecting Attwater's prairie-chicken brood survival provides valuable information to guide management and recovery efforts for this species.

The Attwater's prairie-chicken Tympanuchus cupido attwateri Bendire is endemic to grasslands of southeast Texas and southwest Louisiana (Bendire 1894; Lehmann 1941). It shares its subspecies status with the greater prairie-chicken T. c. pinnatus and the extinct heath hen T. c. cupido (Johnson et al. 2020; Gill et al. 2021). Once abundant on expansive coastal prairie grasslands (Lehmann 1941, 1968; Lehmann and Mauermann 1963), the Attwater's prairie-chicken was listed as endangered with extinction in 1967 pursuant to the U.S. Endangered Species Preservation Act of 1966 (Udall 1967). Habitat conversion to cropland and residential areas and degradation of remaining grasslands by overgrazing, woody species encroachment, and invasion by exotic fauna and flora have driven this subspecies to near extinction (Lehmann 1941; Jurries 1979; U.S. Fish and Wildlife Service [USFWS] 2010; Morrow et al. 2015). Populations have remained below 200 individuals since 1993 despite intensive intervention, including habitat management and release of captive-reared individuals to supplement failing populations (USFWS 2010, 2021; Morrow et al. 2015). The Attwater's prairie-chicken is currently listed as endangered pursuant to the U.S. Endangered Species Act (ESA 1973, as amended; USFWS 2021).

The Attwater's prairie-chicken recovery plan pinpointed poor survival of chicks in the wild as “…the single-most factor limiting significant progress toward recovery” (USFWS 2010:40). Morrow et al. (2015) identified availability of invertebrate food as influenced by the invasive red imported fire ant Solenopsis invicta as a major limiting factor for survival of Attwater's prairie-chicken broods. However, other factors including habitat composition and structure (e.g., Lehmann 1941; Jones 1963; Kessler 1978; Svedarsky 1979), weather (e.g., Lehmann 1941; Jurries 1979; Flanders-Wanner et al. 2004), time of year (Riley et al. 1998; Fields et al. 2006; Matthews et al. 2011), or characteristics of the brood hen such as age or captive-reared versus wild (Moss et al. 1975; Fields et al. 2006; McNew et al. 2012; Rymesova et al. 2013) have been suggested as factors affecting brood survival for Galliformes.

Removing poor chick survival as a bottleneck to Attwater's prairie-chicken recovery requires a thorough understanding of which factors most limit existing populations. Hannon and Martin (2006) identified two critical periods for survival of juvenile grouse: 1) the first 2 wk after hatch when chicks are dependent on the hen for thermoregulation, habitat selection, and protection from predators; and 2) when independent young are dispersing. Hatch-year Attwater's prairie-chickens become capable of independent survival and dispersal from the brood hen at approximately 6 wk of age (Lehmann 1941). Therefore, we evaluated the relative importance of hen attributes, weather, topography, ordinal day, site, and habitat characteristics, including plant phenology, vegetation structure, fire ant suppression, and invertebrate abundance on Attwater's prairie-chicken brood survival to 2 wk posthatch and on the number of chicks per brood alive at 6 wk posthatch.

We conducted our study at the Attwater Prairie Chicken National Wildlife Refuge and on private land in Goliad County, Texas (Figure 1). The 4,265-ha Attwater Prairie Chicken National Wildlife Refuge (APCNWR; Colorado County; 29.7°N, 96.3°W) near Eagle Lake, Texas, is part of the USFWS National Wildlife Refuge System and was established specifically to maintain habitat for Attwater's prairie-chickens (USFWS 2012). The Goliad County study area (28.6°N, 97.3°W) near Goliad, Texas, was located on a 2,670-ha private cattle ranch situated within approximately 20,445 ha of relatively contiguous grasslands (USFWS 2010). The two study areas were located approximately 150 km apart. Both sites were within the gulf prairies and marshes vegetational area of Texas (Hatch et al. 1990) and consisted of managed open grasslands containing mid–tall native grass species. Climate of the region was subtropical, and dominated by warm, moist air masses derived from the Gulf of Mexico (Smeins et al. 1991). Total annual precipitation for Colorado County, Texas, averaged 1,057 mm during 1960–1990, and average daily temperatures ranged from 10.8°C in winter to 27.9°C in summer (Brown 2006). Total annual precipitation for Goliad County, Texas, averaged 1,016 mm from 1971 to 2000, and average daily temperatures ranged from 13°C in winter to 28.9°C in summer (Wiedenfeld 2010). We collected data on Attwater's prairie-chicken broods during 2009–2019 at APCNWR and 2009–2012 at the Goliad County site. Annual precipitation during those periods averaged 1,076 (standard deviation [SD] = 329) mm at the Eagle Lake Agricultural Research Station 11 km southwest of APCNWR and 739 (SD = 314) mm at Coleto Creek Reservoir 18 km northeast of the Goliad County site (Wilson et al. 2015).

Figure 1.

Location of the Attwater Prairie Chicken National Wildlife Refuge (APC NWR) and Goliad County, Texas, study sites within the historical range of the Attwater's prairie-chicken Tympanuchus cupido attwateri as delineated by Lehmann (1941).

Figure 1.

Location of the Attwater Prairie Chicken National Wildlife Refuge (APC NWR) and Goliad County, Texas, study sites within the historical range of the Attwater's prairie-chicken Tympanuchus cupido attwateri as delineated by Lehmann (1941).

Close modal

Cattle stocking rates at both study sites generally averaged 4.8–10.1 ha/animal unit year, and both sites used prescribed burning as a management tool. Prescribed fires at the Goliad County site were accomplished predominantly as whole pasture burns with pre- and postfire grazing deferral, whereas APCNWR implemented patch burns within pastures (Fuhlendorf and Engle 2001; Fuhlendorf et al. 2006) with and without grazing deferral after prescribed fire. Depending on funding availability, managers treated portions of both sites aerially with 1.7 kg/ha of Extinguish Plus™ (0.365% hydramethylnon, 0.25% s-methoprene; Central Life Sciences, Schaumburg, IL) fire ant bait to suppress fire ants and increase abundance of invertebrates required by prairie-chicken broods for food (Morrow et al. 2015). Area and timing of treatment varied somewhat for each site and among years. For the Goliad County site, managers applied repeat treatments of Extinguish Plus™ in 2010 (November) and 2011 (September) on 294 ha of potential prairie-chicken brood habitat. At APCNWR, treatment area size increased from 308 ha in 2009 to a maximum of 2,381 ha in 2014–2015. During 2016–2019, treatment area size at APCNWR remained constant at 2,052 ha. The APCNWR treatments (1·site−1·year−1) occurred in autumn (September–November) or spring (March–early April). Although we targeted prairie-chicken core use areas for fire ant treatment as indicated by long-term telemetry data and lek locations, we did not focus treatments on specific habitats (e.g., brood or nesting habitat).

Brood survival

We evaluated the relationship between Attwater's prairie-chicken brood survival and five categories of variables: 1) weather and topography, 2) habitat, 3) plant phenology, 4) time and site, and 5) hen characteristics (Data S1, Data S2, Table S1 [Supplemental Material]). We equipped hens with poncho-mounted radiotransmitters (Amstrup 1980; Toepfer 2003) with tuned-loop (Telemetry Solutions, Walnut Creek, CA; Advanced Telemetry Systems, Isanti, MN) or whip antennas (American Wildlife Enterprises, Monticello, FL). We coiled or trimmed whip antennas to extend only approximately 6 cm beyond the poncho to avoid potential interference with flight (Marks and Marks 1987). Most (n = 124; Table S1) Attwater's prairie-chicken females at both study locations were released from various captive rearing facilities (USFWS 2010). Although some (n = 53) were after-hatch-year, most (n = 83) brood hens were released as 8–12+ wk-old poults during July–October after spending 2 wk in prerelease acclimation pens at release sites (Table S1). Therefore, releases occurred roughly 5–8 mo in advance of their first reproductive season (March–July) in the wild. We equipped hens with transmitters (12–24 g, ≤ 3% of body mass; expected battery life > 365 d) at the time of transfer from captive rearing facilities. We recaptured survivors by night-lighting in subsequent years as needed to replace transmitters nearing the end of their expected battery life. We placed transmitters on wild-hatched hens (n = 14; Table S1) at ≥ 7 wk of age after homing on their radioed mother at night.

We tracked broods (hen with chicks) using telemetry approximately 10 times (i.e., daily during the 5-d work week) during the first 2 wk posthatch, except when adverse weather hindered obtaining these data or hen movements (> 1.6 km) indicated probable brood loss. We determined brood hen locations (hereafter, brood location) either by triangulation using a vehicle-mounted 6-element Yagi antenna system and computer software (DogTrack, Blacksburg, VA; Locate III, Tatamagouche, NS, Canada), or by circling brood hens at 10–20 m with a hand-held telemetry system and recording location data with a global positioning system (GPS) unit. Radiomarking of hens and brood monitoring activities were authorized by Federal Fish and Wildlife Permit TE051839 and Texas Parks and Wildlife Department Scientific Research Permit SPR-0491-384.

We assessed whether broods survived to 2 wk posthatch in one of two ways: 1) homing on radioed brood hens at dawn and visually checking for the presence of chicks, and 2) playing chick distress calls near brood hens. Through 2015, we approached radioed hens with hand-held telemetry equipment at 14 d posthatch and observed them at dawn before they left night roosts to determine whether any chicks remained with the hen. We encouraged hens to move only if necessary to be absolutely certain of chick presence. We considered a brood successful if we observed ≥ 1 chick with the hen. To reduce disturbance to the hen and brood, we did not attempt to obtain a total count of chicks at the 2-wk assessment. During 2016–2019, we approached within approximately 20–30 m of brood hens during daylight hours and played an Attwater's prairie-chicken chick distress call recorded at the Houston Zoo, Inc. Visible and/or vocal response by the brood hen indicated positive presence of chicks (Healy et al. 1980). If, based on results of playing the chick distress call, we were uncertain whether chicks remained with the brood hen, we followed up by checking at dawn as previously described. At approximately 6–8 wk posthatch, we assessed brood survival again, this time with effort made to count all chicks. We approached brood hens at night using telemetry and head-mounted spotlights, and we counted all surviving chicks. If counts of surviving chicks were assessed more than once, we used the maximum number of chicks observed on any one occasion for our analyses.

Weather and topography variables

We characterized weather metrics for each brood by determining average temperature (°C; “temp”), total rain (cm; “rain”), number of days with rain (“days.rain”), total wind (km/day; “wind”), total pan evaporation (cm; “evap”), and average Keetch–Byram Drought Index (KBDI; Keetch and Byram 1968) over the first 2 and 6 wk posthatch (Tables S1, S2). The Keetch–Byram Drought Index ranges from 0 to 800 and estimates moisture deficits in the deep duff or upper soil layer. Each KBDI unit represents 0.25 mm of moisture deficit. Temperature, precipitation, wind, and evaporation data were downloaded for the Eagle Lake Agricultural Research Station (APCNWR broods; Colorado County; 29.6°N, 96.4°W) and Coleto Creek Reservoir (Goliad County broods; 28.7°N, 97.2°W) weather stations (Wilson et al. 2015). Daily KBDIs were calculated for each site from daily precipitation and maximum temperature (Alexander 1990).

We calculated a Topographic Position Index (TPI) as an index to flooding potential at brood sites during the first 2 wk posthatch based on light detection and ranging (LiDAR) data (Guisan et al. 1999; Tables S1, S2). Data collection for LiDAR occurred in June 2009 and January–February 2011 for Goliad County and APCNWR, respectively, with approximately 4 returns/m2 (Strategic Mapping Program 2009, 2011). We processed the LiDAR data to produce a 3-m Digital Elevation Model, and then calculated the TPI using a 200-m-radius neighborhood (Guisan et al. 1999). We used the 2-wk mean TPI value for each brood in subsequent analyses to evaluate whether topographic position was related to brood survival. We also hypothesized that variation in TPI values might indicate the availability of potential refugia from flooding at the microhabitat level. Therefore, we also evaluated the relationship between brood survival and standard deviation (SD) of TPI (“TPI.SD”).

Habitat variables

We collected habitat data (Tables S1, S2) at brood sites 1–4 d after determining their locations to avoid disturbance to the broods. We collected invertebrate samples with 25 vigorous sweeps of vegetation along an approximately 20-m transect in a random direction from the predetermined brood location with a 38-cm diameter, canvas sweep-net. Each sweep consisted of a single approximately 2-m arc of the net. We did not collect samples when vegetation was wet. We labeled invertebrate samples and froze them until we could determine counts of individuals for each sample. Invertebrates were dried at 60°C for 24 h and then weighed to the nearest 0.0001 g on a digital analytical scale. We determined median invertebrate numbers (“num.inverts”) and mass (“dry.mass”) for each brood to control for pseudoreplication within brood unit. We also included the derived metric “mean.dry.mass” representing total dried mass (g) of invertebrates from samples divided by the count of invertebrates. We hypothesized that a mean mass per invertebrate > 1 mg and < 10 mg would represent prey sizes most available to foraging chicks.

We determined vegetation effective height (cm) using a 1.2 × 1.2-m white pegboard with holes spaced at 2.54 cm. We placed the pegboard at each predetermined brood location, and at four locations 10 m away in each cardinal direction (Kobriger 1965; Toepfer 2003). We took digital photographs at each of the five locations from a distance of 4 m and height of 1 m (Robel et al. 1970). We then imported photographs into ArcGIS Desktop ArcMap (version 10.6.1 and earlier; ESRI 2018) and the height of vegetation that completely obscured all dots below it was determined at 10 equidistant (12.7 cm) points on the board, using the top of the board as a known distance to scale measurements. We used all 50 points (10 points/photograph, 5 photographs/brood site) to determine the mean and coefficient of variation of effective height at each brood location. Finally, we determined mean effective height (cm; “veg.ht”) and coefficient of variation (CV; “veg.ht.cv”) determined for each brood for the 2-wk posthatch period. We included the CV of effective vegetation height to assess the importance of variability in vegetation structure to brood survival in addition to height.

We also included metrics of vegetation height (cm) obtained from a LiDAR canopy height model (Tables S1, S2). We first processed the LiDAR data to provide an estimate of vegetation height above ground level using FUSION software v. 2.80 (McGaughey 2017). We then summarized this layer at multiple scales, using neighborhood analyses to calculate the mean and standard deviation of vegetation height at 9- and 21-m radii from each pixel. We calculated mean values for brood locations taken during the first 2 wk posthatch only; we did not determine specific brood locations after that time. These LiDAR-derived mean values provided information on the extent of vegetation along the vertical axis only. The field-collected effective height measurements we described previously were similar to visual obstruction measurements described by Robel et al. (1970), which were strongly correlated with the biomass of vegetation present in Kansas grasslands. These two vegetation height metrics provided information on different aspects of vegetation structure at Attwater's prairie-chicken brood sites.

Plant phenology, time and site, and hen variables

Remote sensing of land-surface plant phenology can characterize vegetation changes during the growing season and document specific events such as start of the growing season and its duration. These vegetation changes may in turn affect habitat characteristics (e.g., timing and density of invertebrate abundance, thermal cover, shelter from predators). We included metrics of plant phenology as derived from time-series Collection 6 Aqua EROS Moderate Resolution Imaging Spectroradiometer (eMODIS) Normalized Difference Vegetation Index (NDVI) data recorded at 7- to 10-d intervals and a 250-m grid-cell size (Jenkerson et al. 2010). These metrics (Tables S1, S2) consisted of five annual phenological layers derived from NDVI values recorded over the course of the year (Jenkerson et al. 2010): 1) start of season (value at the beginning of measurable photosynthesis; “sosn”), 2) end of season (value at the end of measurable photosynthesis; “eosn”), 3) maximum increase in NDVI above the baseline level (“amp”), 4) maximum level of NDVI (“maxn”), and 5) sum of NDVI (“tin”).

We classified broods according to year (“year”), study site (“site”), and ordinal date of hatch within year (“day”) to evaluate the relative importance of these variables on brood and chick survival (Tables S1, S2). We classified brood hens by age (“age”; second year [SY] or after second year [ASY]), source (“source”; released from captivity or wild-hatched), years in the wild (“years.out”), whether they had successfully nested (“nest.prev”) or fledged (“fledge.prev”; chicks survived to minimum of 6 wk) chicks in previous years, and the total estimated fresh weight (g; “total.egg.mass”) of eggs in the clutch as an index of energy invested by the hen. We weighed all eggs in a clutch (nearest 0.1 g) and measured them (nearest 0.1 mm) with digital calipers, in most cases when hens were off their nests during morning or evening feeding forays. We used these data to estimate fresh egg weights as described by Hoyt (1979) and Burnham (1983). Table S1 provides summaries of data collected on broods and hens in our study.

Data analyses

We used a two-stage approach for analyzing both 2-wk brood survival and the number of chicks per brood at 6 wk. For both sets of analyses, we first fit models for each of the five predictor categories separately (stage one; Table S2, Supplemental Material). Models with support in each stage one category compared with the null model were then combined into additional models (stage two) using the ‘dredge' function from the ‘MuMIn' package (Barton 2020) to produce all possible combinations, provided the variables in these models were not highly correlated (|r| < 0.6). For each model, we examined β values for significance (P ≤ 0.05) and removed models that had uninformative parameters (Arnold 2010). We took this modeling approach because missing values would have necessitated the removal of 16% of the observations had all data been initially analyzed together. By separating the analyses by predictor category, and then combining only terms with support in the final analysis, we were able to minimize the removal of observations with missing data. To aid in model convergence, all variables were scaled by subtracting the mean and dividing by the standard deviation.

Brood survival to 2 weeks.

We modeled survival to 2 wk using generalized linear models (GLM). The response variable was presence or absence of at least one chick at 2 wk posthatch. Although most hens (86%) had only one observation in the data set, there were some hens with more than one observation because they were observed across multiple years: 16 hens with 2 broods and 1 hen with 3. There may have been some lack of independence among observations of the same hen, so we initially ran the models as mixed-effects models with hen as a random effect to account for any pseudoreplication. However, these models had difficulty converging and the estimated variation of the random effect was zero. These models did not support the inclusion of hen as a random effect, so we instead used GLMs for the 2-wk brood survival analysis.

We built GLMs using the ‘glm' function in the Program R statistical system (R version 3.6.3, R Core Team 2020) and specified a binomial distribution with a logit link. We identified top models using change in Akaike's Information Criterion values (ΔAIC; Akaike 1973) and associated Akaike weights (w; Burnham and Anderson 2002). For the final (stage two) combined model analysis we considered the top models to be within 2 ΔAIC. To assess fit for the top model, we report the Hosmer and Lemeshow goodness-of-fit test (Hosmer and Lemeshow 2000). To evaluate performance, we calculated the accuracy of the top model predictions (% of broods with correctly predicted outcomes). We also calculated the Area Under a Receiver Operating Characteristic Curve (AUC; Hanley and McNeil 1982) for the top model. The AUC represents the probability of giving a brood that survived a higher probability of survival than a brood that did not survive. An AUC value of 0.5 indicates the model performs no better than expected by chance whereas a value of 1 indicates perfect predictive ability.

Number of chicks per brood at 6 weeks.

For analysis of chicks per brood at 6 wk, we followed the same procedure as the brood survival analysis at 2 wk with the exception that 6-wk weather metrics were added to candidate models, and the response variable, number of chicks per brood at 6 wk, was modeled with a negative binomial distribution using the ‘glm.nb' function from the ‘MASS' package (Venables and Ripley 2002) in R. We did not determine detailed locations of broods and did not collect habitat metrics after 2 wk posthatch. However, we included habitat metrics collected from 0 to 2 wk in our candidate models because most chick mortality typically occurs during this timeframe (Hannon and Martin 2006), and events during this time period could influence the number of chicks per brood at 6 wk. We used a negative binomial rather than a Poisson model because the response variable, number of chicks, was overdispersed. To evaluate model fit, we used a chi-square test to compare the best supported model with the same model fitted with a Poisson distribution to verify that use of the negative binomial distribution was supported by the data. We also evaluated goodness-of-fit by comparing the observed deviance with the expected deviance under a chi-square distribution.

We monitored 138 (APCNWR n = 115; Goliad n = 23) broods from 120 hens between 2009 and 2019 (Table S1). Of these, 58 (42.0%) survived to 2 wk. Nine (15.5%) brood hens died during weeks 2–6, and we were unable to locate 4 (6.9%) brood hens at 6 wk. Of the 45 surviving hens that we relocated at 6 wk posthatch, 13 (28.9%) had no chicks, and the remaining 32 (71.1%) had ≥ 1 chicks ( = 2.6, maximum = 8). Brood location coordinate data were available for 108 of the 120 hens during the first 2 wk posthatch. Location data from hens were available for all years at APCNWR, but only for 2009 and 2012 at Goliad. The mean number of locations per hen during the first 2 wk after hatch was 7.4 (SD = 3.7, range = 1–18) with 801 locations recorded for all hens combined.

Analysis of brood survival to 2 weeks

There was support for some candidate models in all categories except hen traits (Table 1). For stage one analyses, the best time and site model for 2-wk brood survival contained a quadratic term for day of year (“day2”; w = 0.53). All time and site models with support compared with the null model contained “day” or “day2” (ΔAICc ≤ 4.31; cumulative w = 0.84); neither “year” nor “site” were competitive for predicting 2-wk brood survival. The best stage one weather and topography model contained the 2-wk mean value for KBDI along with a quadratic of mean temperature (“temp2”) over the 2 wk following hatching (w = 0.20; Table 1). The top three (ΔAICc ≤ 2.33; cumulative w = 0.41) weather and topography models contained some form of KBDI or “temp.” The best supported stage one habitat model was a quadratic relationship with median dry mass of invertebrates along with a term for percent of brood locations within fire ant–treated areas (w = 0.51; Table 1). The second best stage one habitat model (ΔAICc = 0.93) also contained quadratic terms for “dry.mass” along with a binary variable indicating whether broods hatched in fire ant–treated areas (“rifa.trt.hatch”). Cumulative weight for these two models was 0.83. The only other models for habitat that were competitive compared with the null model contained variations of these parameters. For the plant phenology metrics, there was some support for a quadratic brood survival relationship with both maximum level of photosynthetic activity (“maxn” + “maxn2”; ΔAICc = 0.00; w = 0.22) in the canopy and level of photosynthetic activity at the beginning of measurable photosynthesis (“sosn” + “sosn2”; ΔAICc = 0.43; w = 0.18; Table 1). No hen variables had support when compared with the null model, nor did any vegetation structure metrics (Table 1).

Table 1.

Two-stage model selection results for predicting Attwater's prairie-chicken Tympanuchus cupido attwateri 2-wk brood survival between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). Results for top models and the null are presented here. K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.

Two-stage model selection results for predicting Attwater's prairie-chicken Tympanuchus cupido attwateri 2-wk brood survival between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). Results for top models and the null are presented here. K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.
Two-stage model selection results for predicting Attwater's prairie-chicken Tympanuchus cupido attwateri 2-wk brood survival between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). Results for top models and the null are presented here. K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.

The best supported stage two model for predicting brood survival to 2 wk posthatch contained day of year (“day2”), percent of brood locations within fire ant–treated areas (“rifa.trt”), median invertebrate dry mass (“dry.mass”), and maximum level of photosynthetic activity (“maxn” + “maxn2”) indicated by the NDVI (w = 0.42; Table 1; Figure 2). The two top stage two models with ΔAICc < 2 both contained the same “dry.mass,” “day,” and “maxn” terms, and both contained terms for fire ant treatment (“rifa.trt,” “rifa.trt.hatch”). These top stage two models accounted for 0.60 of cumulative model weight (Table 1). Of all the variables in the top model, support was strongest (P < 0.001) for median dry mass of invertebrates at brood sites (Table 2). Predicted brood survival increased from 0.38 (95% CI = 0.21–0.58) to 0.96 (95% CI = 0.83–0.99) when the median dry invertebrate biomass at brood locations increased from 0.12 to 1.36 g. Similarly, predicted survival increased from 0.36 (95% CI = 0.20–0.55) to 0.64 (95% CI = 0.48–0.77) for broods with locations taken entirely outside of fire ant–treated areas versus those with locations taken entirely within those areas, respectively (Figure 2). Model diagnostics did not uncover any issues with the final model. The Hosmer and Lemeshow goodness-of-fit test indicated adequate model fit (P = 0.42). The model predicted 74% of observations correctly and the AUC was 0.79, indicating the model had a moderate amount of predictive power.

Figure 2.

Relationship between date of hatch, percent of brood locations within areas treated to suppress invasive red imported fire ants Solenopsis invicta, median dry mass of invertebrates at brood sites, and maximum Normalized Difference Vegetation Index (NDVI) on Attwater's prairie-chicken Tympanuchus cupido attwateri brood survival from 0 to 2 wk posthatch between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranches in Goliad County, Texas. Shaded areas represent 95% confidence intervals.

Figure 2.

Relationship between date of hatch, percent of brood locations within areas treated to suppress invasive red imported fire ants Solenopsis invicta, median dry mass of invertebrates at brood sites, and maximum Normalized Difference Vegetation Index (NDVI) on Attwater's prairie-chicken Tympanuchus cupido attwateri brood survival from 0 to 2 wk posthatch between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranches in Goliad County, Texas. Shaded areas represent 95% confidence intervals.

Close modal
Table 2.

Coefficients for the best-supported combined models of 2-wk brood survival and number of chicks per brood at 6 wk for Attwater's prairie-chickens Tympanuchus cupido attwateri between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables were standardized by subtracting the mean and dividing by the standard deviation.

Coefficients for the best-supported combined models of 2-wk brood survival and number of chicks per brood at 6 wk for Attwater's prairie-chickens Tympanuchus cupido attwateri between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables were standardized by subtracting the mean and dividing by the standard deviation.
Coefficients for the best-supported combined models of 2-wk brood survival and number of chicks per brood at 6 wk for Attwater's prairie-chickens Tympanuchus cupido attwateri between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables were standardized by subtracting the mean and dividing by the standard deviation.

Analysis of number of chicks at 6 weeks

Only three variables were supported in stage one models for predicting number of chicks per brood at 6 wk: average KBDI2 during the first 2 and 6 wk after hatch, and median dry mass2 of invertebrates collected at brood sites during the first 2 wk posthatch (Table 3). Overall, the mean KBDI2 from hatch to 2 wk had the most support in combined stage two models for predicting the number of chicks per brood at 6 wk (w = 0.90; Table 3). No other predictor variables were found to be informative in the final stage two model. The highest number of chicks were predicted when KBDI values ranged from 200 to 400 (Figure 3), indicating conditions that were neither excessively wet nor dry. Model diagnostics revealed no issues with model fit. The negative binomial model, which estimated the dispersion parameter, was more appropriate (P < 0.001) than the Poisson model. Model diagnostics indicated an adequate fit of the model to the data (P = 0.36).

Table 3.

Two-stage model selection results for predicting the number of Attwater's prairie-chicken Tympanuchus cupido attwateri chicks per brood between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best-supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.

Two-stage model selection results for predicting the number of Attwater's prairie-chicken Tympanuchus cupido attwateri chicks per brood between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best-supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.
Two-stage model selection results for predicting the number of Attwater's prairie-chicken Tympanuchus cupido attwateri chicks per brood between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best-supported models in each Type category (stage one) were combined for evaluation in subsequent stage two analyses (Type = Combined). Candidate models and variables are described in Table S2 (Supplemental Material). K = number of parameters estimated, ΔAICc = change in Akaike's Information Criterion corrected for small sample sizes, and w = model weight.
Figure 3.

Relationship between Keetch–Byram Drought Index (KBDI; Keetch and Byram 1968) during the first 2 wk after hatching and the number of Attwater's prairie-chicken Tympanuchus cupido attwateri chicks per brood at 6 wk posthatch between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranches in Goliad County, Texas. Shaded areas represent 95% confidence intervals. The KBDI estimates soil moisture depletion and ranges from 0 (fully saturated soil) to 800 (maximum depletion).

Figure 3.

Relationship between Keetch–Byram Drought Index (KBDI; Keetch and Byram 1968) during the first 2 wk after hatching and the number of Attwater's prairie-chicken Tympanuchus cupido attwateri chicks per brood at 6 wk posthatch between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranches in Goliad County, Texas. Shaded areas represent 95% confidence intervals. The KBDI estimates soil moisture depletion and ranges from 0 (fully saturated soil) to 800 (maximum depletion).

Close modal

We observed 42.0% survival for Attwater's prairie-chicken broods to 2 wk posthatch during our 11-y study. In comparison, McNew et al. (2011) observed 47–54% (n = 15) survival of Kansas greater prairie-chicken broods to 14 d posthatch, and Matthews et al. (2011) observed 50% survival of Nebraska greater prairie-chicken broods (n = 36) to 10 d (38% when extrapolated to 14 d assuming constant daily survival). Pratt (2010) reported 69% survival for 83 Minnesota greater prairie-chicken broods to 2 wk posthatch. Not including 4 (2.9%) broods of unknown fate, ≥ 23.9% of brood units in our study survived to 6 wk compared with 28–38% reported by others for lesser T. pallidicinctus, greater, and Attwater's prairie-chickens (Fields et al. 2006; Pratt 2010).

Posthatch survival of broods represents a critical stage in the life cycle of grouse and is potentially influenced by a myriad of biological and environmental factors (Hannon and Martin 2006; Manzer and Hannon 2008). It is important to evaluate the relative influence of as many of those factors as possible so management strategies may be formulated within an efficient and logical framework where feasible. We evaluated 20 environmental factors and 6 characteristics of brood hens on Attwater's prairie-chicken brood survival (Tables 13). Broods were most likely to survive the first 2 wk posthatch if they hatched between early and late May, were located within areas treated to suppress red imported fire ants where vegetation produced intermediate values for maximum NDVI, and that supported high invertebrate biomass (Figure 2). Numerous studies have identified the importance of invertebrate abundance to survival of prairie-chicken broods (e.g., Lehmann 1941; Jones 1963; Hagen et al. 2005; Morrow et al. 2015). Morrow et al. (2015) identified invertebrate abundance as a limiting factor specifically for Attwater's prairie-chicken brood survival, and in turn implicated red imported fire ants in limiting invertebrate abundance. However, that study did not explore the relative importance of other factors (e.g., habitat, weather, hen characteristics) that may influence brood survival.

Surprisingly, field-collected metrics related to habitat structure (“veg.ht,” “veg.ht.cv”) were not supported as predictive of Attwater's prairie-chicken brood survival within the range of habitat conditions we observed (Tables 1, S2). This suggests that other resources or environmental conditions (i.e., invertebrate abundance, ordinal date, fire ant management) were more limiting to Attwater's prairie-chicken broods than was habitat structure in our study. These findings are consistent with those of Matthews et al. (2011) and Flanders-Wanner et al. (2004), who also failed to find support for habitat variables in explaining daily survival rates of greater prairie-chickens broods or sharp-tailed grouse Tympanuchus phasianellus production indices, respectively. Lutz and Silvy (1980) suggested that for nesting Attwater's prairie-chickens, vegetation characteristics beyond critical minimum thresholds may have little influence on susceptibility of nests to predation. Our data, along with those of Flanders-Wanner et al. (2004) and Matthews et al. (2011) suggest the same principle may apply to prairie grouse brood habitat as well.

However, it is also possible that we did not collect field data on habitat variables most pertinent to brood survival. Maximum NDVI, collected remotely by satellite sensors, was included in our final model for predicting 2-wk brood survival. The NDVI is derived from canopy reflectance of the red and near-infrared wavebands, and serves as an indicator of canopy structure, green biomass, nitrogen content, and potential photosynthetic activity of vegetation (Gamon et al. 1995). The quadratic relationship for maximum NDVI indicates that 2-wk brood survival was highest at intermediate values (0.05–0.15) and declined sharply on either side of those values (Figure 2). In addition to providing an abundant supply of invertebrate food, prairie-chicken brood habitat must allow for chick movement, provide concealment from predators, and provide shelter from the weather (Lehmann 1941; Kessler 1978; Svedarsky et al. 2003). Thus, to maximize brood survival a balance must be achieved that optimizes provision of food, concealment, and shelter, while facilitating chick movements. The relationship of 2-wk brood survival to maximum NDVI we observed suggests that overhead canopy structure may be more important to brood survival than are the vertical effective vegetation heights we measured in the field at brood sites. It is also possible that maximum NDVI may be related to potential invertebrate abundance supported by green vegetation. However, if that were the case, we would expect the maximum NDVI–2-wk brood survival relationship to become asymptotic as the ability of the habitat to support invertebrate biomass reached saturation levels necessary to maximize survival.

We found little support for Topographic Position Index in predicting Attwater's prairie-chicken brood survival. This may be due to the relative lack of variability in topographic relief for our study areas, which are typical of much of the Attwater's prairie-chicken's historical range. In a Nebraska study, where topographic relief is more prominent than in the coastal prairie habitat of Attwater's prairie-chickens, greater prairie-chicken brood hens showed a strong preference for intermediate topographic positions, but topographic position was not supported for predicting brood survival (Matthews et al. 2011). Topographic position may be relevant to brood survival only when extreme precipitation results in flooding.

Consistent with other studies (e.g., Fields et al. 2006; Matthews et al. 2011), our analyses indicated time of year (ordinal date) was an important predictor of brood survival to 2 wk posthatch (Tables 1, 2; Figure 2). The probability of a brood surviving to 2 wk posthatch peaked when nests hatched in mid-May, and then precipitously declined thereafter (Figure 2). Fields et al. (2006) and Matthews et al. (2011) also observed declining survival for prairie-chicken broods as the season progressed and has been observed for other Galliformes including grey partridge Perdix perdix (Panek 1992) and ring-necked pheasant Phasianus colchicus (Riley et al. 1998). It is likely that time-of-year effects are confounded with the influence of other factors including insect availability, habitat quality, temperature, precipitation patterns, and hen condition (Riley et al. 1998; Flanders-Wanner et al. 2004; Fields et al. 2006; Matthews et al. 2011).

Various aspects of weather are known to influence brood survival for Galliformes (e.g., Lehmann 1941; Moss 1986; Panek 1992; Riley et al. 1998). For example, previous studies have highlighted the importance of rainfall received during the early brooding period on survival of young prairie-chickens, with extremes in both directions being potentially detrimental to survival (e.g., Lehmann 1941; Jurries 1979; Morrow et al. 1996; Schole et al. 2011). Temperature effects have also been reported for survival of galliform broods including gray partridge (Panek 1992), ring-necked pheasants (Riley et al. 1998), and sharp-tailed grouse (Flanders-Wanner et al. 2004). Although our first-stage analyses indicated support for some weather parameters (rain, KBDI, temperature) in predicting 2-wk brood survival, these variables were not supported in combined stage two models (Table 1). Fields et al. (2006) also failed to find support for weather variables in predicting brood survival for greater and lesser prairie-chickens.

The lack of support we observed for weather variables in predicting survival of the brood unit to 2 wk posthatch should not imply that these variables are unimportant to chick survival. We observed the highest number of 6-wk-old chicks per brood when KBDI was at intermediate levels during the first 2 wk after hatch. The Keetch–Byram Drought Index is determined by local rainfall and temperatures (Alexander 1990). Additionally, ordinal date was important in predicting 2-wk brood survival, and it is highly correlated with temperature. Rather, our findings suggest that some chicks within a brood likely perish during extreme values of KBDI during the first 2 wk, resulting in fewer chicks per brood at 6 wk, but food as indicated by invertebrate abundance and fire ant treatment determines whether the brood unit collectively survives to 2 wk.

Even though our candidate models included weather metrics for 0–6 wk posthatch, the number of chicks per brood at 6 wk posthatch was best predicted by the mean KBDI from 0 to 2 wk posthatch (Table 2; Figure 3). This finding emphasizes the importance of the first 2 wk posthatch for galliform chicks (Newell et al. 1987; Panek 1992; Hannon and Martin 2006; Schole et al. 2011). The Keetch–Byram Drought Index ranges from 0 (no moisture deficiency) to 800 (severe drought). Each KBDI unit represents 0.25 mm of soil moisture depletion (Keetch and Byram 1968). Therefore, values of KBDI from 200 to 400, at which our most supported final model predicts maximum chicks per brood (Table 2; Figure 3), represents 50–100 mm of precipitation needed to fully saturate soil. Our findings are consistent with those of Lehmann (1941:33), who concluded 1) rainfall in May is of greater significance than other months because most Attwater's prairie-chicken chicks hatch in May, and 2) production of chicks is highest when May rainfall is approximately 3.8 cm below average. Values for KBDI are inversely related to precipitation received (Keetch and Byram 1968; Alexander 1990). Therefore, our finding of maximized brood survival at moderate KBDI values and that reported by Lehmann (1941) under conditions of slightly below average rainfall likely reflects a “happy medium” for newly hatched chicks, whereby danger is low from both precipitation and flooding, or desiccation due to dry conditions.

Intrinsic differences among brood hens may also contribute to posthatch survival of chicks (e.g., Moss et al. 1981; Fields et al. 2006; Buner et al. 2011). A potential difference among hens in our study of particular interest was whether they had been reared in the wild by Attwater's prairie-chicken hens or had been hand-raised in captivity. Poor breeding success in the wild is a commonly reported malady for captive-bred animals (e.g., Parish and Sotherton 2007; Buner et al. 2011; Rymesova et al. 2013). However, we did not find evidence for any of the hen characteristics we hypothesized might predict Attwater's prairie-chicken brood survival (Tables 1, S2). Small sample sizes for some variables may have limited our ability to detect differences in brood survival among hen traits that we evaluated (Table S1). This was especially the case for captive-reared (n = 124) vs. wild (n = 14) status, but consistent with our findings, Buner et al. (2011:599) concluded regarding captive-bred grey partridge: “…once the released hens successfully hatch chicks, their chick-rearing behavior is normal. It also indicates that despite many generations of captive breeding, released stock with a game farm background maintains its natural breeding potential.” Our observations suggest this may be the case for captive-reared Attwater's prairie-chickens as well. Despite relatively large sample sizes of SY (n = 85) and ASY (n = 53) Attwater's prairie-chicken hens in our data set, hen age class was not competitive in predicting 2-wk brood survival, consistent with observations by McNew et al. (2012) for greater prairie-chickens and Riley et al. (1998) for ring-necked pheasants. In contrast, Fields et al. (2006) found that broods reared by ASY prairie-chickens were 9.8× more likely to survive to 60 d posthatch than were those reared by SY hens, and Hannon and Martin (2006) stated that older female ptarmigan Lagopus spp. raised more chicks to independence than do first- or second-time breeders.

We examined a comprehensive list of variables hypothesized to influence Attwater's prairie-chicken brood survival and identified five that were important in predicting survival through 6 wk posthatch. Invertebrate abundance (dry mass), treatment for red imported fire ants, ordinal date, and maximum NDVI were most important for predicting survival of the brood unit to 2 wk of age. Management actions with demonstrated efficacy for increasing invertebrate abundance include fire ant suppression (Morrow et al. 2015), soil disturbance to encourage forbs that support insects (Jones 1963), and patch burning (Engle et al. 2008). Maintenance of high-quality nesting habitat (e.g., USFWS 2010; Starns et al. 2020) and predator management, including the use of predator-deterrent fences (Morrow and Toepfer 2020) are actions that managers can take to increase success of early nests and mitigate for the reductions in survival we observed for late season broods (Figure 2). Finally, we observed the highest number of chicks at 6 wk posthatch when KBDI values were intermediate (not too dry, not too wet) during the 2 wk after hatch. Although management cannot control precipitation patterns in the short-term, actions can be taken to ensure that runoff efficiently drains from brood habitat. In the long-term, climate predictions for Texas indicate “unprecedented” drought risk resulting from climate change driven by greenhouse gas emissions (Cook et al. 2015). Our data suggest that increases in frequencies of severe droughts will lead to substantially fewer Attwater's prairie-chicken chicks surviving to independence and may further complicate recovery of this species.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Data set used in the analysis of factors affecting Attwater's prairie-chicken Tympanuchus cupido attwateri brood survival from 2009 to 2019 on the Attwater Prairie Chicken National Wildlife Refuge (APCNWR; Colorado County, Texas) and on private ranchlands in Goliad County, Texas.

Available: https://doi.org/10.3996/JFWM-21-054.S1 (55 KB XLSX)

Data S2. Description of variables in Data S1 used in the analysis of factors affecting Attwater's prairie-chicken Tympanuchus cupido attwateri brood survival from 2009 to 2019 on the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and on private ranchlands in Goliad County, Texas.

Available: https://doi.org/10.3996/JFWM-21-054.S2 (26 KB DOCX)

Table S1. Variables hypothesized to affect Attwater's prairie-chicken Tympanuchus cupido attwateri brood survival on the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and on private ranchlands in Goliad County, Texas, 2009–2019. Data were summarized by 1) whether the brood survived the first 2 wk, and 2) whether there were ≥ 1 chicks detected at 6 wk for those broods that survived the first 2 wk. For variables with missing values, the sample size is given in the row labeled n; otherwise, the sample size is given in the header row.

Available: https://doi.org/10.3996/JFWM-21-054.S3 (211 KB DOC)

Table S2. Stage one candidate models for predicting Attwater's prairie-chicken Tympanuchus cupido attwateri 2-wk brood survival and number of chicks per brood at 6 wk posthatch between 2009 and 2019 at the Attwater Prairie Chicken National Wildlife Refuge (Colorado County, Texas) and private ranchlands in Goliad County, Texas. Variables from best supported models in each category were selected for evaluation in subsequent stage two analyses.

Available: https://doi.org/10.3996/JFWM-21-054.S4 (70 KB DOC)

Reference S1. Jenkerson CB, Maiersperger T, Schmidt G. 2010. eMODIS: user-friendly data source. U.S. Geological Survey Open-File Report 2010–1055.

Available: https://doi.org/10.3996/JFWM-21-054.S5 (1.989 MB PDF) and https://pubs.usgs.gov/of/2010/1055/pdf/OF2010-1055.pdf

Reference S2. Jurries RW. 1979. Attwater's prairie chicken. Austin: Texas Parks and Wildlife Department, F.A. Series No. 18.

Available: https://doi.org/10.3996/JFWM-21-054.S6 (2.336 MB PDF)

Reference S3. Keetch JJ, Byram GM. 1968. A drought index for forest fire control. Asheville, North Carolina: U.S. Forest Service Southeastern Forest Experiment Station Research Paper SE-38.

Available: https://doi.org/10.3996/JFWM-21-054.S7 (2.094 MB PDF) and https://www.srs.fs.usda.gov/pubs/rp/rp_se038.pdf

Reference S4. Lutz RS, Silvy NJ. 1980. Predator response to artificial nests in Attwater's prairie chicken habitat. Pages 48–51 in Vohs PA Jr, Knopf FL, editors. Proceedings of the prairie grouse symposium. Stillwater: Oklahoma State University.

Available: https://doi.org/10.3996/JFWM-21-054.S8 (10.725 MB PDF)

Reference S5. McNew LB, Sandercock BK, Pitman JC. 2011. Impacts of alternative grassland management regimes on the population ecology of grassland birds. Project W-67 Second Quarter Report to Kansas Department of Wildlife and Parks.

Available: https://doi.org/10.3996/JFWM-21-054.S9 (113 KB PDF)

Reference S6. Newell JA, Toepfer JE, Rumble MA. 1987. Summer brood-rearing ecology of the greater prairie chicken on the Sheyenne National Grasslands. Pages 24–43 in Bjugstad AJ, technical coordinator. Prairie Chickens on the Sheyenne National Grasslands. Fort Collins, Colorado: U.S. Forest Service Rocky Mountain and Range Experiment Station General Technical Report RM-159.

Available: https://doi.org/10.3996/JFWM-21-054.S10 (103 KB PDF) and https://www.fs.fed.us/rm/pubs_rm/rm_gtr159/rm_gtr159_024_031.pdf

Reference S7. Svedarsky WD, Toepfer JE, Westemeier RL, Robel RJ. 2003. Effects of management practices on grassland birds: greater prairie-chicken. Jamestown, North Dakota: Northern Prairie Wildlife Research Center.

Available: https://doi.org/10.3996/JFWM-21-054.S11 (279 KB PDF) and https://doi.org/10.3133/70159589

Reference S8. Toepfer JE. 2003. Prairie chickens & grasslands: 2000 and beyond. Report to the Council of Chiefs, Society of Tympanuchus Cupido Pinnatus. Elm Grove, Wisconsin.

Available: https://doi.org/10.3996/JFWM-21-054.S12 (21.537 MB PDF)

Reference S9.U.S. Endangered Species Preservation Act of 1966, Pub. L. No. 89-669, 80 Stat. 926 (Oct. 15, 1966).

Available: https://doi.org/10.3996/JFWM-21-054.S13 (1.070 MB PDF) and https://uscode.house.gov/statutes/pl/89/669.pdf

Reference S10.[USFWS] U.S. Fish and Wildlife Service. 2010. Attwater's prairie-chicken recovery plan. 2nd revision. Albuquerque, New Mexico: U.S. Fish and Wildlife Service.

Available: https://doi.org/10.3996/JFWM-21-054.S14 (2.862 MB PDF) and https://ecos.fws.gov/docs/recovery_plan/100426.pdf

Reference S11.[USFWS] U.S. Fish and Wildlife Service. 2012. Attwater Prairie Chicken National Wildlife Refuge comprehensive conservation plan and environmental assessment. Albuquerque, New Mexico: U.S. Fish and Wildlife Service.

Available: https://doi.org/10.3996/JFWM-21-054.S15 (55.165 MB PDF)

Reference S12.[USFWS] U.S. Fish and Wildlife Service. 2021. Attwater's greater prairie-chicken (Tympanuchus cupido attwateri) 5-year review: summary and evaluation. Attwater Prairie Chicken National Wildlife Refuge and Texas Coastal Ecological Services.

Available: https://doi.org/10.3996/JFWM-21-054.S16 (1.286 MB PDF) and https://ecos.fws.gov/docs/tess/species_nonpublish/995.pdf

We thank staff and interns at the Attwater Prairie Chicken National Wildlife Refuge for assistance with data collection, and private landowners for their support and access to their properties. We also gratefully acknowledge the support and wise counsel of the late Dr. J. E. Toepfer. The Society of Tympanuchus Cupido Pinnatus, Ltd., the Minnesota Prairie Chicken Society, the National Fish and Wildlife Foundation, the U.S. Fish and Wildlife Service, and the Nature Conservancy of Texas provided logistical and financial support. Funders had no influence on the content of the manuscript and did not require approval prior to publication. This research was authorized by permits from the U.S. Fish and Wildlife Service (TE051839) and Texas Parks and Wildlife Department (SPR-0491-384). Finally, we appreciate comments and suggestions by the Associate Editor and anonymous reviewers, which greatly improved earlier versions of this paper.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Akaike
H.
1973
.
Information theory as an extension of the maximum likelihood principle. Pages 267–281in
Petrov
BN,
Csaki
F,
editors.
Second international symposium on information theory
.
Budapest
:
Akademiai Kiado
.
Alexander
ME.
1990
.
Computer calculation of the Keetch–Byram drought index—programmers beware
.
Fire Management Notes
51
(4)
:
23
25
.
Amstrup
SC.
1980
.
A radio-collar for game birds
.
Journal of Wildlife Management
44
:
214
217
.
Arnold
TW.
2010
.
Uninformative parameters and model selection using Akaike's Information Criterion
.
Journal of Wildlife Management
74
:
1175
1178
.
Barton
K.
2020
.
MuMIn: Multi-Model Inference. R package version 1.43.17
.
Bendire
CE.
1894
.
Tympanuchus americanus attwateri Bendire. Attwater's or southern prairie hen
.
Auk
11
:
130
132
.
Brown
SE.
2006
.
Soil survey of Colorado County, Texas. U.S. Department of Agriculture, Natural Resources Conservation Service
.
Buner
FD,
Browne
SJ,
Aebischer
NJ.
2011
.
Experimental assessment of release methods for the re-establishment of a red-listed galliform, the grey partridge (Perdix perdix)
.
Biological Conservation
144
:
593
601
.
Burnham
W.
1983
.
Artificial incubation of falcon eggs
.
Journal of Wildlife Management
47
:
158
168
.
Burnham
KP,
Anderson
DR.
2002
.
Model selection and multimodel inference: a practical information-theoretic approach. 2nd edition
.
New York
:
Springer
.
Cook
BI,
Ault
TR,
Smerdon
JE.
2015
.
Unprecedented 21st century drought risk in the American Southwest and Central Plains. Science Advances, 1:e1400082
.
Engle
DM,
Fuhlendorf
SD,
Roper
A,
Leslie
DM
2008
.
Invertebrate community response to a shifting mosaic of habitat
.
Rangeland Ecology and Management
61
:
55
62
.
[ESA] U.S. Endangered Species Act of 1973, as amended, Pub. L. No. 93-205, 87 Stat. 884 (Dec. 28, 1973)
.
ESRI.
2018
.
ArcGIS desktop 10.6.1
.
Redlands, California
:
ESRI
.
Fields
TL,
White
GC,
Gilgert
WC,
Rodgers
RD.
2006
.
Nest and brood survival of lesser prairie-chickens in west central Kansas
.
Journal of Wildlife Management
70
:
931
938
.
Flanders-Wanner
BL,
White
GC,
McDaniel
LL.
2004
.
Weather and prairie grouse: dealing with effects beyond our control
.
Wildlife Society Bulletin
32
:
22
34
.
Fuhlendorf
SD,
Engle
DM.
2001
.
Restoring heterogeneity on rangelands: ecosystem management based on evolutionary grazing patterns
.
BioScience
51
:
625
632
.
Fuhlendorf,
SD,
Harrell
WC,
Engle
DM,
Hamilton
RG,
Davis
CA,
Leslie
DM
2006
.
Should heterogeneity be the basis for conservation? Grassland bird response to fire and grazing
.
Ecological Applications
16
:
1706
1716
.
Gamon
JA,
Field
CB,
Goulden
ML,
Griffin
KL,
Hartley
AE,
Joel
G,
Peñuelas
J,
Valentini
R.
1995
.
Relationships between NDVI, canopy structure, and photosynthesis in three Californian vegetation types
.
Ecological Applications
5
:
28
41
.
Gill
F,
Donsker
D,
Rasmussen
P,
editors.
2021
.
International Ornithological Committee world bird list (v11.2). doi:10.14344/IOC.ML.11.2
.
Available: www.worldbirdnames.org (April 2022)
Guisan
A,
Weiss
SB,
Weiss
AD.
1999
.
GLM versus CCA spatial modeling of plant species distribution
.
Plant Ecology
143
:
107
122
.
Hagen
CA,
Salter
GC,
Pitman
JC,
Robel
RJ,
Applegate
RD.
2005
.
Lesser prairie-chicken brood habitat in sand sagebrush: invertebrate biomass and vegetation
.
Wildlife Society Bulletin
33
:
1080
1091
.
Hanley
JA,
McNeil
BJ.
1982
.
The meaning and use of the area under a receiver operating characteristic (ROC) curve
.
Radiology
143
:
29
36
.
Hannon
SJ,
Martin
K.
2006
.
Ecology of juvenile grouse during the transition to adulthood
.
Journal of Zoology
269
:
422
433
.
Hatch
SL,
Gandhi
KN,
Brown
LE.
1990
.
Checklist of the vascular plants of Texas
.
College Station
:
Texas Agricultural Experiment Station, MP-1655
.
Healy
WM,
Kimmel
RO,
Holdermann
DA,
Hunyadi
W.
1980
.
Attracting ruffed grouse broods with tape-recorded chick calls
.
Wildlife Society Bulletin
8
:
69
71
.
Hosmer
DW,
Lemeshow
S.
2000
.
Applied logistic regression
.
New York
:
John Wiley and Sons
.
Hoyt
DF.
1979
.
Practical methods of estimating volume and fresh egg weight of bird eggs
.
Auk
96
:
73
77
.
Jenkerson
CB,
Maiersperger
T,
Schmidt,
G.
2010
.
eMODIS: user-friendly data source. U.S. Geological Survey Open-File Report 2010–1055
(see Supplemental Material, Reference S1).
Johnson
JA,
Schroeder
MA,
Robb
LA.
2020
.
Greater prairie-chicken (Tympanuchus cupido). Version 1.0. in Poole AF, editor. Birds of the world. Ithaca, New York: Cornell Lab of Ornithology.
Jones
RE.
1963
.
Identification and analysis of lesser and greater prairie chicken habitat
.
Journal of Wildlife Management
27
:
757
778
.
Jurries
RW.
1979
.
Attwater's prairie chicken. Austin: Texas Parks and Wildlife Department, F.A. Series No. 18
(see Supplemental Material, Reference S2).
Keetch
JJ,
Byram
GM.
1968
.
A drought index for forest fire control
.
Asheville, North Carolina
:
U.S. Forest Service Southeastern Forest Experiment Station Research Paper SE-38 (see Supplemental Material, Reference S3)
.
Kessler
WB.
1978
.
Attwater's prairie-chicken ecology in relation to agricultural and range management practices. Doctoral dissertation
.
College Station
:
Texas A&M University
.
Kobriger
GD.
1965
.
Status, movements, habitats, and foods of prairie grouse on a Sandhills refuge
.
Journal of Wildlife Management
29
:
788
800
.
Lehmann
VW.
1941
.
Attwater's prairie chicken, its life history and management
.
North American Fauna
57
:
1
65
.
Lehmann
VW.
1968
.
The Attwater prairie chicken, current status and restoration opportunities
.
Transactions of the North American Wildlife Conference
33
:
398
407
.
Lehmann
VW,
Mauermann
RG.
1963
.
Status of Attwater's prairie chicken
.
Journal of Wildlife Management
27
:
713
725
.
Lutz
RS,
Silvy
NJ.
1980
.
Predator response to artificial nests in Attwater's prairie chicken habitat
.
Pages
48
51
in
Vohs
Jr,
PA
Knopf
FL,
editors.
Proceedings of the prairie grouse symposium
.
Stillwater
:
Oklahoma State University
(see Supplemental Material, Reference S4).
Manzer
DL,
Hannon
SJ.
2008
.
Survival of sharp-tailed grouse Tympanuchus phasianellus chicks and hens in a fragmented prairie landscape
.
Wildlife Biology
14
:
16
25
.
Marks
JS,
Marks
VS.
1987
.
Influence of radio collars on survival of sharp-tailed grouse
.
Journal of Wildlife Management
51
:
468
471
.
Matthews
TW,
Tyre
AJ,
Taylor
JS,
Lusk
JJ,
Powell
LA.
2011
.
Habitat selection and brood survival of greater prairie-chickens
.
Pages
21
32
in
Sandercock
BK,
Martin
K,
Segelbacher
G,
editors.
Ecology, conservation, and management of grouse
.
Berkeley
:
University of California Press
.
Studies in Avian Biology Number 39.
McGaughey
RJ.
2017
.
FUSION/LDV: software for LIDAR data analysis and visualization v2.80. USDA Forest Service
.
Portland, Oregon
:
Pacific Northwest Research Station
.
McNew
LB,
Gregory
AJ,
Wisely
SM,
Sandercock
BK.
2012
.
Demography of greater prairie-chickens: regional variation in vital rates, sensitivity values, and population dynamics
.
Journal of Wildlife Management
76
:
987
1000
.
McNew
LB,
Sandercock
BK,
Pitman
JC.
2011
.
Impacts of alternative grassland management regimes on the population ecology of grassland birds. Project W-67 Second Quarter Report to Kansas Department of Wildlife and Parks
(see Supplemental Material, Reference S5).
Morrow
ME,
Adamcik
RS,
Friday
JD,
McKinney
LB.
1996
.
Factors affecting Attwater's prairie-chicken decline on the Attwater Prairie Chicken National Wildlife Refuge
.
Wildlife Society Bulletin
24
:
593
601
.
Morrow
ME,
Chester
RE,
Lehnen
SE,
Drees
BM,
Toepfer
JE.
2015
.
Indirect effects of red imported fire ants on Attwater's prairie-chicken brood survival
.
Journal of Wildlife Management
79
:
896
906
.
Morrow
ME,
Toepfer
JE.
2020
.
Use of predator-deterrent fences to increase Attwater's prairie-chicken nest success
.
Journal of Fish and Wildlife Management
11
:
455
462
;
e1944-687X. Available: https://doi.org/10.3996/112019-JFWM-099 (April 2022)
Moss
R.
1986
.
Rain, breeding success and distribution of capercaillie Tetrao urogallus and black grouse Tetrao tetrix in Scotland
.
Ibis
128
:
65
72
.
Moss
R,
Watson
A,
Parr
R.
1975
.
Maternal nutrition and breeding success in red grouse (Lagopus lagopus scoticus)
.
Journal of Animal Ecology
44
:
233
244
.
Moss
R,
Watson
A,
Rothery
P,
Glennie
WW.
1981
.
Clutch size, egg size, hatch weight and laying date in relation to early mortality in red grouse Lagopus lagopus scoticus chicks
.
Ibis
123
:
450
462
.
Newell
JA,
Toepfer
JE,
Rumble
MA.
1987
.
Summer brood-rearing ecology of the greater prairie chicken on the Sheyenne National Grasslands
.
Pages
24
43
in
Bjugstad AJ, technical coordinator. Prairie Chickens on the Sheyenne National Grasslands. Fort Collins, Colorado: U.S. Forest Service Rocky Mountain and Range Experiment Station General Technical Report RM-159
(see Supplemental Material, Reference S6).
Panek
M.
1992
.
The effect of environmental factors on survival of grey partridge (Perdix perdix) chicks in Poland during 1987–89
.
Journal of Applied Ecology
29
:
745
750
.
Parish
DMB,
Sotherton
NW.
2007
.
The fate of released captive-reared grey partridges Perdix perdix: implications for reintroduction programmes
.
Wildlife Biology
13
:
140
149
.
Pratt
AC.
2010
.
Evaluation of the reintroduction of Attwater's prairie-chickens in Goliad County, Texas. Master's thesis
.
Kingsville
:
Texas A&M University
.
R Core Team.
2020
.
R: a language and environment for statistical computing
.
Vienna
:
R Foundation for Statistical Computing
.
Available: https://www.R-project.org/ (April 2022)
Riley
TZ,
Clark
WR,
Ewing
DE,
Vohs
PA.
1998
.
Survival of ring-necked pheasant chicks during brood rearing
.
Journal of Wildlife Management
62
:
36
44
.
Robel
RJ,
Briggs
JN,
Dayton
AD,
Hulbert
LC.
1970
.
Relationships between visual obstruction measurements and weight of grassland vegetation
.
Journal of Range Management
28
:
295
297
.
Rymesova
D,
Tomasek
O,
Salek
M.
2013
.
Differences in mortality rates, dispersal distances and breeding success of commercially reared and wild grey partridges in the Czech agricultural landscape
.
European Journal of Wildlife Research
59
:
147
158
.
Schole
AC,
Matthews
TW,
Powell
LA,
Lusk
JJ,
and
Taylor
JS.
2011
.
Chick survival of greater prairie-chickens
.
Pages
247
254
in
Sandercock
BK,
Martin
K,
Segelbacher
G,
editors.
Ecology, conservation, and management of grouse
.
Berkeley
:
University of California Press
.
Studies in Avian Biology Number 39.
Smeins
FE,
Diamond
DD,
Hanselka
CW.
1991
.
Coastal prairie
.
Pages
269
290
in
Coupland
RT,
editor.
Ecosystems of the world 8A—natural grasslands—introduction and western hemisphere
.
New York
:
Elsevier Press
.
Starns
HD,
Fuhlendorf
SD,
Elmore
RD,
Twidwell
D,
Thacker
ET,
Hovick
TJ,
Luttbeg
B.
2020
.
Effects of pyric herbivory on prairie-chicken (Tympanuchus spp) habitat. PLoS ONE 15(6):e0234983
.
Strategic Mapping Program (StratMap).
2009
.
Goliad, McMullen, & Zapata Counties Lidar, 2009-06-20
.
Strategic Mapping Program (StratMap).
2011
.
Austin, Grimes, & Walker Counties Lidar, 2011-02-03
.
Svedarsky
WD.
1979
.
Spring and summer ecology of female greater prairie chickens in northwestern Minnesota. Doctoral dissertation
.
Grand Forks
:
University of North Dakota
.
Svedarsky
WD,
Toepfer
JE,
Westemeier
RL,
Robel
RJ.
2003
.
Effects of management practices on grassland birds: greater prairie-chicken
.
Jamestown, North Dakota
:
Northern Prairie Wildlife Research Center (see Supplemental Material, Reference S7)
.
Toepfer
JE.
2003
.
Prairie chickens & grasslands: 2000 and beyond. Report to the Council of Chiefs, Society of Tympanuchus Cupido Pinnatus. Elm Grove, Wisconsin
(see Supplemental Material, Reference S8).
Udall
SL.
1967
.
Native fish and wildlife endangered species. Federal Register 32:4001
.
U.S. Endangered Species Preservation Act of 1966,
Pub. L. No. 89-669, 80 Stat. 926 (Oct. 15, 1966)
(see Supplemental Material, Reference S9).
[USFWS] U.S. Fish and Wildlife Service.
2010
.
Attwater's prairie-chicken recovery plan. 2nd revision
.
Albuquerque, New Mexico
:
U.S. Fish and Wildlife Service (see Supplemental Material, Reference S10)
.
[USFWS] U.S. Fish and Wildlife Service.
2012
.
Attwater Prairie Chicken National Wildlife Refuge comprehensive conservation plan and environmental assessment
.
Albuquerque, New Mexico
:
U.S. Fish and Wildlife Service (see Supplemental Material, Reference S11)
.
[USFWS] U.S. Fish and Wildlife Service.
2021
.
Attwater's greater prairie-chicken (Tympanuchus cupido attwateri) 5-year review: summary and evaluation. Attwater Prairie Chicken National Wildlife Refuge and Texas Coastal Ecological Services.
(see Supplemental Material, Reference S12).
Venables
WN,
Ripley
BD.
2002
.
Modern applied statistics with S. 4th edition
.
New York
:
Springer
.
Wiedenfeld
JK.
2010
.
Soil survey of Goliad County, Texas. U.S. Department of Agriculture, Natural Resources Conservation Service
.
Wilson
LT,
Yang
Y,
Wang
J.
2015
.
Integrated Agricultural Information and Management System (iAIMS): world climatic data. April 2015
.

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Author notes

Citation: Morrow ME, Lehnen SE, Chester RE, Pratt AC, Sesnie SE, Kelso J, Feuerbacher CK. 2022. Factors affecting survival of Attwater's prairie-chicken broods. Journal of Fish and Wildlife Management 13(2):359–374; e1944-687X. https://doi.org/10.3996/JFWM-21-054

Supplemental Material