ABSTRACT Previous studies of Whooping Crane demography used estimates of fecundity rates based on females in captivity, and breeding success was estimated based on either the number of unbanded pairs nesting or the number arriving in wintering areas with chicks. We analyzed demographic data from 12 cohorts of banded Whooping Cranes (Grus americana) from the Aransas National Wildlife Refuge/Wood Buffalo National Park (ANWR/WBNP) population that had not been compiled previously into a single data base and had not been included in previous population analyses. We estimated age-specific survival and natality, parameterized an age-structured density-dependent model, and projected population sizes assuming two different estimates of wintering ground carrying capacity. Sixty-seven of 132 birds banded between 1977 and 1988 formed nesting pairs, females first produced hatchlings when 3 to 7 yrs old, and the annual proportion of mature females that reproduced successfully ranged from 0.566 to 0.606. Population projections indicate that the down-listing criterion of reaching a population size of 1000 individuals might be attained considerably later than the target year (2035) indicated in the Whooping Crane recovery plan. Even assuming that all suitable habitat within a ∼100-km radius of their current winter range could be occupied, projections suggest that population size may be ∼700 in 2035, and might not reach 1000 individuals until the mid-2060s. Based on their territorial behavior on the wintering grounds, long generation time and faithfulness to their migratory route, we suspect that the population growth rate may decrease markedly in the near future and the ANWR/WBNP population may remain below the target down-listing size of 1000 individuals.
Estudios previos de demografía de la Grulla Blanca usaron estimativos de fecundidad usando hembras en cautiverio, y el éxito reproductivo fue estimado basado en el numero de parejas anidando no anilladas o en el número de parejas arribando a las áreas de invierno con juveniles. Analizamos datos demográficos de 12 cohortes anilladas de G. americana en el refugio nacional de vida silvestre de Aransas y en el parque nacional Wood Buffalo (ANWR/WBNP), estos datos no habían sido compilados anteriormente en una sola base de datos y no habían sido incluidos en previos análisis poblacionales. Estimamos la supervivencia y natalidad a edades especificas, parametrizamos una estructura de edad en un modelo dependiente de la densidad y proyectamos tamaños poblacionales asumiendo dos estimados diferentes de capacidad de carga en las áreas de invierno. Sesenta y siete de las 132 aves anilladas entre 1977 y 1988 formaron parejas reproductivas, las hembras produjeron polluelos por primera vez cuando tenían de 3 a 7 años, y la proporción anual de hembras que se reprodujeron exitosamente varió desde 0,566 hasta 0,606. Las proyecciones poblacionales indican que la población de 1000 individuos, criterio establecido para mover a la G. americana a una categoría de amenaza más baja, puede alcanzarse considerablemente más tarde que el año meta (2035), indicado en el plan de recuperación de G. americana. Incluso asumiendo que todo el hábitat adecuado en un radio de ∼100-km del actual rango invernal pueda ser ocupado, proyecciones sugieren que el tamaño poblacional puede ser de ∼700 en el 2035, y puede que no alcance los 1000 individuos hasta mediados del 2060. Basándonos en el comportamiento territorial en sus areas de invierno, el largo tiempo generacional y la fidelidad a su ruta migratoria, sospechamos que la tasa de crecimiento poblacional puede disminuir marcadamente en un futuro cercano y que la población de ANWR/WBNP puede continuar por debajo del objetivo de 1000 individuos para moverlos a una categoría de amenaza más baja.
The only non-reintroduced population of endangered Whooping Cranes (Grus americana), after declining from 88 birds in 1912 to a low of 15 in 1941 (Allen 1952), had increased to 266 individuals by the winter of 2007–2008 (Stehn 2009). The United States Fish and Wildlife Service (USFWS) and the Canadian Wildlife Service (CWS) still list Whooping Cranes as an endangered species, but hope to down-list the species from endangered to threatened by 2035 (CWS and USFWS 2007). Because one of the alternative criteria for possible down-listing is attaining a population of 1000 individuals, including at least 250 breeding pairs, there is much interest in projections of future population growth (CWS and USFWS 2007).
Whooping Cranes winter at Aransas National Wildlife Refuge (ANWR; 28°12′ N, 96°54′ W), Matagorda Island National Wildlife Refuge, and adjacent private lands in Texas (hereafter referred to as Aransas wintering grounds), and breed at Wood Buffalo National Park (WBNP; 60°46′ N, 112°14′ W) on the border of Alberta and the Northwest Territories in Canada. Fall migration occurs during October and early November and the migration route follows the Central Flyway corridor, covering ∼4300 km, with stopover sites in southern Saskatchewan, the central Platte River in Nebraska, the Quivira National Wildlife Refuge in Kansas, and other locations (Lewis 1995; Fig. 1). Cranes arrive at ANWR in November and remain there for ∼5 mo. They depart by April to return to WBNP along the same migratory route (Lewis 1995). Sources of mortality on the breeding grounds include mammalian predators (Kuyt 1981, Boyce 1987, Kuyt and Goossen 1987, Bergeson et al. 2001), migration-related mortality factors include severe weather, predation, shooting, and collisions with fences and electric transmission lines (Chavez-Ramirez 2004, CWS and USFWS 2007, Stehn and Wassenich 2008), and sources of mortality during winter include illegal shooting, disease, and predation (Lewis et al. 1988, Lewis 1995). Despite these hazards during their annual cycle, the ANWR–WBNP population has exhibited exponential growth over the past seven decades, notwithstanding a few short periods of decline (Fig. 2).
Although habitat availability on the breeding grounds appears capable of supporting continued growth (Tischendorf 2004), habitat availability on the wintering grounds could become limiting (Stehn and Prieto 2010; Fig. 1). Mated pairs of Whooping Cranes establish territories each winter to maintain pair-bonds, protect foraging and roost sites, reduce intraspecific competition, decrease risk of predation and human disturbance, and guarantee resources for their young-of-the-year (Allen 1952, Fretwell and Lucas 1969, Alonso et al. 2004). New pairs of cranes must find unoccupied spaces between established territories, creating the potential for density-dependent aggressive interactions (Stehn and Johnson 1987, Stehn and Prieto 2010). Based on an estimated average minimum winter territory size of 172 ha and the habitat occupied by 266 individuals in 2007, Stehn and Prieto (2010) estimated a carrying capacity between 329 and 576 individuals on ANWR and a maximum carrying capacity of 1156 individuals if the population expanded into the adjacent, apparently suitable salt marsh habitat within a 111-km radius of ANWR.
We compiled into a single data base and analyzed demographic data from 12 cohorts of banded Whooping Cranes provided by the CWS and the USFWS. These data had not been included in previous population analyses. We estimated age-specific survival and natality, parameterized an age-structured density-dependent model, and projected population sizes assuming two different estimates of wintering ground carrying capacity. We also compared our results with those of previous studies and interpreted our results within the context of the published down-listing criterion related to population size.
We compiled demographic data (Gil de Weir 2006) from the ANWR–WBNP population provided by the CWS (L. Craig-Moore and B. Johns, unpubl. data) and the USFWS (Stehn 2004). During ground surveys of WBNP from 1977 to 1988, CWS-USFWS personnel banded 132 juveniles. Data provided by the CWS were collected at WBNP from aerial and ground surveys each May and June from 1977 to 2004. Each survey consisted of ∼25 hrs of observations over ∼927 km2 of WBNP and adjacent areas. Data included estimates of number of nests, number of eggs per nest, and number of chicks fledged per nest in June and August for the entire population. In addition, nests of banded adults were visited three times each year, once during the incubation period (May/June), once during the hatching period (June/July), and once during the fledgling period (August). Data provided by the USFWS were collected at ANWR during ground and aerial surveys conducted approximately weekly from mid-October through April from 1977 to 2004 (Fig. 2). Data included censuses of juveniles, subadults, and adults; observations of banded birds; and reports of mortality, with the latter usually inferred from the disappearance of an individual from its territory (Stehn 2004). Each banded individual was positively resighted each year at ANWR, WBNP, or both via ground or aerial surveys from the banding year to 2004. Some banded cranes were also sighted during spring or fall migration (Jobman and Tacha 2004). Detailed descriptions of field methods during banding and surveys are provided elsewhere (Kuyt 1979, Stehn and Taylor 2008).
We estimated age-specific survival (lx, proportion of individuals surviving to the beginning of age x) based on data from 132 banded birds (Supplementary Table S1). Because we were confident that banded birds were observed at least once each year during their life (only one banded bird was ever resighted after not being observed during the preceding year), we used the Kaplan–Meier method (Kaplan and Meier 1958, Bart et al. 2000, Kleinbaum and Klein 2005) programmed in SPSS (v. 15.0; Hair et al. 2007) rather than more commonly used capture–recapture statistics (e.g., Program MARK, White and Burnham 1999) that are required if the probability of resighting is less than 1. We established the number of individuals at age 0 in each cohort as the number of eggs laid in those nests where individuals hatched and were banded. Thus, estimates of survivorship for age 0 (l0) accounted for egg, chick, and juvenile mortality. The Kaplan–Meier method does adjust survival estimates to account for marked individuals whose age at death is not yet known. In this case, banded individuals who were still alive were censored from age classes they had not yet attained (e.g., individuals from the 1988 cohort still alive in 2004, at age 16, were censored from age classes >16). Because this resulted in unreasonably high survival estimates (equal to 1.0 in most cases) based on small sample sizes for age classes ≥16 (only seven birds were monitored for more than 5 yrs beyond age 16), we assumed that survival rates for age classes ≥16 were equal to the mean survival rate for age classes 6–15 (annual survival rates for ages 6 ≤x≤ 15 did not decrease significantly with age; P≥ 0.52).
We estimated age-specific natality (bx, number of females/female per year) for x≤ 19 directly using data from 33 banded females that survived to reproductive maturity (Supplementary Table S2). We estimated age-specific fecundity for x > 19 by extrapolating the linear regression of bx on x for 14 ≤x≤ 19 (bx= 1.460–0.033*x, r2= 0.96, SE = 0.004) because each of these older age classes was only represented by one or two females.
Based on our estimates of lx and bx, we developed an age-structured density-dependent population projection model:
where nt is a population vector containing age-specific abundances and At is an annual transition matrix containing age-specific survival rates in subdiagonal elements and age-specific natality in the first row. Age-specific natality is a product of the breeding proportion (proportion of mature females that breed successfully), average clutch size, and average survival rate of offspring during their first year.
We hypothesized that when competition for resources in the wintering ground becomes intense, some females may not reproduce the following breeding season (Fretwell and Lucas 1969, Both and Visser 2003). Therefore, we assumed that the breeding proportion (α) is affected by population size. This density-dependent term is expressed as (see Caswell 2001)
where p represents the maximum proportion of mature females that breed successfully and c determines how fast that proportion declines as population size increases. We parameterized the function for the breeding proportion such that population sizes were 72 and 266 in 1977 and 2007, respectively, assuming each of two wintering ground carrying capacities (K). Published estimates of K range from 329 to 576 if population size is assumed to be limited by the current winter range and available contiguous salt marsh habitats, and the estimate increases to 1156 if it is assumed that the population could also occupy all suitable habitats within a 111-km radius of ANWR (Stehn and Prieto 2010); we used K= 576 and K= 1156.
Of 132 marked birds, 67 (33 males and 34 females) formed pairs and nested in or near WBNP, and the oldest surviving bird (in 2004) was 27 yrs old (this female died in 2007 at age 30, establishing the maximum observed longevity in the wild). Ages when marked females first produced hatchlings ranged from 3 to 7 yrs, including two females at age 3, 13 at age 4, eight at age 5, seven at age 6, and four at age 7. Most nests had two eggs (405/431, or 94%); 3.7% of all nests (16/431) had one egg; and 2.3% of all nests were not found (10/431). Proportions of reproductively mature females that reproduced successfully (number of nests where a chick fledged/total number of nests monitored) ranged from 0.56 (10/18) in 1977 to 0.61 (20/33) in 1989.
Age-specific annual survival (lx+1/lx) ranged from 0.87 to 1.0 for ages 1–26, and was 0.42 for age class 0, while age-specific natality (bx) increased from 0.06 at age 3 to 0.9 at age 7, remained above 0.9 through age 16, above 0.8 through age 22, and decreased to 0.65 at age 27 (Table 1). Estimated parameters for the breeding proportion function were: (1) p= 0.869 and c= 0.0027 when the carrying capacity was K= 576, and (2) p= 0.702 and c= 0.0009 when the carrying capacity was K= 1156. Model projections under these two scenarios suggested that population sizes would be (1) 492 and (2) 715 in the year 2035, respectively, and would surpass 90% of carrying capacity in (1) 2043 and (2) 2071, respectively, with subsequent population growth within the indicated areas greatly slowed (Fig. 3).
Table 1. Age-specific survival (lx, proportion of individuals surviving to the beginning of age x) and fecundity (bx, number of females born/female per year) estimates for the only non-reintroduced population of Whooping Cranes based on data provided by the CWS (B. Johns and L. Craig-Moore, CWS, unpubl. data) and the USFWS (Stehn 2004).
Our population projections suggest that the down-listing criterion population of 1000 individuals might be attained considerably later than the target year (2035) established in the Whooping Crane recovery plan (Mirande et al. 1997, CWS and USFWS 2007). Our projection, based on the assumption that all suitable habitat within a 111-km radius of ANWR could be occupied, indicates that the population may be about 715 in 2035, and might not reach 1000 individuals until the mid-2060s (2064). If the population were limited to the current winter range and available contiguous salt marsh habitats, a population of 1000 may not be attainable and the population growth rate may slow markedly by 2035 (Fig. 3). Our projections assumed that: (1) the average minimum winter territory size for a family group is 172 ha (current size range is 101 to 304 ha), (2) there are 27,729 ha of suitable marsh habitat that are not contiguous with ANWR that could support an additional 580 Whooping Cranes on 161 territories (assuming 3.6 individuals per territory; Stehn and Prieto 2010), (3) behavior of Whooping Cranes on the wintering grounds will continue as observed to date at ANWR, with offspring, especially males of resident pairs, tending to establish territories next to their parents (Stehn and Johnson 1987), (4) no major shifts in land use will affect the availability of suitable habitat, and (5) environmental factors (e.g., precipitation, temperature, and fresh water inflow to salt marsh habitats) will continue within ranges observed over the past several decades.
Our last two assumptions may be questionable and, if incorrect, limit the interpretation of our model projections to a much greater degree than uncertainties associated with Whooping Crane ecology. Land use changes and changes in land cover are regarded as the primary sources of global environmental change (Millennium Ecosystem Assessment 2005), and climate change is prompting policy responses impacting resource conservation and management worldwide (IPCC Climate Change 2007). Land use changes affecting the quality and availability of wetland habitats along the Texas coast include those associated with hydrocarbon production (White and Morton 1997) and federal wetland alteration permits for recreational, commercial, and industrial development (pursuant to Section 404 of the Clean Water Act; Brody et al. 2008), as well as those driven by perceived regional economic consequences of federal resource policies, such as changes in federal crop subsidies for rice production (Musacchio and Grant 2002). Approximately 30,000 ha of coastal wetlands have been lost along the Texas coast during the last 40 yrs (Morton et al. 2004). It is likely that conversion of wetlands to other uses will also alter the structure and composition of remaining wetlands (Liu and Cameron 2001), thus potentially reducing further the amount of suitable marsh habitat accessible (within ∼100 km of ANWR) to Whooping Cranes.
Global climate change may increase flooding along the Texas coast, as has been projected worldwide (Nicholls 2004), thereby affecting the quality and availability of wetland habitats. Flooding reduces the amount of marsh edge (marsh within 1 m of open water), greatly reducing the abundance of penaeid shrimps, blue crabs, and other nekton (Minello and Rozas 2002) that are important food sources for Whooping Cranes (Lewis 1995, Jay et al. 1996). Production of another important food resource, wolfberries (Lycium carolinianum), may be affected by recent shifts in more regional weather patterns. Whooping Cranes depend heavily on wolfberry fruits when they arrive on the wintering grounds in the fall, and there is a dramatic decrease in wolfberry productivity during summer as temperatures and estuarine salinity rise (Butzler and Davis 2006). Recent shifts in regional weather have been pronounced, e.g., the summer of 2011 (June–August) was the hottest and driest on record in the state of Texas (data covering ∼100 yrs), with an average temperature of ∼1.7˚C hotter than recorded previously, and total precipitation ∼2.5 cm lower than previously recorded (Dolce and Erdman 2011). Declines in the availability of food resources may result in increases in minimum winter territory sizes, with a corresponding decrease in wintering ground carrying capacity.
In view of these assumptions, we want to make clear how we believe that our model projections and our comment regarding reaching the down-listing target related to population size should be interpreted. First, we are sure that Whooping Crane population growth will not continue its long-term historical trend indefinitely because there are limits to growth. Second, our model projections should be viewed as plausible trajectories the Whooping Crane population might follow against which to compare future data and to debate management options. We suspect that the population may not reach 1000 individuals by 2035 (the target year for meeting this down-listing criterion), primarily because the population would fall about 100 individuals short of this goal even if the historical, essentially exponential, growth rate continued, but also because we believe that the growth rate, even assuming no change in environmental conditions, would begin to slow before the population reached carrying capacity.
Dawson et al. (2011) proposed that the vulnerability of a species, or the extent to which it is threatened with decline, reduced fitness, genetic loss, or extinction, should be viewed in terms of sensitivity to change, adaptive capacity, and exposure to change as affected by barriers to dispersal. Based on these criteria, species could be placed within a three-dimensional graph space, divided into broad management zones of preparedness (monitor environmental conditions and population levels, and prepare contingency plans with increasing levels of intervention), low-intensity intervention, and intensive intervention to inform decisions about appropriate research, monitoring, and management strategies. Although the ANWR/WBNP Whooping Crane population does not appear to be threatened with imminent decline or extinction (or, as far as we are aware with reduced fitness or genetic loss), there are reasons to suspect that the population growth rate may decrease markedly and the population may remain precariously close to the target down-listing size. Based on their territorial behavior on the wintering grounds (sensitivity to crowding), long generation time (low adaptive capacity) and faithfulness to their migratory route (in effect, imposing a barrier to dispersal), we suggest that the ANWR/WBNP population falls within the preparedness/low-intensity intervention zone of Dawson et al. (2011) and will remain there for the near future.
We thank all members of the recovery and banding team (CWS and USFWS), and especially B. Johns and L. Craig-Moore of the CWS at WBNP, T. Stehn of the USFWS at ANWR, and W. Jobman and M. Tacha of the USFWS in Nebraska for providing access to internal reports and long-term data sets on Whooping Cranes. We also thank K. Winemiller, E. Weir, and several anonymous reviewers for their comments on earlier versions of the manuscript, and P. Johnsgard and G. Ritchison for their editorial suggestions. Financial support for this project was provided by the San Antonio Water System, the San Antonio River Authority, and the Guadalupe-Blanco River Authority as part of the Lower Guadalupe Water Supply Project, and Texas A&M Research Grant Office of Graduate Studies. The Crane Trust provided much appreciated support.