Natal dispersal of Whooping Cranes in the reintroduced Eastern Migratory Population

Abstract Natal dispersal is a key demographic process for evaluating the population rate of change, especially for long‐lived, highly mobile species. This process is largely unknown for reintroduced populations of endangered avian species. We evaluated natal dispersal distances (NDD) for male and female Whooping Cranes (Grus americana) introduced into two locations in central Wisconsin (Necedah National Wildlife Refuge, or NNWR, and the Eastern Rectangle, or ER) using a series of demographic, spatial, and life history‐related covariates. Data were analyzed using gamma regression models with a log‐link function and compared using Akaike information criterion corrected for small sample sizes (AICc). Whooping Cranes released in the ER dispersed 261% further than those released into NNWR, dispersal distance increased 4% for each additional nesting pair, decreased about 24% for males as compared to females, increased by 21% for inexperienced pairs, and decreased by 3% for each additional year of age. Natal philopatry, habitat availability or suitability, and competition for breeding territories may be influencing observed patterns of NDD. Whooping Cranes released in the ER may exhibit longer NDD due to fragmented habitat or conspecific attraction to established breeding pairs at NNWR. Additionally, sex‐biased dispersal may be increasing in this population as there are more individuals from different natal sites forming breeding pairs. As the population grows and continues to disperse, the drivers of NDD patterns may change based on individual or population behavior.

reduce competition for resources and mates (Ferriere et al., 2000;Greenwood, 1980;Rockwell & Barrowclough, 1987). Dispersal patterns of individuals in the newly established populations can also be used to evaluate habitat suitability of an area and the utility of introducing individuals to bolster small, local populations (Krištín et al., 2006).

Characteristics of individuals including their sex and source
population may influence dispersal patterns in a population. Greenwood (1980) hypothesized that sex-biased dispersal can be due to inbreeding avoidance, and the direction of bias depends on the mating system of the species. In birds, dispersal is typically female-biased (one notable exception is waterfowl) and is often associated with monogamy and resource defense by males (Greenwood, 1980). Dispersal patterns in a population can also depend on whether it is reintroduced, translocated, or naturally occurring (Butler et al., 2005;Calvete & Estrada, 2004;Margalida et al., 2013;Skjelseth et al., 2007). Some studies have found longer dispersal distances in reintroduced populations, which has been attributed to the tendency of reintroduced individuals to search for new territories in unfamiliar habitat or the lack of conspecific attraction due to a low population density (Margalida et al., 2013;Martín et al., 2008;Stamps, 2001)

Population (EMP). Prior to reintroduction efforts, no Whooping
Cranes remained in this part of their range although historic records occurred (Allen, 1952;Austin et al., 2019). Additionally, in 2011 state and federal agencies began reintroducing Whooping Cranes to establish another population of Whooping Cranes in Louisiana, known as the Louisiana Non-migratory Population (LNMP). It is important to understand dispersal patterns and population range expansion for ongoing reintroductions of Whooping Cranes to direct habitat conservation efforts and inform release strategies for captive-reared cranes.
We report natal dispersal distances (NDD) of Whooping Cranes in the EMP and compare them with those reported for other populations of Whooping Cranes and other crane species. We also explore the potential relationship between NDD for Whooping Cranes in the reintroduced EMP and a variety of demographic (age, sex), spatial (number of nesting pairs, release area), and life history (rearing method, release method) variables that might help explain the observed pattern of natal dispersal. If cranes first establish a territory then wait to find a mate, we expect shorter NDD for cranes that start breeding at an older age, compared with younger cranes that have paired and disperse further with their mate to breed. Based on a small population size (100 individuals as of May 2019) in the EMP and a lack of sex-biased dispersal in a similarly small AWBP (185 individuals as of 2002, during the time of Johns et al., 2005 study), we expected cranes in the EMP to also show no sex-biased dispersal (Whooping Crane Eastern Partnership, 2019). Small naturally occurring or reintroduced populations may exhibit a lack of sex-biased dispersal if individuals are coming from a single breeding area and one individual of each sex disperses to a breeding site and therefore have equivalent NDD. As the EMP's breeding density has increased over time, thereby also increasing intraspecific competition for territories, we expected NDD to increase with the number of nesting pairs in the area. Lastly, rearing and release methods may affect a crane's site fidelity or familiarity with an area and potentially their NDD. For example, cranes released using different methods spend varying amounts of time in the area prior to release (approximately 0-123 days), which may affect imprinting on the area, site fidelity, and NDD. We expect cranes that fly and those that spend more time in the area prior to release to have shorter NDD than cranes that spend little to no time at the release area or cannot fly prior to release.

| Reintroduction techniques
Whooping Cranes in this study hatched in the wild or were raised in captivity by either costumed caretakers (costume-reared)  Figure 1; Van Schmidt et al., 2014) to attempt to increase reproductive success and minimize nest abandonments due to black flies (Simulium spp.), which have been problematic at NNWR (Barzen et al., 2018;Converse et al., 2013;. From 2011 to 2012, DAR birds were raised at NNWR until they had fledged, when they were transferred to HNWR, where they were eventually released. This method is known as the modified Direct Autumn Release program, or mDAR. After 2012, DAR cranes were raised only at HNWR or WRM and the mDAR technique was discontinued. Additionally, in 2013 the Partnership began releasing parent-reared juveniles into breeding territories of adult cranes. Parent-reared juveniles typically spent zero or very little time in a release pen and were released directly into adult territories (hard release). Initially, all parent-reared juveniles were released at NNWR, then as pairs became established at locations scattered throughout the range of the EMP, juveniles were released in the ER and other areas outside of NNWR ( Figure 1). As of 2019, captive-reared Whooping Cranes continued to be released in the ER.
Both release areas, NNWR and the ER, were comprised of wetland and upland habitats; however, there were some key differences with respect to size, habitat fragmentation, and wetland characteristics. NNWR is a contiguous 17,683-ha refuge property, with sedge (Carex spp.) meadow wetlands, emergent marshes, prairies, oak (Quercus spp.) savanna, and oak-pine (Pinus spp.) forest. The ER is a large 2,021,800-ha region which includes many separate wetland properties, including HNWR and WRM. Soils in the ER were more productive than the sandy soils of NNWR. The ER had more row crop agriculture and human development and fewer forested areas than NNWR. Unlike the sedge meadows of NNWR, wetlands in the ER tended to be dominated by cattail (Typha spp.) vegetation.

| Banding information and monitoring
Prior to release, each Whooping Crane was uniquely marked with colored plastic leg bands and leg-band-mounted VHF radio transmitters (Advanced Telemetry Systems, Isanti, MN) which enabled long-term monitoring of individuals in the population (Urbanek, 2018). Remote transmitters (Platform Terminal Transmitters or PTTs, Global System for Mobile Communications or GSMs, Microwave Telemetry, Columbia, MD) were also deployed on the leg bands of 2-13 individuals from each cohort (130 total deployed). Whooping Cranes were monitored throughout their lives by a combination of opportunistic resightings, remote telemetry locations, and aerial or ground surveys. Each spring, biologists conducted intensive surveys to locate nesting Whooping Cranes using VHF radio telemetry from the ground as well as from a plane and recorded the identities of the individual cranes at each nest. Biologists visited accessible nest sites after eggs had hatched, were abandoned, or the pair incubated full term (30 days) without successful hatching. During nest visits, the location of each nest was collected using handheld GPS units. In areas where nests were inaccessible, coordinates were taken from a plane during an aerial survey, using the plane's GPS. Sex of each crane was determined from blood samples taken prior to release for captive-reared juveniles, and at banding for wild-hatched juveniles, using genetic techniques (Duan & Fuerst, 2001;Griffiths et al., 1998).

| Statistical analysis
We assessed the influence of multiple traits of individual Whooping Cranes in the EMP on their NDD during 2005-2019. We defined an individual's natal site as either the nest where it hatched in the case of wild-hatched individuals or the site where it was released in the case of captive-reared individuals. We then measured the distance from the natal site to the individual's first nesting location using the "Near" tool in ArcMap version 10.6.1 (ESRI, 2011). In our study, we also included one female-female nesting pair and one male Whooping Crane who nested with a Sandhill Crane (Antigone canadensis), as these individuals attempted to breed, although it was not with a Whooping Crane of the opposite sex. We then examined the influence of the age of the bird at first breeding, the rearing method (costume, parent, wild), its release method (ultralight-led, DAR or soft, mDAR, hard release, or wild-hatched), year, release location (NNWR or ER), and whether it was establishing a new nesting territory with an inexperienced mate or was filling a gap in a previously held territory with an experienced mate (hereafter, mate experience). We used the number of breeding pairs in a given year as a proxy for breeding density. Using Pearson's product-moment correlations, we determined whether our independent variables were correlated (Dormann et al., 2013). No two independent variables with a correlation of more than |r > .50| were included within the same model (Dormann et al., 2013).
We generated gamma-generalized linear models with a log-link function, comprised of all uncorrelated variables as well as a null model, using the "glm" function in R (R Core Team, 2019). We generated 20 a priori models to examine how suites of spatially, demographically, or life history-related variables, or combinations thereof, impacted NDD. To compare models, we used AIC c and the "model. sel" tool in the "MuMIn" package in R ( Barton, 2019;Burnham & Anderson, 2002;R Core Team, 2019). We used conditional model averaging for all models with an Akaike weight of 0.10 or higher using the "model.avg" tool in the "MuMIn" package (Barton, 2019;Burnham & Anderson, 2002;Wagenmakers & Farrell, 2004). We transformed parameter estimates to percent change observed in the log-transformed dependent variable per unit increase in the independent variable following Benoit (2011). We present median values that better represent non-normal data, as well as estimates of the mean and standard error for comparisons with other studies. All statistical analyses were done in R 3.6.0 (R Core Team, 2019).  (Figure 1). On average, the age at first breeding was 4.9 ± 0.3 years old for males, and 3.7 ± 0.2 years old for female Whooping Cranes. Mean NDD for all individuals in this population was 28.7 ± 4.7 km (Table 1). Mean NDD for male Whooping Cranes was 22.9 ± 6.0 km and 34.1 ± 7.3 km for female cranes (Table 1). Due to a few long dispersers, median NDD were shorter than mean NDD (median distance for all birds = 12.4 km, males = 11.7 km, females = 13.4 km, Table 1).

| RE SULTS
The "spatial" model best predicted NDD of Whooping Cranes in the EMP and included the number of nesting pairs in the population and the individual's release location as independent variables (AIC weight = 0.254, Table 2). However, four more models, which included the spatial model with additional demographic and/or life history covariates, were within AIC c delta 2 and had a model weight higher than 0.10 and therefore warrant consideration ( Table 2).
The second-best model included the spatial model plus sex and had a nearly identical AIC weight to the top model (0.232, Table 2).
Conditional average parameter estimates from models within AIC delta 2 suggested Whooping Cranes dispersed 4% further for each additional breeding pair in the population. Cranes released in the ER dispersed 261% further than cranes released at NNWR (Figure 2).
Male Whooping Cranes first nested 24% closer to their natal site than female Whooping Cranes ( Figure 3). As first breeding age increased by one year, dispersal distances shortened 3%. Lastly, individuals establishing a new territory with an inexperienced mate dispersed 21% further than those filling a gap in a previously held territory. Typically, ER cranes that nested outside of the ER tended to establish territories closer to NNWR (Figure 1). Spatial models outperformed life history and demographic models, yet some variables (sex, age at first breeding, mate experience) from life history and demographic models demonstrated value when added to spatial models. However, rearing and release methods were not included in any models with a delta weight above 0.10 predicting Whooping Crane NDD in the EMP (Table 2).

| D ISCUSS I ON
Overall, NDD of Whooping Cranes in the EMP were comparable to those reported for other Whooping Crane populations. Cranes in the EMP dispersed slightly further (mean NDD = 28.7 ± 4.7 km, median = 12.4 km, Table 1) than cranes in the AWBP (mean NDD = 16.6 ± 1.8 km, median = 11.9 km, range = 0.3-54.8 km, n = 61, Johns et al., 2005). Median dispersal distances for the EMP and AWBP were similar; however, values were more positively skewed for the EMP population as evidence of the notably higher mean. This suggests that there were more extreme high values, or lengthy dispersals, within the EMP, but otherwise the dispersal distances were very similar. These extreme values seemed to result from females released into the ER in particular (Figures 2 and 3).  potentially altered habitat or find mates across scattered or isolated populations.
Two models within AIC c delta 2 included sex, suggesting that there is likely some sex-biased dispersal within the EMP. From a traditional hypothesis-testing perspective, sex was not a statistically significant predictor within the second-best model (p = .169, Table 2).
However, from an information-theoretic approach, given that this variable improved the model it likely has a measurable influence on NDD (Burnham & Anderson, 1998. The lack of significance from a hypothesis-testing perspective is due to the high variation across dispersal distances for both male (SD = 45.2) and female (SD = 56.4) Whooping Cranes. To summarize, it is likely that there is some sex-biased dispersal within the EMP but that other variables such as the number of nesting pairs and release site account for significantly more of the variation in NDD than sex (Appendix S1).

This pattern of sex-biased dispersal in the EMP differs from
Whooping Cranes in the AWBP (Johns et al., 2005), as well as the reintroduced LNMP (mean male dispersal distance = 49.9 ± 10.9 km, mean female dispersal distance = 44.9 ± 10.2 km, E. K. Szyszkoski, Notably, 70 of the 117 breeding Whooping Cranes in the EMP paired with mates who were released from the same location; thus, both members of a pair dispersed the same distance from their core release locations to their nest location, resulting in no difference in NDD. Therefore, sex-biased dispersal in the EMP could become more apparent over time as more individuals pair with mates from different natal areas. Whitfield et al. (2009) found the same pattern of NDD at the beginning of a reintroduction of released White-tailed Eagles (Haliaeetus albicilla) in western Scotland. Natal dispersal distances of male White-tailed Eagles did not change over the 25+ year study; however, female NDD increased as the population expanded (Whitfield et al., 2009). In recent years, more EMP Whooping Cranes have hatched in the wild and the Partnership released captive-bred birds at new sites, thus expanding the distribution of birds throughout Wisconsin. As birds from different release areas form pairs, we are able to measure differences in NDD between males and females.
It appears females are beginning to disperse further than males, which could lead to increasing sex-biased dispersal patterns in the future. It is also possible that sex-biased dispersal distances have pothesized that small population size and decreased opportunity to find mates with increased dispersal distances were driving the lack of sex-biased dispersal. Movements by juvenile birds prior to nesting, however, were not well documented. There was no evidence of depredation, competition, or habitat changes influencing patterns of sex-biased dispersal in the AWBP (Johns et al., 2005). Alternatively, the abundance of appropriate breeding habitat in and around Wood Buffalo National Park may be sufficient to limit the need for distant dispersals, while suboptimal or fragmented habitat in the EMP may promote longer distance dispersals (Divoky & Horton, 1995).
The number of nesting pairs in the population and the age at which a crane first nested affected NDD. In years with more breeding pairs in the population, first-time nesters dispersed further from their natal area, suggesting territories closer to release sites were occupied, and individuals had to move further to find suitable, unclaimed breeding habitat. However, if an individual waited to breed and nested for the first time at an older age, they had a shorter NDD. Nesbitt and Tacha (1997) hypothesized that Sandhill Cranes must first occupy a territory, then wait for an available mate. There may be three strategies for cranes to find a mate and a high-quality breeding territory: (1) occupy a territory close to the natal site and wait to find a mate, potentially breeding at an older age but with a shorter NDD, (2) first find a mate, then search together for a vacant breeding territory, potentially breeding sooner but further from the release area with a longer NDD, or (3) remain near the natal site until they can out-compete another male for an established mate and territory. Relatedly, Pasinelli and Walters (2002) found that male Redcockaded Woodpeckers (Picoides borealis) were more likely to defer breeding in favor of remaining as a helper in higher quality territories given an increased probability of eventually inheriting that breeding site. Though Whooping Cranes have a very different social system, it is possible that remaining nearer to the natal site has a cost in terms of age at first breeding but a benefit in terms of habitat quality. However, as a long-lived species, the cost in age at first breeding may be relatively small for Whooping Cranes compared with smaller short-lived species.  Gouar et al., 2008).

Release location had the largest effect on NDD of Whooping
ER cranes selecting territories near NNWR, either due to conspecific attraction or a flexibility of this population to seek out appropriate breeding habitat, may have population-level consequences.
Due to avian-feeding black flies (Simulium spp.) contributing to widespread nest-abandonment at NNWR, continued establishment of Whooping Crane territories in that area could continue to limit self-sustainability in the EMP (Barzen et al., 2018;Converse et al., 2013;. With continued releases of captive-reared individuals in the ER, there may be a stronger influence of conspecific attraction on breeding territory establishment and shorter NDD of ER cranes in the future. Ultimately, a shift in high density breeding areas from NNWR to the ER may contribute to greater productivity in the EMP. Continuing to monitor NDD as the number of breeding pairs in the ER increases will help us better understand the influence of conspecific attraction as well as habitat on crane behavior across the two core release areas. The information gathered in this study will help inform managers of this endangered species with regard to identifying appropriate nesting habitat as well as the logistics of future releases of captive-reared individuals into this population.

ACK N OWLED G M ENTS
The

CO N FLI C T O F I NTE R E S T
None declared.