Study Area. 

The Armendaris Ranch comprises approximately 1439 km² in Sierra and Socorro counties in south-central New Mexico, east of the Rio Grande. The ranch lies within the Jornada del Muerto basin in the northern Chihuahuan Desert. Topography consists mostly of an open valley plain, bordered by the Fra Cristobal Mountains to the west, the Jornada del Muerto Volcano and lava fields to the north, the White Sands Missile Range to the east, and the Jornada Experimental Range to the south. Ranch habitats consisted primarily of Chihuahuan Desert scrub and desert grassland as classified by Dick-Peddie (1993).

The Armendaris is privately owned and managed by Turner Enterprises, Inc., for livestock production of plains bison (Bison bison). The bison herd includes approximately 500 individuals, mostly cows and calves, which are managed to have minimal impact on the grasslands. To promote high quality hunting for several quail species, water and grain stations are common throughout the ranch. Despite the presence of these water and grain stations, avifauna surveys conducted at the Armendaris (Henry 1995, 1998) revealed an apparent breeding-season avian prey biomass not significantly different than other areas of the Chihuahuan Desert (Truett 2002). A habitat suitability model based on vegetative characteristics indicated that the Armendaris contained highly suitable habitat for Aplomado Falcons (Fig. 1; Young et al. 2002, 2005). This included areas of open savannah scattered with tree-yucca (Fig. 2).

Figure 1. 

Habitat suitability model for Aplomado Falcons on the Armendaris Ranch. Map subset from model detailed in report (Young et al. 2002).

Figure 2. 

Open grassland savanna with scattered tree yuccas in the central portion of the Armendaris Ranch.

Falcon Releases. 

In 2006, the U.S. Fish and Wildlife Service, New Mexico Department of Game and Fish, The Peregrine Fund, Turner Endangered Species Fund, and Turner Enterprises, Inc., initiated a reintroduction project as part of a larger effort to restore the Aplomado Falcon to New Mexico. The Armendaris was chosen as the first release site because it offered secure and extensive suitable habitat (Young et al. 2005), seemingly sufficient prey populations compared to other areas of the Chihuahuan Desert (Truett 2002), and had historically supported a small breeding population of aplomados (Ligon 1961).

The Peregrine Fund employed standard raptor hacking (reintroduction) procedures for releasing 102 captive-bred Aplomado Falcons at the Armendaris Ranch from 2006 through 2011 in June, July, and August (Hunt et al. 2013). This procedure involved holding juvenile aplomados, with color VID (visual identification) and aluminum U.S.G.S. bands, in a hack box, for 7–10 d, on an elevated platform erected in suitable habitat, and then releasing them at an age when they would naturally fledge (Mutch et al. 2001). On release day, the box was opened remotely allowing the birds to emerge on their own, thus simulating natural fledging conditions. To promote survival and encourage the aplomados to establish residency near the release site, site attendants provided supplemental food (freshly thawed Japanese quail [Coturnix japonica]), twice a day for approximately 40 d post-release.

Monitoring. 

During and after the release period, Turner Endangered Species Fund personnel conducted three to seven driving surveys per week to locate Aplomado Falcons on the Armendaris and environs (L. Sweikert unpubl. data and J. Truett pers. comm.). Widespread nesting surveys were conducted during the spring of each year, occasionally with the use of airplanes, to determine the existence of any aplomado nests. Additionally, trail cameras were deployed in 2010 to monitor activity at nest platforms and feeding sites.

Extended Supplemental Feeding Program. 

We hypothesized that known survival of recently released aplomados on the Armendaris would increase if the supplemental feeding program was extended beyond the standard 40-d period. We surmised that the aplomados were still relatively naïve, even after 40 d of free-flying experience, owing to their captive origins. We proposed that extending the feeding program would allow these birds to gain more experience, without being food-stressed, which could increase their survival.

The extended supplemental feeding programs varied in length from year to year. Each year feeding continued until late spring of the first year following release, as long as falcons visited the feeding station or unless there was evidence of falcons being targeted by predators at the stations. In an attempt to improve the cost-effectiveness of the program, the schedule of feedings was changed from every day, during 2006 through 2008, to every other day, during 2009 through 2011. For each feeding, one fewer quail than the number of expected falcons was provided in the evenings. Very infrequently, scheduled feedings were missed due to weather or other circumstances.

We recorded which birds attended feedings through direct observations. Observations were continued for the duration of the time the aplomados were eating at the station during a feeding. In 2007, of the total 88 d that food was provided, identifications of birds attending feedings were made only on 43 d, beginning on the fourteenth day food was provided. In all other years, barring technical difficulties, birds were identified on all days that food was provided.

Known Survival. 

We defined “known survival” as the number of days up to the last day a bird was recorded attending a feeding station, or was known to be alive by some other monitoring method. Given the naïveté of the released birds, those who ceased to attend feedings, and were not found via other monitoring methods, were assumed to be dead. This assumption was consistent with The Peregrine Fund’s conclusion that “excessive mortality is implicit, however, in the current lack of a breeding population and in the overall rarity with which the species is seen and reported” (Hunt et al. 2013). However, it is possible that these birds persisted in unknown locations. Unfortunately, we were unable to differentiate between likely mortality and actual mortality because no birds were radio-tagged until 2011. In 2011, 10 aplomados released at the Armendaris were equipped with VHF radio transmitters to document movements and mortality as part of a larger study by The Peregrine Fund (Hunt et al. 2013).

Statistical Analysis. 

All data were analyzed using Microsoft Excel (version 14.4.1) and Minitab (version 16) (Microsoft 2010, Minitab, Inc. 2014). We used logistic regression to predict the probability of known survival of aplomados that participated in the every-day feeding programs versus the every-other-day feeding programs and males versus females. A two-proportion hypothesis test was used to determine if the proportion of birds known to survive to day 30, out of the total number of birds known to participate in every-day feedings, was significantly different than the proportion known to participate in every-other-day feedings. A two-proportion hypothesis test was also used to compare the proportion of males that survived to day 30 to females that survived to day 30. Because the length of the feeding programs varied a great deal from year to year (Table 1), we considered only the known survival of birds for 30 d following the standard 40-d feeding period for logistic regression and hypothesis test of two proportions.

Table 1. 

Number of Aplomado Falcons released in New Mexico, on the Armendaris Ranch, and White Sands Missile Range, and the duration, participation, and percent of time food was available during the extended supplemental feeding program (2006–2011).

Survival analysis, a class of statistical method for studying the occurrence and timing of events, was used to compare the amount of time an aplomado was known to be alive, when fed every day compared to every other day. In this case, the event of interest was the last day the falcon was known to be alive and the response variable was the number of days of known survival up to that last day. Falcons still attending feedings at day 60 were right-censored. Despite its name, researchers frequently use survival analysis to analyze events other than death, such as the time to entry into marriage or divorce (Diekmann 1989, Goldstein 1999). The Log-Rank and Wilcoxon methods of comparison were used for this survival analysis. Because we were comparing the amount of time each individual bird was known to take advantage of the feeding program, we considered the periods of known survival of birds for 60 d following the standard 40-d feeding period.

Finally, a weighted regression was used to determine whether there was a significant difference between feeding-program usage by birds that were offered food every day versus every other day. For this analysis, the response variable was the proportion of times an aplomado fed out of the total number of times food was available while that individual was known to be alive. Because the number of feeding opportunities available to each individual varied by the frequency of feeding and the year, the response variable was weighted by the square root of the number of feeding opportunities while the bird was known to be alive.

Extended Supplemental Feeding and Falcon Reproduction. 

Annual spring surveys revealed that releases at the Armendaris led to the formation of single nesting pairs at the ranch in 2007, 2009, and 2011. These were the only recorded nesting pairs in New Mexico from 2006 through 2011, despite the release of 102 birds at the Armendaris and 203 more birds elsewhere in the state during that time. All nesting individuals had previously attended the extended supplemental feeding programs. Indeed, one bird that attended 124 feedings out of 148 feeding opportunities, during the 2006 feeding program, was part of the 2007 nesting pair, the first nesting pair resulting from the reintroduction effort in New Mexico.

The 2007 pair fledged two young, the 2009 nesting attempt failed, and the 2011 pair fledged three young. Because none of the fledglings were banded, it is difficult to speculate on their survival. However, from 2007 to October 2010, an unbanded female Aplomado Falcon resided at the Armendaris and regularly attended extended supplemental feedings. Circumstantial evidence suggested that this female was one of the young fledged in 2007 and, as an adult, was a part of the failed nesting attempt in 2009 (T. Waddell, Turner Enterprises, Inc. pers. comm.).

Radiotelemetry Study. 

Unfortunately, eight of the ten VHF transmitters placed on aplomados released at the Armendaris malfunctioned, by failing to emit a signal, turning off and on sporadically, or emitting an incorrect signal (i.e., signaling that the aplomado was alive when it was dead). Six aplomados were re-trapped and their transmitters adjusted or replaced with a new VHF transmitter or satellite transmitter. Although the majority of the VHF transmitters continued to malfunction, three fatalities were recorded by tracking two signals, one in mortality mode and one not, and one satellite transmitter. Two of these mortalities were attributed to avian predators and one to a mammalian predator.

Participation in Extended Supplemental Feeding Programs. 

Of the 102 Aplomado Falcons released at the Armendaris, 56 birds participated in the extended supplemental feeding program (Table 1). An additional nine falcons released at the White Sands Missile Range, about 48 km from the Armendaris feeding stations, also participated in the extended supplemental feeding program. Of the 65 total birds that used extended supplemental feedings, 42 were involved in every-day feedings and 23 were involved in every-other-day feedings.

The length of the feeding programs varied every year, from 257 d in 2008 to 37 d in 2011. Most falcons attended feedings in the beginning of the extended supplemental feeding program; however, one falcon, ID 2008-6, did not begin attending feedings until the 82^nd day (Fig. 3). The length (number of days) of falcon attendance at feedings varied greatly.

Figure 3. 

The number of days released juvenile Aplomado Falcons were known to be alive and the number of days each individual participated in the feeding program on the Armendaris Ranch in New Mexico after the conclusion of the standard 40 d supplemental feeding period from 2006–2011.

Effect of Frequency of Feeding on Known Survival. 

The frequency of feeding, i.e., every day or every other day, was a significant predictor of whether the recently released aplomados were known to be alive at day 30 (logistic regression, P  =  0.020, χ²  =  5.41, df  =  2), but the sex of the falcons was not a significant predictor no matter the frequency of feeding (P  =  0.378, χ²  =  0.78, df  =  2). Specifically, the probability of a known falcon surviving to day 30, when provided food every day, was 3.6 times greater (95% CI: 1.173, 11.064) than the probability of known survival to day 30 for a falcon that was provided food every other day. The probability of a male falcon surviving to day 30, regardless of the feeding frequency, was 1.6 times higher than a female’s (95% CI: 0.566, 4.466).

When food was provided daily, 57% of the recently released aplomados were known to be alive at day 30, compared to only 26% of the aplomados when food was provided every other day. A two-proportion hypothesis test indicated the proportion of birds known to be alive at day 30 when food was provided every other day was significantly lower (−0.31, 95% CI: −0.544, −0.077) than those provided food every day (P  =  0.009, Z  =  −2.60).

A two-proportion hypothesis test showed no significant difference (−0.142, 95% CI: −0.382, 0.098) between the proportion of female falcons known to be alive at day 30, compared to the proportion of male falcons known to be alive at day 30 (P  =  0.245, Z  =  −1.16). Specifically, 39% of the females were known to survive to day 30 and 53% of the males were known to survive.

Survival analysis revealed a significantly greater percentage of birds were known to survive to day 60 when fed every day, compared to birds fed every other day (P < 0.001, Fig. 4, M  =  16.27, 95% CI: 10.47, 22.07). The Log-Rank method of comparison had χ²  =  24.96 and the Wilcoxon method of comparison had χ²  =  19.84. Specifically, none of the birds that were offered food every other day were known to survive to day 60, whereas about 50% of the birds fed every day survived to day 60. Longer-term monitoring revealed that although some of these birds persisted into the first and second years following their initial release, most disappeared within a few months of release (L. Sweikert and M. Phillips unpubl. data, Hunt et al. 2013).

Figure 4. 

Known survival (%) of released juvenile Aplomado Falcons fed every day (n = 42 birds) and every other day (n = 23 birds) on the Armendaris Ranch in New Mexico over a 60-d period after the conclusion of the standard 40-d supplemental feeding period from 2006–2011.

Effect of Frequency of Feedings on Aplomado Falcon Attendance at Feeding Stations. 

A weighted regression showed that falcons, during the years when food was offered every day, attended more feedings than falcons during the years when food was offered every other day (P  =  0.001, t  =  −3.37). This result supports the rationale for providing food every day as opposed to every other day.

Known Survival Trends as a Function of Feeding Frequency. 

The results suggest that when daily supplemental feeding was extended, following the end of the standard 40-d supplemental feeding protocol, survival of recently released juvenile Aplomado Falcons increased through at least day 60. Furthermore, when food was provided more frequently, aplomados used the feeding program more. Therefore, it appears that aplomados derived greater benefit from a daily food source than from an every-other-day food source, demonstrating a possible reliance on the feeding program. All outcomes indicate that the extended supplemental feeding program promoted known survival of recently released juvenile Aplomado Falcons. However, as a caveat, we note that because the practice of providing food every day versus every other day occurred in different years, it is possible that variations in weather, naturally available prey, and presence of predators also contributed to these results.

Effect of Extended Supplemental Feeding on Population Establishment. 

Nearly all of the 102 birds released at the Armendaris, with the exception of the members of the nesting pairs and a few others, did not become permanent or even infrequent residents. Similarly, of the aplomados that used the ranch in the year following their release, only a few were known to inhabit the ranch during the second year following their release. Although the final fates of the released birds were largely unknown, there was no evidence that they established themselves elsewhere (Hunt et al. 2013, U.S. Fish and Wildlife Service unpubl. data). In agreement with the Peregrine Fund’s conclusion, it seems likely that most did not survive to reproduce, although it is possible they dispersed to areas with limited human populations or were unobserved (Hunt et al. 2013).

Supplemental feeding programs for other reintroduced, captive-born, carnivores, including the red wolf (Canis lupus rufus; Phillips et al. 2003), Mexican wolf (Canis lupus baileyi; Romo et al. 2013), and the black-footed ferret (Mustela nigripes; L. Sweikert unpubl. data), have promoted increased survival rates for recently released individuals. Unfortunately, and in contrast to these other species, improved short-term prospects for Aplomado Falcons did not appear to translate into improved long-term survival or retention in the general area of the releases for population establishment.

Given the success of the extended daily supplemental feeding program, it remains unclear why the release of over 100 Aplomado Falcons did not lead to the establishment of at least a few birds on the Armendaris Ranch. Releases elsewhere in New Mexico were similarly unsuccessful. What is clear is that the daily-extended supplemental feeding program increased known short-term survival at the Armendaris. These results suggest that a lack of available food is a primary limiting factor to the survival of released aplomados in this area.

Implications for Recovery. 

Hunt et al. (2013) emphasized the importance of drought, and its effects on prey populations, as a principal cause for the lack of success in New Mexico and west Texas. Furthermore, a decrease in prey due to drought might encourage predators to target aplomados as an alternate prey source; continuous yearly drought would cumulatively contribute to a decline in habitat quality (Steenhof and Kochert 1988, Macías-Duarte et al. 2004). Moreover, another applicable factor that may have affected the long-term success of reintroduced Aplomado Falcons at the Armendaris Ranch is that the ranch is at the northernmost edge of the aplomados’ known historical range. Although historical records of breeding aplomados exist for the area around ranch headquarters (Ligon 1961), it is reasonable to conclude that populations of any species at the edge of their range might naturally expand and contract over time (Peterson 2001, Laliberte and Ripple 2004). In addition to the above possible limiting factors at the Armendaris and environs, we propose that changes to the structure of habitat are sufficiently substantial to affect the hunting success of recently released aplomados, in effect, further reducing the amount of food locally available.

Historically, the Aplomado Falcon inhabited open grasslands with prominent widely scattered tall woody vegetation (Ligon 1961, Keddy-Hector 1988, 2000). Nearby woodland, shrubland, and wetland areas may be important for providing aplomados with sufficient prey, species that construct suitable nesting platforms, and suitable roosting locations (Keddy-Hector 2000, Macías-Duarte et al. 2004, C. Perez pers. comm.). Most of the formerly open grasslands have, however, undergone significant vegetation changes in the last century (Grover and Musick 1990). In southern New Mexico, grass cover declined 70% during the preceding century (York and Dick-Peddie 1969). Brush increased proportionately with the decrease in grass cover, in the Jornada Basin, directly south of the Armendaris Ranch (Gibbens et al. 2005). Indeed, on the Armendaris Ranch, grasslands are interspersed with thick patches of brush cover, primarily creosote (Larrea tridentate), that support abundant populations of quail (Baxter and Wolfe 1972, Rollins et al. 2006). Factors contributing to this vegetation shift throughout the grasslands include overgrazing by cattle, fire suppression, and changes in climate (Brown et al. 1997, Curtin et al. 2002, Truett 2010). Additionally, the once common prairie dogs that maintained large brush-free areas are now functionally extinct (Truett et al. 2006).

Regardless of the mechanism, despite the high suitability of the Armendaris according to the habitat suitability model (Young et al. 2002, 2005) and the historical occurrence of breeding aplomados on the ranch (Ligon 1961), portions of grasslands at the Armendaris and elsewhere in the Chihuahuan Desert now support a more obvious brush component that may hinder successful hunting by aplomados. Indeed, although brush encroachment can increase the abundance of prey birds (Macías-Duarte et al. 2009), it may also decrease their availability (Truett 2002). “Aplomados may be sensitive both to the spacing of woody vegetation (because of its effect on feasible pursuit speeds), and the density of ground cover (because of its effect on impenetrability of prey refugia)” according to Keddy-Hector (1988). A dense vegetation structure may provide obstacles for prey to use to avoid capture, mid-chase, and may also provide cover for escape (Janes 1985). These changes in vegetation structure may also promote an increased presence of large avian predators (Hunt et al. 2013). It is noteworthy that reintroduced aplomados released in coastal Texas appeared to avoid even small live oaks (Quercus virginianus), a type of large spreading oak, associated with avian predators such as the Great Horned Owl (Hunt et al. 2013).

Unfortunately, drought, climate change, and historical range contraction are all factors possibly limiting the establishment of aplomados in New Mexico that cannot be easily controlled; however, providing daily supplemental food and reducing the shrub component of desert grasslands might improve prospects for restoring the species. The latter could be achieved through prescribed fire, mechanical removal, and/or chemical application (Fisher 1950, Westoby et al. 1989). The Bureau of Land Management has treated over 400 km² of brush-infested land nearby in the Jornada Basin since 2007 through the application of herbicides, deferred grazing for 5 yr, and prescribed fire (B.L.M. 2009). It is probable aplomados, and other grassland species, would benefit from expanding these restoration efforts throughout the Chihuahuan Desert.

Because of the failure of released aplomados to persist at the Armendaris Ranch, no falcons were released there, or elsewhere in New Mexico for the same reason, in 2013. No releases are planned for the state in 2014 and beyond. The U.S. Fish and Wildlife Service is currently conducting a comprehensive review of the New Mexico reintroduction program, and a similar program in west Texas, to determine if one or both should be permanently terminated. Such terminations would focus recovery efforts on the coastal plains of south Texas where notable challenges exist to further improve circumstances for the species (Hunt et al. 2013). Although much progress has been achieved there on behalf of the species, after 11 yr of reintroductions involving over 800 birds (Jenny et al. 2004), population growth has stalled at 30 or so nesting pairs, likely due to a lack of suitable habitat (Hunt et al. 2013). These results stand in contrast to those conducted in New Mexico (Hunt et al. 2013). Even though far fewer aplomados (337) were released in New Mexico over a shorter period of time (7 yr), not a single known pair persisted.

Even if conditions for aplomados along the south Texas coastal plains improve sufficiently to support 60 nesting pairs, thus satisfying the current criterion for downlisting the species to Threatened status, such a population would almost certainly fall short of that needed for recovering the species in a significant portion of its range and removing it from the Endangered Species List entirely. Though the current Aplomado Falcon recovery plan does not present delisting criterion (U.S.F.W.S. 1990), such criterion would be more demanding than the downlisting criterion.

Prospects are dim for restoring a population of aplomados to the Armendaris Ranch and environs unless those factors responsible for limiting the survival of recently released birds are eliminated or substantially ameliorated. Considering it took 11 yr to establish the existing populations of aplomados in coastal Texas, where prey biomass is much greater, we recommend that the U.S. Fish and Wildlife Service complete a comprehensive assessment of reintroductions in New Mexico before officially terminating restoration work there. If future reintroductions are scheduled at the Armendaris and/or areas with a similar prey biomass, our work supports the inclusion of an extended daily supplemental feeding program through spring following a falcon’s release year if possible, to promote known survival. Additionally, a concerted effort to reduce the abundance of brush in the grasslands, to improve access to prey, may be valuable.