The What, Where, and How of Altitudinal Migration

Unlike obligate, latitudinal migrations, altitudinal migrations typically (1) involve short distances, (2) are controlled facultatively, and often (3) consist of partially migratory populations. These attributes make biological sense; ecological gradients change quickly over short distances with elevational change (Körner 2007), meaning that birds may not need to move as far as do latitudinal migrants to experience similar fitness payoffs from migration (e.g., escaping unfavorable climatic conditions, exploiting food resources in areas with different phenological patterns). Lower movement costs likely reduces selection for morphological and physiological adaptations required to complete those movements. Thus, the expression of migration can be more flexible, varying among individuals (e.g., Gillis et al. 2008) or within individuals under different circumstances during their lifetimes (Norbu et al. 2013).

Altitudinal migrant birds live in all major mountain ranges of the Earth (Figure 1). Asia may be home to the greatest number of altitudinal migrants in the world where, purportedly, up to 65% of high-elevation–breeding species in the Himalayas engage in such movements. Altitudinal migrants are distributed throughout the avian tree of life with representatives in orders as morphologically and ecologically distinct as the Charadriiformes (Ferrari et al. 2008) and Psittaciformes (Chassot and Monge-Arias 2012). Altitudinal migrants also represent a broad range of foraging guilds. Neotropical nectarivores and frugivores are well-known altitudinal migrants (Stiles 1980, Strewe 2006), but nectarivores elsewhere (Grant and Grant 1967), insectivores (Nocedal 1994), carnivores (Bildstein 2004), piscivores (Mackas et al. 2010), granivores (Horvath and Sullivan 1988, Borras et al. 2010), and birds of other dietary guilds (e.g., Martínez del Rio et al. 2009) also migrate altitudinally. They range in body size from under a few grams to several kilograms (Boyle 2011). Altitudinal migrants live in forests, deserts, shrublands, and terrestrial aquatic environments. They dwell at all levels from the ground (e.g., Dark-eyed Juncos; Rabenold and Rabenold 1985) up to the canopy (e.g., Great Green Macaws; Chassot and Monge-Arias 2012), are flocking (e.g., Mountain Quail; Brennan et al. 1987) or solitary (e.g., Andean Condors; Sick 1993), and represent diverse mating systems.

FIGURE 1.

Altitudinal bird migration around the world. Numbers in red ovals correspond to references listed below documenting altitudinal migration in most major mountain ranges globally. Tallies of community- or guild-level prevalence of altitudinal migration (reported in dark boxes) provide a comparison to the results of this study in North America (inset). [1] Rabenold and Rabenold 1985; [2–4] Cade and Hoffman 1993, Connelly 1988, Dixon and Gilbert 1964; [5, 6] Dunning and Bowers 1984, Horvath and Sullivan 1988; [7–11] Garwood et al. 2009, Gillis et al. 2008, Grinnell and Miller 1944, Laymon 1989, Lorenz and Sullivan 2009; [12, 13] Des Granges 1979, Solórzano et al. 2000; [14–20] Boyle 2011, Chassot and Monge-Arias 2012, Chaves-Campos 2004, Chaves-Campos et al. 2003, Fraser et al. 2008, Loiselle and Blake 1991, Stiles 1980; [21–27] Newsome et al. 2015, Tinoco et al. 2009, Strewe and Navarro 2003, Hobson et al. 2003, Hilty 1997, Hughes 1980, 1984; [28] Beebe 1947; [29–31] Galetti 2001, Bencke and Kindel 1999, Areta and Bodrati 2010; [28, 32–34] Beebe 1947, Bildstein 2004, Alves 2007, Capllonch 2007; [35, 36] Borras et al. 2010, Klemp 2003; [37, 38] Brambilla and Rubolini 2009, De La Hera et al. 2014; [39] Burgess and Mlingwa 2000; [40, 41] Brown 2006, Johnson and Maclean 1994; [42] Lu et al. 2010; [43, 44] Kimura et al. 2001, Ryan 2012; [45, 46] Dingle 2004, 2008; [47] Nocedal 1994; [48] Stiles and Skutch 1989; [32] Bildstein 2004; [39] Burgess and Mlingwa 2000; [41] Johnson and Maclean 1994; [49] Barve et al. 2016; [45] Dingle 2004; [50] Tidemann et al. 1988; [51] Dixit et al. 2016.

Although species differ in the elevations they occupy, many migrants breed at higher elevations than where they spend their nonbreeding season in both tropical and temperate regions (Dixon and Gilbert 1964, Ramos-Olmos 1983, Norment and Green 2004, Powell and Bjork 2004). Many exceptions do occur, however. For example, the endangered Nēnē or Hawaiian Goose (Branta sandvicensis) breeds and molts at low elevations, migrating to high elevations during the nonbreeding season (Hess et al. 2012). A few species apparently make both uphill and downhill movements post-breeding; in the New England region of New South Wales, Australia, Eastern Spinebills (Acanthorhynchus tenuirostris) migrate uphill in winter (Ford and Pursey 1982), whereas farther south in the same mountain range, this species migrates to lower elevations in winter (Tidemann et al. 1988). In partially migratory populations, the elevation where nonmigrant individuals reside year-round can be at either the upper or lower end of the elevational gradient. In species where nonmigrants spend the nonbreeding season at high elevation, the migrants and residents generally breed together but spend the nonbreeding season apart (e.g., Resplendant Quetzals; Solórzano et al. 2000). In contrast, where nonmigrants spend the nonbreeding season at low elevations, then migrants and residents breed separately, meaning that the potential for assortative mating is strong (e.g., American Dippers; Middleton et al. 2006).

We know little regarding how birds make altitudinal migratory movements. Mountain Quail apparently traverse >2,000 m elevational gradients primarily on foot (Gutiérrez and Delehanty 1999). Whether small songbirds make altitudinal migrations in nocturnal flights as do many of their long-distance counterparts is unknown. Likewise, for most species, we do not know whether migrations occur in a single directed flight or many short flights. In Central American Three-wattled Bellbirds (Procnias tricaunculata), radio-tagged birds typically departed their high-elevation, Pacific-slope breeding areas in a single sustained flight to the Caribbean lowlands. A few months later, they recrossed the Continental Divide in multiple flights to the Pacific lowlands before returning to high-elevation breeding areas (Powell and Bjork 2004). Satellite-tagged Nēnē exhibit considerable inter-individual flexibility in the duration and distance of migratory journeys, with some individuals making nonstop flights while others make multiple stopovers (Leopold and Hess 2014). In some tropical passerines, it seems likely that migratory flights can be fast and directed given the rapid increase in capture rates of altitudinal migrants following storm events (Ramos-Olmos 1983, Boyle et al. 2010). In this way, some altitudinal migrations parallel the unpredictable, short-term elevational movements of birds escaping from inclement weather (Hahn et al. 2004, O'Neill and Parker 1978); a key distinction is that, once in the lowlands, tropical altitudinal migrants remain for the rest of the nonbreeding season.

Why Do Birds Migrate Altitudinally?

The causes of altitudinal migration have only been studied in a few systems and species. Soon after altitudinal migration was described for tropical birds (Beebe 1947, Slud 1964), the idea that such movements were driven by spatial and temporal variation in food resources gained traction (Stiles 1980, 1988; Wheelwright 1983). This hypothesis made strong intuitive sense, especially given broad species-level associations between short-distance intra-tropical migration and dependence on nectar or fruit (Levey and Stiles 1992). Several studies tested predictions of this hypothesis to explain community-level (Loiselle and Blake 1991) and species-level patterns of abundance and movement (e.g., Rosselli 1994, Chaves-Campos 2004). Frequently, while part of the annual movement cycle was consistent with birds seeking elevations with high food abundance, the timing and direction of return movements did not match food-based predictions. For example, Bare-necked Umbrellabirds (Cephalopterus glabricollis) bred in locations and seasons of high food abundance, but migrated downhill to elevations having lower food availability (Chaves-Campos et al. 2003). Species-pairs of frugivorous migrants and nonmigrants revealed that neither dietary specialization nor competitive asymmetries could explain differences in migration, but even within genera, migrants consumed more fruit relative to arthropod prey (Boyle et al. 2011a). This finding suggested that energetic constraints driven by dependence on nutritionally dilute foods may underlie migratory tendency in some tropical species.

Evidence exists for alternative processes driving altitudinal migration including predation risk (Fretwell 1980, Greenberg 1980) and climatic constraints (Cox 1985). For instance, differences in nest predation risk could potentially explain the movements of birds between middle and high elevations (Boyle 2008a). However, in many species, the temporal and spatial patterns predicted by multiple ecological processes overlap (Ornelas and Arizmendi 1995). Climatic constraints have received relatively little attention in tropical regions despite Skutch (1969: 85) writing, “The Bellbird's wanderings appear to be caused partly by changes in weather….” However, Nocedal (1994) conjectured that in western Mexico, interannual variation in migratory propensity may well reflect individual differences in climatic tolerances and interannual variation in weather.

A mechanistic link between dietary guild, climate, and migration exists; dependence on low-protein, low-fat, dilute foods means that frugivores and nectarivores must forage frequently. Intense rain storms or cooler weather that increase thermogenic costs will affect small birds most acutely as they have narrow energetic safety margins and require high food intake rates. Storms can potentially limit the number of foraging hours to such an extent that such birds with high metabolic rates risk starvation, not because there is not enough food, but because the climate constrains their ability to forage (Boyle 2008b). Evidence in support of this hypothesis comes from Costa Rican White-ruffed Manakins (Corapipo altera); heavy rainfall is perceived by birds as a stressor and leads to increased fuel storage and dependence on stored energy for survival (Boyle et al. 2010). Additionally, in a diverse set of species, more altitudinal migrants reach the lowlands in years with wetter wet seasons (Boyle 2011). Thus, it seems likely that, at least in some tropical systems, the relationship between diet and migratory strategy are not driven by simple food availability, but via interactions with climate and energetics to shape individual and species-level migratory behavior.

In temperate North America, interest in altitudinal migration dates back farther than in tropical systems (e.g., Presnall 1935, Grinnell and Miller 1944, Dixon and Gilbert 1964). Climate has always made intuitive sense in explaining why birds might “escape” from temperate high elevations during winter. However, for most dietary guilds, seasons of climatic harshness coincide with food scarcity, and few authors have attempted to determine whether direct physiological challenges imposed by weather or insufficient food is more important in shaping movement. Carolina Dark-eyed Juncos engage in female-biased downhill migration away from high-elevation breeding areas (Rabenold and Rabenold 1985), as would be predicted by the dominance, body-size, and arrival time hypotheses proposed to explain differential and partial latitudinal migration (Ketterson and Nolan 1983). Interannual variation in the identity and abundance of high-elevation residents implicates climate in mediating some costs of residency and that, likely, competition for scarce winter food explains why many individuals depart breeding grounds. In other well-studied temperate systems, competition also appears to be important in shaping migratory strategies. American Dippers breed at both high and low elevations in southwestern Canada and winter together along low-elevation rivers. Snow and ice restricts access to food during winter at high elevations. Because fecundity is lower in migrants, and survival is only a little higher, the benefits of breeding at high elevation are less clear (Green et al. 2015) and, possibly, altitudinal migration in this system is driven by intraspecific competition for food-rich low-elevation territories. In contrast, Yellow-eyed Juncos (Junco phaeonotus) breeding at high elevations of Southern Arizona's sky islands appear to be limited in their ability to remain at high elevations year-round due to direct and indirect energetic consequences of cold temperatures and snow limiting foraging opportunities during winter (Lundblad 2014).

Generalization regarding the causes of altitudinal migration within major biogeographic regions or among them is extremely difficult due to the few species that have been the subject of detailed investigations. Given the diversity of patterns exhibited by altitudinal migrants, multiple ecological processes almost certainly shape this behavior. One important emergent conclusion is that altitudinal migrations are often facultative, at least within the Americas. Thus, the fitness costs and benefits of alternate strategies can be interpreted in the context of current ecological drivers, making them excellent systems in which to understand the evolutionary pressures that have shaped less labile movement strategies.

A Systematic Survey

I systematically assessed the prevalence of altitudinal migration in the landbirds breeding in North America north of the Mexico border (i.e. Canada and the USA) and Hawaii. I used the Birds of North America (Rodewald 2015) as it represents the most comprehensive set of life histories available for this avifauna presented in a common format. I accessed this source and tabulated data during the summer and fall of 2014. I included all 607 extant, native landbirds breeding in this region, excluding the orders Pelecaniformes, Procellariformes, Gaviiformes, Podicipediformes, Phaethontiformes, Suliformes, and the families Stercorariidae and Alcidae as these groups are almost exclusively aquatic with limited opportunities for altitudinal movements. Nomenclature and species concepts follow the American Ornithologists' Union Check-list and supplements (American Ornithologists' Union 1998, Chesser et al. 2013).

I read the Introduction, Migration, Distribution, and Habitat sections for each species. I tabulated information for all species for which any populations were mentioned migrating altitudinally, referencing original sources. Use of the word “migration” varied among authors of life history accounts. Accordingly, I included species referenced as undertaking “elevational movements,” “downhill dispersal,” or other similar terms when they referred to seasonal movements between breeding and nonbreeding areas. I coded the strength of evidence qualitatively. I considered evidence to be “strong” when accounts reported results of movement-focused studies or if reports from multiple populations reported altitudinal migration with no ambiguity in the language describing altitudinal movements. I considered the strength of evidence to be “intermediate” when references consisted of more anecdotal information, but different authors working in different populations reported such movements and authors did not use ambiguous wording. I considered the evidence to be “weak” if movement descriptions were accompanied by words denoting uncertainty (e.g., “In montane areas, Pileated Woodpeckers may move to lower elevations in winter (Burleigh 1972; Simpson and Pratt 1992)”). I coded each putative altitudinal migrant species as completely or partially migratory, and if partially migrant, whether that variation occurred among or within populations. For partially migratory populations, I coded individuals as either breeding or wintering together. I noted species that included latitudinal migrant populations, and whether any population overwinters within the continental USA and Canada. I coded species for geographic overlap in breeding and wintering range and, where present, whether that overlap occurred within the continental USA and Canada.

I classified species into one of eight broad geographic distributions: “western” and “eastern” for species whose ranges primarily lie west or east of the 100th meridian, “widespread” for species occurring commonly in both eastern and western regions, “northern” for species occurring wholly or primarily in boreal or more northerly regions, “southwestern” for species only occurring in North America in deserts or sky islands of the Southwest or extreme southern Texas, “Hawaii” for species not found elsewhere in North America, and “other” for species not easily classifiable into any of these categories. I excluded extinct or introduced birds in the summaries below, but read and tabulated data for such species' accounts when they were available. Although there were few missing species accounts, life histories varied in the date of initial publication and subsequent revision, and it is possible that new studies may have been published since tabulation. Given constraints of study design and the relative paucity of studies on altitudinal migration in North America, the estimates presented below should be viewed as extremely conservative minima.

Why Study Altitudinal Migration?

There are at least 4 main reasons to fill this gap in knowledge.

(1) Embrace the messy. No avifauna, including that of North America, consists of species falling neatly into tidy migrant and resident categories (e.g., Seifert et al. 2016), or even the arbitrary “long-distance” and “short-distance” migrant groups—terms that mean different things to different authors (e.g., short-distance = 10s of km (Nilsson et al. 2008), or up to 1,000 km (Kondo and Omland 2007)). Furthermore, other movement categorizations including irruptions and nomadism belie the vast variation found in both latitudinal (e.g., Strong et al. 2015) and altitudinal movements (e.g., White-headed Woodpecker, Leuconotopicus albolarvatus; Richards 1924). Embracing the sometimes messy, facultative, partial migration behaviors brings opportunities to understand the causes of many potential movement responses arising from environmental selective pressures and constraints. In a recent example, Himalayan altitudinal migrants provided the opportunity to contrast physiological adaptations to high elevations in year-round residents compared to more mobile species that must be phenotypically flexible to cope with changing atmospheric conditions over their annual cycle (Barve et al. 2016). Additionally, the study of “messy” movements can lead to results that challenge dogma arising from study of a biased sample of movement behaviors. An example of this is the molt-migration of many Western songbirds, a pattern important to the understanding of interspecific differences in behavior, evolution, and conservation that has only been described and appreciated in the past few decades (Rohwer and Manning 1990, Leu and Thompson 2002).

A second argument is that studies of “messy” migrations are necessary to fully understand classic obligate migrations. For example, Ramenofsky et al. (2012) highlighted the paucity of studies describing the physiological regulation of facultative migrations, and explained how all migrations are likely governed by combinations of mechanisms. The hypothalamic–pituitary–gonad axis (HPG) is the primary pathway controlling the timing and physiological preparation for obligate migrations, while the hypothalamic–pituitary–adrenal axis (HPA) is probably involved in cueing facultative migrations. However, both types of migrations likely are controlled by both the HPG and HPA to different degrees and/or under different circumstances. As geolocator data is making increasingly obvious (e.g., Stutchbury et al. 2016), obligate migrants frequently engage in nonbreeding movements that do not fit neatly into the paradigm of tidy flights between breeding ranges and wintering ranges via discrete, short-term stopover sites (Newton 2012). A better understanding of the control of more flexible migration systems would inform how and why obligate migrants move in the way they do during these facultative phases of their migratory cycle.

(2) Variation is the fodder of evolution. Many hypotheses have been proposed for how long-distance, obligate migration arises in birds (Alerstam and Enckell 1979, Cox 1985, Lundberg 1988, Holmgren and Lundberg 1993). All share a common feature: individuals vary in their movement behavior, and selection acts on that variation. Altitudinal migration and other short-distance facultative migrations have been proposed to be precursors of long-distance migration in Neotropical lineages (Levey and Stiles 1992). Understanding the selective costs and benefits in obligate long-distance migrants requires that we adopt a comparative approach and make inferences regarding the migratory behavior of ancestral species (e.g., Outlaw and Voelker 2006). However, we know that migration is extremely labile; European Starlings (Sturnus vulgaris) in North America have altered their migratory propensity since being introduced to the continent in 1890 (Bitton and Graham 2015), and House Finches (Haemorhous mexicanus) transported to eastern North America have “evolved” migratory behavior over the past 8 decades (Able and Belthoff 1998). Thus, inferences regarding the selective pressures that have tipped fitness cost–benefit trade-offs toward migrating may be obscured by saturation of gains and losses in phylogenetic contexts. By studying altitudinal migrants and other partial, “messy” migrations, we can study ecological selective pressures in the here and now. In effect, such systems permit us to conduct empirical tests of hypotheses regarding the evolution of migration as it is happening.

(3) Efficiency. We live in an age of competitive job markets (Marshall et al. 2009) and scarce funding ( https://debblog.nsfbio.com/category/deb-numbers/), factors that favor high research efficiency per unit time and dollar. Studying migration over large spatial scales frequently involves costly travel and technology. If such studies take place in wintering areas, they can also involve environmentally costly air travel and time-intensive acquisition of foreign research permits. In altitudinal migration systems, we can usually test the same hypotheses with fewer logistical challenges at the expense of fewer grant dollars relative to comparable studies with latitudinal migrants (e.g., Boyle 2008a vs. McKinnon et al. 2010). Such considerations are particularly relevant to graduate student research where careful choice of study system can permit full annual cycle field-based migration studies, even while completing academic requirements (e.g., Morrissey 1997, Lundblad 2014).

It is possible that the shorter-distance, facultative movements have been bypassed because overlapping summer and winter ranges make broad movement patterns harder to pin down. New tracking technologies are overcoming many of these logistical constraints as ever-smaller tracking devices come on the market, increasing the volume, precision, and accuracy of spatial data. Additionally, altitudinal gradients typically vary in isotopic signatures over short distances (Hobson 2005). By judicious choice of tissue sampling, detecting individual differences in altitudinal migration using isotopes is now possible in birds of any size at relatively low cost (Hobson et al. 2003, Boyle et al. 2011b).

(4) Save the birds. Just as we cannot understand the causes of population declines or implement effective conservation for long-distance migrants if we do not know where they go, we cannot ensure the future of the rest of our avifauna if we are ignorant of their movements. Declines in migratory birds are not limited to long-distance latitudinal migrants. In some cases, critically endangered birds undergo seasonal altitudinal migrations; understanding where, when, and how they move may prove crucial to protecting their future (Guevara et al. 2015). To date, no comparison exists of population trends of altitudinal migrants relative to nonmigrants or long-distance migrants. However, we do know that montane species are subject to unique challenges as climates change (La Sorte and Jetz 2010), and that the biology of montane birds is not as well understood as lowland species (Boyle and Martin 2015). Explicitly designing studies to allow detection and documentation of the patterns and prevalence of altitudinal migration at community or population levels should be a priority in future ornithological studies.