Altitude sickness in pollinators: Skyward emigration holds consequences for a native bee

Abstract

In the era of climate change, organisms globally are being challenged to adapt to increasingly extreme stressors (Lloret et al., 2012; Walters et al., 2012). Such climate stressors are often typified by heat spikes, severe drought, flooding, and fire (Wagner, 2020). For animals requiring snow cover (i.e., the space between snow and soil, also known as the subnivium) to survive winter, this refuge space has been retreating skyward (higher in altitude) and poleward (higher in latitude) as climates warm (Pauli et al., 2013). With the increasing severity of climate stressors, the elevational ranges of plant species are pushing skyward (Jump et al., 2009; Kelly & Goulden, 2008), as are many insect populations (Hodkinson, 2005). Indeed, an emerging biogeographic pattern associated with climate change is the skyward and poleward redistribution of plant and animal populations (Hodkinson, 2005; Jump et al., 2009). Among pollinator communities, climate change has been linked to continental-scale redistributions of bee taxa (Ghisbain et al., 2023), reductions in bee size over the last three decades (Herrera et al., 2023), and precipitous declines in general abundance (Wagner, 2020). The channeling and redistribution of species are expected to increase in frequency and magnitude globally (Hodkinson, 2005; Jump et al., 2009).

High-elevation habitats may serve as near-term biodiversity reservoirs (“sky islands”), particularly for pollinator communities (Wagner, 2020), as climate stressors reduce the viability of populations at lower elevations (Kelly & Goulden, 2008; Lloret et al., 2012). Pollinators and other organisms seeking refuge from excessive heat may find skyward dispersal more expedient than poleward dispersal, given that a 1°C decline in temperature can be achieved (on average) with a 167-m increase in altitude, while the same drop in temperature would necessitate a 145 km increase in latitude (Jump et al., 2009). However, any elevational gradient will be associated with decreasing air pressure (Peacock, 1998), which will impose a degree of hypoxia on colonizing organisms (Hoback & Stanley, 2001; Hodkinson, 2005). Insect species developing under depleted oxygen (O₂) concentrations in controlled laboratory conditions are known to exhibit smaller adult sizes, reduced reproductive capacity, and lower survival rates (Harrison et al., 2018), all of which represent major fitness consequences for insects (Honěk, 1993; Kingsolver & Huey, 2008). This begs the question as to whether altitudinal hypoxia might produce the same types of consequences as artificially reduced O₂ in laboratory experiments.

With this question in mind, we observed the development of a native bee species, Osmia lignaria (Figure 1), across an elevational gradient in the Rocky Mountains, USA. Known commonly as the blue orchard bee, O. lignaria is a univoltine (single generation per year) solitary (non-social) bee that nests within hollow stems and reeds (Bosch & Kemp, 2001). On 10 May 2023, fully provisioned reeds (representing a cohort of wild O. lignaria eggs) were collected in the foothills of the Rocky Mountains in Salt Lake City, UT, at an elevation of approximately 1320 m. The cohort of bees represented a total of 76 eggs, each provisioned in a brood cell (Figure 1c) within the reeds. On 12 May 2023, the reeds were distributed among three sites representing an elevational gradient: 1300 m (Kaysville, UT), 1900 m (Sundance Mountain Resort ticket office, Sundance, UT) and 2500 m (Sundance Mountain Resort ski patrol office). Three reeds were established at 1300 m, three at 1900 m, and four at 2500 m. The fully provisioned reeds (N = 10 mud-capped reeds, with 6–9 brood cells per reed) were placed within unsealed plastic containers in climate-controlled indoor spaces at each of the three elevations (Appendix S1: Section S1.1). Importantly, all the reeds used in this study had been provisioned at a single location (a private residence in the foothills of Salt Lake City, UT) by the resident O. lignaria females. Given that the adult female bees had foraged recently (within days) from the same locally available pool of floral resources, their larvae likely consumed pollen-nectar diets similar in quantity and quality. After they had developed at their respective elevations for approximately six weeks, the reeds were retrieved in late June 2023, when the bees were pupating.

Here we report a series of observations that reveal a surprising pattern of uneaten pollen-provisions (Figure 1f), increased mortality (Figure 2a), and markedly reduced sizes among the bees at higher elevations (Figure 2b). The incidence of uneaten pollen increased with rising altitude, while pupal (or adult) masses declined with altitude, suggesting that with increasing altitude, the bee larvae ceased feeding prematurely, becoming smaller pupae/adults. From the lowest elevation (1300 m) to the higher elevations (1900 and 2500 m), the bees experienced significantly different rates of survivorship (Figure 2a) (Appendix S1: Section S1.2), despite having experienced similar ranges of temperatures and humidities (Appendix S1: Section S1.1). At the highest elevation, 71% of the bee larvae had died, followed by approximately 10% at the intermediate elevation, and only 4% at the lowest elevation (Figure 2a). Among the survivors, mean pupal/adult mass decreased with increasing elevation (Figure 2b). The magnitude of mass loss at each elevational increment was significant for each sex (Appendix S1: Section S1.2), and the pattern of declining mass across the elevational gradient was consistent for each sex (regression statistics for females: F_1,28 = 20.58, p < 0.001; for males: F_1,18 = 37.21, p < 0.001).

Our finding that male and female bee sizes scaled predictably with elevation was mirrored by a trend in their larval feeding patterns (Figure 3a). Specifically, there was an increasing incidence of unfinished pollen provisions with increasing elevation (Appendix S1: Section S1.2). This phenomenon—reduced feeding at higher elevations—represents one mechanism by which the bees developing at higher elevations would have been less capable of optimal growth. At the highest elevation, none of the surviving larvae finished their provisions, which was associated with the greatest reduction in adult size (Figure 3a). At the intermediate elevation, approximately 30% of the brood cells had unfinished pollen provisions, and these bees registered an intermediate mass. At the lowest elevation, all of the larvae finished their pollen provisions, resulting in the heaviest adult mass. It needs to be emphasized how extraordinary it is that the higher-elevation larvae would consistently leave food unfinished. O. lignaria larvae, as well as most solitary bee larvae, receive a single dietary provision, all of which is typically consumed by the developing larva (Bosch & Kemp, 2001), as demonstrated by the larvae at the lowest elevation in our study (Figure 1e). The fact that the surviving bees at the higher elevations left food unfinished suggests there was some aspect of altitude that had altered their behavior/phenotype.

Reduced feeding rates caused by hypoxic conditions are known to constrain adult size in insects (Farzin et al., 2014), and the effects of this dynamic can be amplified by warmer temperatures (Frazier et al., 2001). Indeed, it does appear that increased temperatures can create hypoxic conditions within the hemolymph of insects, deriving from the fact that metabolic demand for O₂ rises with warmer temperatures while the solubility of O₂ in body fluids decreases (Atkinson, 1994; Frazier et al., 2001; Harrison et al., 2018). The bees in our study exhibited many of the same fitness consequences as insect fauna reared under artificially reduced oxygen conditions at low elevations (Farzin et al., 2014; Frazier et al., 2001; Greenberg & Ar, 1996; Harrison et al., 2018). The commonalities among the prior studies and our current findings are threefold: insects developing under hypoxic conditions exhibited (1) greater mortality, (2) reduced larval feeding, and (3) reduced pupal/adult size. Reductions in adult size represent significant fitness consequences among all major insect orders (Honěk, 1993).

Temperature is known to play a role in mediating adult size, and much past work has been devoted to illuminating the temperature-size relationship for ectotherms, converging on the observation that with increasing temperature, ectotherm body size tends to decrease (Atkinson, 1994). There are many exceptions to this temperature-size relationship (Hodkinson, 2005; Horne et al., 2018; Kingsolver & Huey, 2008; Mousseau, 1997), which are themselves instructive, notably in high-elevation habitats, where air pressure is reduced, seasons often shortened, and temperatures cooler (Ohlberger, 2013; Walters & Hassall, 2006). At higher altitudes, adult sizes would be expected to be larger (based on the hotter = smaller temperature-size rule), but this is not often the case (Hodkinson, 2005; Horne et al., 2018; Kingsolver & Huey, 2008), nor was it the case with the bees in our study.

Past work testing explicitly the interactive effects of hypoxia and temperature on insect size found that both have strong effects (Frazier et al., 2001). Higher temperatures can dramatically increase the metabolic demand for O₂ while reducing the oxygen-delivery capacity of a larval insect (Harrison et al., 2018). In one study, there was strong evidence that higher temperatures had created hypoxic conditions within larval insects by elevating the O₂ demand relative to supply, even when the insects were developing under normoxic conditions (Frazier et al., 2001). Interestingly, under ambient hypoxia, adult size was reduced (compared to normoxic conditions) across all temperature treatments, providing further evidence that hypoxia (whether imposed from outside the insect or created within) strongly influences adult size (Frazier et al., 2001). Not surprisingly, the O₂ supply-and-demand relationship is dependent on the normative temperature range and species-specific hypoxia tolerance of an insect (Harrison et al., 2018; Ohlberger, 2013; Walters & Hassall, 2006). Exemplifying this dynamic is a study of temperature in solitary bees (the red mason bee, Osmia bicornis), which found that relatively hot temperatures resulted in smaller adult bees (Kierat et al., 2017). The bee larvae in our study experienced temperatures that were broadly similar to each other (means: 20–22°C; ranging from 18 to 30°C), and all were within the normative temperature range for this bee species (Bosch & Kemp, 2001). If the bees had been subjected to the cooler temperatures typical of higher altitudes, the temperature-size rule would predict larger adult sizes (Atkinson, 1994), yet we found the opposite. Further, O. lignaria emerges in mid-spring and develops to adulthood by early summer, indicating the bees in our study had ample time to complete their growth and development. Given the similarities in temperature and ample development time across the three elevations, the bee larvae were not constrained by time or heat, nor were they subjected to cold/heat extremes. Temperature, therefore, was unlikely to be a significant determinant of the body size reductions or rates of unfinished provisions at the higher elevations.

The mechanisms driving the size reductions and unusual feeding patterns exhibited by the bees in our study were likely centered around the effects of altitudinal hypoxia. Decades of studies looking at how hypoxia can alter insect development suggest that chronic hypoxia expedites insect ontogeny and imposes major fitness consequences (Greenberg & Ar, 1996; Greenlee & Harrison, 2005; Harrison et al., 2018; Harrison & Haddad, 2011; Hodkinson, 2005; Lundquist et al., 2018; Wilmsen & Dzialowski, 2023). An early study of beetle development across a wide range of O₂ concentrations reported that chronic hypoxia not only decreased adult size but also dramatically increased the number of molts that the young beetles required to reach maturity (Greenberg & Ar, 1996). As an insect larva feeds within a developmental stage (i.e., an instar), the volume of its biomass increases, but its respiratory system does not, creating a problem in which the capacity to meet its increasing O₂ demand (during that stage) declines (Greenlee & Harrison, 2005). However, this O₂ supply-and-demand problem is readily resolved by molting, because with each molt, the larva gains a larger exoskeleton and a more reticulate tracheal system (respiratory apparatus), allowing the insect to breathe more effectively (Greenlee & Harrison, 2005; Wigglesworth & Lee, 1982). It has been repeatedly hypothesized that a low O₂ concentration within an insect's body represents the first cue triggering a molt (Greenlee & Harrison, 2005; Lundquist et al., 2018; Wilmsen & Dzialowski, 2023). Referred to as the “oxygen-dependent induction of molting” (ODIM) hypothesis (Wilmsen & Dzialowski, 2023), it suggests that tolerance for hypoxia has a limit within a larval insect, and when exceeded, the molting process is initiated. Intolerance for hypoxia, therefore, represents a physiological threshold that mediates the timing of a molt (Greenlee & Harrison, 2005; Wilmsen & Dzialowski, 2023). Molting thresholds can differ from temperature-mediated growth thresholds (Hodkinson, 2005; Kingsolver & Huey, 2008; Ohlberger, 2013). For this reason, a mismatch may emerge between an insect's ontogenetic rate and its growth rate, thus if its ontogeny has been expedited by hypoxia, the insect may be fast-tracked into adulthood before it has reached its optimal size (Frazier et al., 2001; Hodkinson, 2005; Ohlberger, 2013).

We suspect that such mismatches between ontogeny and growth can be driven by the chronic hypoxia attendant to higher altitudes (Figure 3b). Across the elevational gradient in our study, the one factor that differed markedly between elevations was ambient air pressure. The inverse relationship between air pressure and altitude clearly reduced O₂ availability for the higher-elevation larvae, and they responded by feeding less before pupating (Figure 3a). They did so despite having similar dietary resources, similar temperature/humidity ranges, and ample time in which to complete their development. This dynamic might be commonly exhibited by pollinators as well as other insect fauna developing at high elevations, particularly if such elevations are not part of their typical biogeographical range.

Insect populations that have recently emigrated skyward (without the capacity to accommodate greater hypoxia) might be inadvertently subjecting their progeny to fitness consequences—analogous to altitude sickness—as each generation develops under chronically hypoxic conditions. Reduced fitness as a function of altitude has been reported within butterfly, grasshopper, and fruit fly populations distributed across elevational gradients (Hodkinson, 2005; Horne et al., 2018). In a review of altitude versus adult size among insect fauna, a consistent relationship was not found, but O₂ availability had not been examined explicitly by the studies in the review, and the authors hypothesized that hypoxia was likely an important yet overlooked factor (Horne et al., 2018). The lack of a strict relationship between altitude and insect size may be partially explained by evidence of directional selection favoring tolerance for high altitude (Kingsolver & Huey, 2008), which, over evolutionary timescales, might eventually allow a population to thrive at higher elevations. Until a population develops such tolerance, skyward emigration may force some species into a trade-off wherein they “pay” for their immediate O₂ requirements by “spending” future reproductive currency.

The capacity to endure climate stressors varies widely among pollinator taxa (Ghisbain et al., 2021), and as pollinator populations emigrate skyward, these communities will likely exhibit a diversity of responses to altitudinal hypoxia, with each species potentially experiencing its own “hypoxia ceiling.” Given that bees are key players in the foundational mutualisms between microbes and flowering plants (Steffan et al., 2024) and that declining pollinator diversity has been linked to reduced plant diversity (Biesmeijer et al., 2006), the basic functioning of high-elevation ecosystems may be constrained by the ceilings imposed upon pollinator populations.

Shawn A. Steffan conceived the study. Shawn A. Steffan and Prarthana S. Dharampal designed and conducted the experiments. Shawn A. Steffan compiled and analyzed the data. Shawn A. Steffan and Prarthana S. Dharampal drafted the manuscript.

The authors declare no conflicts of interest.