Demographic buffering of life histories? Implications of the choice of measurement scale

Recommendations for quantitative uncertainty consideration in ecology and evolution

Ecological and evolutionary studies are currently failing to achieve complete and consistent reporting of model-related uncertainty. We identify three key barriers - a focus on parameter-related uncertainty, obscure uncertainty metrics, and limited recognition of uncertainty propagation - which have led to gaps in uncertainty consideration. However, these gaps can be closed. We propose that uncertainty reporting in ecology and evolution can be improved through wider application of existing statistical solutions and by adopting good practice from other scientific fields. Our recommendations include greater consideration of input data and model structure uncertainties, field-specific uncertainty standards for methods and reporting, and increased uncertainty propagation through the use of hierarchical models.

Unclear relationships between mean survival rate and its environmental variance in vertebrates

Current environmental changes may increase temporal variability of life history traits of species thus affecting their long-term population growth rate and extinction risk. If there is a general relationship between environmental variances (EVs) and mean annual survival rates of species, that relationship could be used as a guideline for analyses of population growth and extinction risk for populations, where data on EVs are missing. For this purpose, we present a comprehensive compilation of 252 EV estimates from 89 species belonging to five vertebrate taxa (birds, mammals, reptiles, amphibians and fish) covering mean annual survival rates from 0.01 to 0.98. Since variances of survival rates are constrained by their means, particularly for low and high mean survival rates, we assessed whether any observed relationship persisted after applying two types of commonly used variance stabilizing transformations: relativized EVs (observed/mathematical maximum) and logit-scaled EVs. With raw EVs at the arithmetic scale, mean-variance relationships of annual survival rates were hump-shaped with small EVs at low and high mean survival rates and higher (and widely variable) EVs at intermediate mean survival rates. When mean annual survival rates were related to relativized EVs the hump-shaped pattern was less distinct than for raw EVs. When transforming EVs to logit scale the relationship between mean annual survival rates and EVs largely disappeared. The within-species juvenile-adult slopes were mainly positive at low (<0.5) and negative at high (>0.5) mean survival rates for raw and relativized variances while these patterns disappeared when EVs were logit transformed. Uncertainties in how to interpret the results of relativized and logit-scaled EVs, and the observed high variation in EV's for similar mean annual survival rates illustrates that extrapolations of observed EVs and tests of life history drivers of survival-EV relationships need to also acknowledge the large variation in these parameters.

Adaptation to the High-Arctic island environment despite long-term reduced genetic variation in Svalbard reindeer

Typically much smaller in number than their mainland counterparts, island populations are ideal systems to investigate genetic threats to small populations. The Svalbard reindeer (Rangifer tarandus platyrhynchus) is an endemic subspecies that colonized the Svalbard archipelago ca. 6,000-8,000 years ago and now shows numerous physiological and morphological adaptations to its arctic habitat. Here, we report a de-novo chromosome-level assembly for Svalbard reindeer and analyze 133 reindeer genomes spanning Svalbard and most of the species' Holarctic range, to examine the genomic consequences of long-term isolation and small population size in this insular subspecies. Empirical data, demographic reconstructions, and forward simulations show that long-term isolation and high inbreeding levels may have facilitated the reduction of highly deleterious-and to a lesser extent, moderately deleterious-variation. Our study indicates that long-term reduced genetic diversity did not preclude local adaptation to the High Arctic, suggesting that even severely bottlenecked populations can retain evolutionary potential.

Harvesting can stabilise population fluctuations and buffer the impacts of extreme climatic events

Harvesting can magnify the destabilising effects of environmental perturbations on population dynamics and, thereby, increase extinction risk. However, population-dynamic theory predicts that impacts of harvesting depend on the type and strength of density-dependent regulation. Here, we used logistic population growth models and an empirical reindeer case study to show that low to moderate harvesting can actually buffer populations against environmental perturbations. This occurs because of density-dependent environmental stochasticity, where negative environmental impacts on vital rates are amplified at high population density due to intra-specific resource competition. Simulations from our population models show that even low levels of harvesting may prevent overabundance, thereby dampening population fluctuations and reducing the risk of population collapse and quasi-extinction following environmental perturbations. Thus, depending on the species' life history and the strength of density-dependent environmental drivers, low to moderate harvesting can improve population resistance to increased climate variability and extreme weather expected under global warming.

Context-dependent fitness costs of reproduction despite stable body mass costs in an Arctic herbivore

The cost of reproduction on demographic rates is often assumed to operate through changing body condition. Several studies have found that reproduction depresses body mass more if the current conditions are severe, such as high population densities or adverse weather, than under benign environmental conditions. However, few studies have investigated the association between the fitness components and body mass costs of reproduction. Using 25 years of individual-based capture-recapture data from Svalbard reindeer Rangifer tarandus platyrhynchus, we built a novel Bayesian state-space model that jointly estimated interannual change in mass, annual reproductive success and survival, while accounting for incomplete observations. The model allowed us to partition the differential effects of intrinsic and extrinsic factors on both non-reproductive mass change and the body mass cost of reproduction, and to quantify their consequences on demographic rates. Contrary to our expectation, the body mass cost of reproduction (mean = -5.8 kg) varied little between years (CV = 0.08), whereas the between-year variation in body mass changes, that were independent of the previous year's reproductive state, varied substantially (CV = 0.4) in relation to autumn temperature and the amount of rain-on-snow in winter. This body mass loss led to a cost of reproduction on the next reproduction, which was amplified by the same environmental covariates, from a 10% reduction in reproductive success in benign years, to a 50% reduction in harsh years. The reproductive mass loss also resulted in a small reduction in survival. Our results show how demographic costs of reproduction, driven by interannual fluctuations in individual body condition, result from the balance between body mass costs of reproduction and body mass changes that are independent of previous reproductive state. We illustrate how a strong context-dependent fitness cost of reproduction can occur, despite a relatively fixed body mass cost of reproduction. This suggests that female reindeer display a very conservative energy allocation strategy, either aborting their reproductive attempt at an early stage or weaning at a relatively constant cost. Such a strategy might be common in species living in a highly stochastic and food limited environment.

Breeding responses to environmental variation are age- and trait-dependent in female Nazca boobies

Age and environment are important determinants of reproductive parameters in long-lived organisms. These factors may interact to determine breeding responses to environmental change, yet few studies have examined the environmental dependence of aging patterns across the entire life span. We do so, using a 20-yr longitudinal data set of reproductive phenotypes in long-lived female Nazca boobies (Sula granti), a monogamous seabird breeding in the eastern tropical Pacific. Young and old females may suffer from inexperience and senescence, respectively, and/or practice reproductive restraint. Breeding performance (for breeding participation, breeding date, clutch size, egg volume, and offspring production) was expected to be lower in these age classes, particularly under environmental challenge, in comparison with middle-aged breeders. Sea surface temperature anomalies (SSTA) represented interannual variation in the El Niño-Southern Oscillation (ENSO) and were one proxy for environmental quality (a population count of clutch initiations was a second). Although only females lay eggs, both sexes care for eggs and nestlings, and the male partner's age, alone or in interaction with female age, was evaluated as a predictor of breeding performance. Middle-aged females performed better than young and old birds for all reproductive traits. Pairing with a young male delayed breeding (particularly for old females) and reduced clutch size, and pairing with an old male reduced offspring production. Challenging environments increased age effects on breeding probability and breeding date across young to middle ages and for offspring production across middle to old ages. However, important exceptions to the predicted patterns for clutch size and fledging success across young to middle ages suggested that trade-offs between fitness components may complicate patterns of trait expression across the life span. Relationships between breeding participation, environment, and individual quality and/or experience in young females may also contribute to unexpected patterns for clutch size and fledging success, traits expressed only in breeders. Finally, independent of age, breeding responses of female Nazca boobies to the ENSO did not follow expectations derived from oceanic forcing of primary productivity. During El Niño-like conditions, egg-laying traits (clutch size, breeding date) improved, but offspring production declined, whereas La Niña-like conditions were "poor" environments throughout the breeding cycle.

The Demographic Buffering Hypothesis: Evidence and Challenges

In (st)age-structured populations, the long-run population growth rate is negatively affected by temporal variation in vital rates. In most cases, natural selection should minimize temporal variation in the vital rates to which the long-run population growth is most sensitive, resulting in demographic buffering. By reviewing empirical studies on demographic buffering in wild populations, we found overall support for this hypothesis. However, we also identified issues when testing for demographic buffering. In particular, solving scaling problems for decomposing, measuring, and comparing stochastic variation in vital rates and accounting for density dependence are required in future tests of demographic buffering. In the current context of climate change, demographic buffering may mitigate the negative impact of environmental variation and help populations to persist in an increasingly variable environment.

Variation in actuarial senescence does not reflect life span variation across mammals

The concept of actuarial senescence (defined here as the increase in mortality hazards with age) is often confounded with life span duration, which obscures the relative role of age-dependent and age-independent processes in shaping the variation in life span. We use the opportunity afforded by the Species360 database, a collection of individual life span records in captivity, to analyze age-specific mortality patterns in relation to variation in life span. We report evidence of actuarial senescence across 96 mammal species. We identify the life stage (juvenile, prime-age, or senescent) that contributes the most to the observed variation in life span across species. Actuarial senescence only accounted for 35%-50% of the variance in life span across species, depending on the body mass category. We computed the sensitivity and elasticity of life span to five parameters that represent the three stages of the age-specific mortality curve-namely, the duration of the juvenile stage, the mean juvenile mortality, the prime-age (i.e., minimum) adult mortality, the age at the onset of actuarial senescence, and the rate of actuarial senescence. Next, we computed the between-species variance in these five parameters. Combining the two steps, we computed the relative contribution of each of the five parameters to the variance in life span across species. Variation in life span was increasingly driven by the intensity of actuarial senescence and decreasingly driven by prime-age adult mortality from small to large species because of changes in the elasticity of life span to these parameters, even if all the adult survival parameters consistently exhibited a canalization pattern of weaker variability among long-lived species than among short-lived ones. Our work unambiguously demonstrates that life span cannot be used to measure the strength of actuarial senescence, because a substantial and variable proportion of life span variation across mammals is not related to actuarial senescence metrics.

More frequent extreme climate events stabilize reindeer population dynamics

Extreme climate events often cause population crashes but are difficult to account for in population-dynamic studies. Especially in long-lived animals, density dependence and demography may induce lagged impacts of perturbations on population growth. In Arctic ungulates, extreme rain-on-snow and ice-locked pastures have led to severe population crashes, indicating that increasingly frequent rain-on-snow events could destabilize populations. Here, using empirically parameterized, stochastic population models for High-Arctic wild reindeer, we show that more frequent rain-on-snow events actually reduce extinction risk and stabilize population dynamics due to interactions with age structure and density dependence. Extreme rain-on-snow events mainly suppress vital rates of vulnerable ages at high population densities, resulting in a crash and a new population state with resilient ages and reduced population sensitivity to subsequent icy winters. Thus, observed responses to single extreme events are poor predictors of population dynamics and persistence because internal density-dependent feedbacks act as a buffer against more frequent events.

Immature survival, fertility, and density dependence drive global population dynamics in a long-lived species

Disentangling the influence of demographic parameters and the role of density dependence on species' population dynamics is a challenge, especially when fractions of the population are unobservable. Additionally, due to the difficulty of gathering data at large spatial scales, most studies ignore the global dynamic of a species, which would integrate local heterogeneity dynamics and remove the noise of dispersal. We developed an integrated population model (IPM) at a global scale to disentangle the main demographic drivers of population dynamics in a long-lived species. We used 28 yr of Audouin's Gull demographic data encompassing 69 local patches (comprising 90% of the world population). Importantly, we took into account the unobservable fraction of non-breeders and also assessed the strength of density dependence for this fraction of the population. As predicted by life histories of long-lived organisms, temporal random variation in survival was highest for immature individuals (1.326, 95% credible interval [CRI] 1.290-1.940) and lowest for adults (0.499, 95% CRI 0.487-0.720). Large temporal fluctuations in the probability of taking a reproductive sabbatical would partly explain the consistency in adult survival, with individuals most likely refraining from breeding when environmental conditions were harsh. Immature survival and fertility were the main drivers of population dynamics during the study period (r²  = 0.83, 0.77-0.87 and 0.73, 0.63-0.79, respectively). We found strong evidence of density dependence, not only due to the number of breeders (r²  = -0.34, -0.43 to -0.24) but also due to individuals on sabbatical (r²  = -0.18, -0.33 to -0.01). From a conservation point of view, the species shows a 5% annual global decrease during the last 10 years, and we propose an update of its conservation status. Even though population dynamics of long-lived organisms are very sensitive to changes in adult survival, we show here that, in the absence of strong environmental perturbations affecting this vital rate, fluctuations in population density are mainly driven by variations in survival of immature individuals and fertility. Integrated models based on long-term monitoring at a global scale may enhance our ecological and evolutionary understanding of how demographic drivers influence population dynamics.

Transient LTRE analysis reveals the demographic and trait-mediated processes that buffer population growth

Temporal variation in environmental conditions affects population growth directly via its impact on vital rates, and indirectly through induced variation in demographic structure and phenotypic trait distributions. We currently know very little about how these processes jointly mediate population responses to their environment. To address this gap, we develop a general transient life table response experiment (LTRE) which partitions the contributions to population growth arising from variation in (1) survival and reproduction, (2) demographic structure, (3) trait values and (4) climatic drivers. We apply the LTRE to a population of yellow‐bellied marmots (Marmota flaviventer) to demonstrate the impact of demographic and trait‐mediated processes. Our analysis provides a new perspective on demographic buffering, which may be a more subtle phenomena than is currently assumed. The new LTRE framework presents opportunities to improve our understanding of how trait variation influences population dynamics and adaptation in stochastic environments.

Demographic drivers of a refugee species: large-scale experiments guide strategies for reintroductions of hirola

Effective reintroduction strategies require accurate estimates of vital rates and the factors that influence them. The hirola (Beatragus hunteri) is the rarest antelope on Earth, with a global population size of <500 individuals restricted to the Kenya-Somali border. We estimated vital rates of hirola populations exposed to varying levels of predation and rangeland quality from 2012 to 2015, and then built population matrices to estimate the finite rate of population change (λ) and demographic sensitivities. Mean survival for all age classes and population growth was highest in the low-predation-high-rangeland-quality setting (λ = 1.08 ± 0.03 [mean ± SE]), and lowest in the high-predation-low-rangeland-quality setting (λ = 0.70 ± 0.22). Retrospective demographic analyses revealed that increased fecundity (the number of female calves born to adult females annually) and female calf survival were responsible for higher population growth where large carnivores were absent. In contrast, variation in adult female survival was the primary contributor to differences in population growth attributable to rangeland quality. Our analyses suggest that hirola demography is driven by a combination of top-down (predation) and bottom-up (rangeland quality) forces, with populations in the contemporary geographic range impacted both by declining rangeland quality and predation. To enhance the chances of successful reintroductions, conservationists can consider rangeland restoration to boost both the survival and fecundity of adult females within the hirola's historical range.

Evolution of stochastic demography with life history tradeoffs in density-dependent age-structured populations

We analyze the stochastic demography and evolution of a density-dependent age- (or stage-) structured population in a fluctuating environment. A positive linear combination of age classes (e.g., weighted by body mass) is assumed to act as the single variable of population size, [Formula: see text], exerting density dependence on age-specific vital rates through an increasing function of population size. The environment fluctuates in a stationary distribution with no autocorrelation. We show by analysis and simulation of age structure, under assumptions often met by vertebrate populations, that the stochastic dynamics of population size can be accurately approximated by a univariate model governed by three key demographic parameters: the intrinsic rate of increase and carrying capacity in the average environment, [Formula: see text] and [Formula: see text], and the environmental variance in population growth rate, [Formula: see text] Allowing these parameters to be genetically variable and to evolve, but assuming that a fourth parameter, [Formula: see text], measuring the nonlinearity of density dependence, remains constant, the expected evolution maximizes [Formula: see text] This shows that the magnitude of environmental stochasticity governs the classical trade-off between selection for higher [Formula: see text] versus higher [Formula: see text] However, selection also acts to decrease [Formula: see text], so the simple life-history trade-off between [Formula: see text]- and [Formula: see text]-selection may be obscured by additional trade-offs between them and [Formula: see text] Under the classical logistic model of population growth with linear density dependence ([Formula: see text]), life-history evolution in a fluctuating environment tends to maximize the average population size.

Contrasting effects of summer and winter warming on body mass explain population dynamics in a food-limited Arctic herbivore

The cumulative effects of climate warming on herbivore vital rates and population dynamics are hard to predict, given that the expected effects differ between seasons. In the Arctic, warmer summers enhance plant growth which should lead to heavier and more fertile individuals in the autumn. Conversely, warm spells in winter with rainfall (rain-on-snow) can cause 'icing', restricting access to forage, resulting in starvation, lower survival and fecundity. As body condition is a 'barometer' of energy demands relative to energy intake, we explored the causes and consequences of variation in body mass of wild female Svalbard reindeer (Rangifer tarandus platyrhynchus) from 1994 to 2015, a period of marked climate warming. Late winter (April) body mass explained 88% of the between-year variation in population growth rate, because it strongly influenced reproductive loss, and hence subsequent fecundity (92%), as well as survival (94%) and recruitment (93%). Autumn (October) body mass affected ovulation rates but did not affect fecundity. April body mass showed no long-term trend (coefficient of variation, CV = 8.8%) and was higher following warm autumn (October) weather, reflecting delays in winter onset, but most strongly, and negatively, related to 'rain-on-snow' events. October body mass (CV = 2.5%) increased over the study due to higher plant productivity in the increasingly warm summers. Density-dependent mass change suggested competition for resources in both winter and summer but was less pronounced in recent years, despite an increasing population size. While continued climate warming is expected to increase the carrying capacity of the high Arctic tundra, it is also likely to cause more frequent icing events. Our analyses suggest that these contrasting effects may cause larger seasonal fluctuations in body mass and vital rates. Overall our findings provide an important 'missing' mechanistic link in the current understanding of the population biology of a keystone species in a rapidly warming Arctic.

Divergent demographic strategies of plants in variable environments

One of the best-supported patterns in life history evolution is that organisms cope with environmental fluctuations by buffering their most important vital rates against them. This demographic buffering hypothesis is evidenced by a tendency for temporal variation in rates of survival and reproduction to correlate negatively with their contribution to fitness. Here, we show that widespread evidence for demographic buffering can be artefactual, resulting from natural relationships between the mean and variance of vital rates. Following statistical scaling, we find no significant tendency for plant life histories to be buffered demographically. Instead, some species are buffered, whereas others have labile life histories with higher temporal variation in their more important vital rates. We find phylogenetic signal in the strength and direction of variance-importance correlations, suggesting that clades of plants are prone to being either buffered or labile. Species with simple life histories are more likely to be demographically labile. Our results suggest important evolutionary nuances in how species deal with environmental fluctuations.