Sand lizard (Lacerta agilis) phenology in a warming world

Phenological shifts have been observed in many species and populations in response
to contemporary changes in the global climate and this is predicted to continue as
climate change proceeds 1]–3]. The hitherto observed consequences of these changes are variable and ultimately
depend on how well organisms are able to track climatically induced shifts in optimal
phenotype, thereby avoiding a loss of fitness. In this study we explored the effect
of spring temperature on oviposition timing in a high latitude population of sand
lizards as a first step towards understanding the potential impacts of climate change
on taxon-specific phenology. Other long-term studies of wild reptile populations have
reported advancements in time of reproduction in warmer years, such as earlier parturition
dates of the aspis viper (Vipera aspis) 21], the spotted snow skink (Niveoscincus ocellatus) 37] and the common lizard (Zootoca vivipara) 38], and earlier breeding of the sleepy lizard (Tiliqua rugosa) 20]. Testing for population-level plasticity, we investigated how increasing spring temperature
is likely to affect egg-laying date in this population in the near future. In addition,
we also explored potential long-term effects by testing for among-individual variation
in the average response and plasticity, a prerequisite for adaptive evolution. One
other study, by Schwanz and Janzen 39], has looked at individual variation in plasticity in a phenological trait in a reptile.
They investigated the potential effects of climate change on painted turtles (Chrysemys picta) by examining whether individual plasticity in the timing of nesting has the capacity
to offset sex ratio effects due to rising temperatures.

Our analyses showed that temperatures over the reproductive period have a significant
effect on oviposition date in this population; female sand lizards laid eggs earlier
in relatively warmer years. This means that the females respond to fairly rapid (annual)
changes in ambient thermal conditions, and hence that there is population-level plasticity
in laying date. Within cohorts, earlier egg laying has previously been shown to enhance
offspring fitness and survival in these lizards 29], as in reptilian offspring at large (e.g., 30]–33]), which suggests that higher spring temperatures benefit components of fitness in
this population, and hence that global warming may involve positive effects short
term. Although contrasting with projections by Sinervo et al. 26], showing a global increase in extinction risk for lizards as a result of climate
change, other studies of wild lizard populations have also presented positive predictions
(see 6] for a summary). For example, Cadby et al. 37] reported earlier hatching in warmer years in two populations of the spotted snow
skink (Niveoscincus ocellatus), which is related to increased survival in the first year of life 33]. In the common lizard (Lacerta vivipara), rising temperatures led to larger body size in female neonates and adults over
an 18 year period, with a concomitant increase in reproductive output 30]. Specifically, in our population of L. agilis, higher temperature during the reproductive period has been associated with higher
mating rates and number of sires per clutch, leading to increased sperm competition
with positive effects on offspring survival 13]. As these findings only concern components of individual fitness, studies on the
whole demography of these species might, however, yield different projections. An
important selective force driving trait evolution is natural selection, which acts
on among-individual differences in fitness-related traits. According to our results,
individual females vary significantly in their average response to spring temperature
when differences in laying date associated with body size and random variation due
to annual fluctuations in environmental conditions are taken into account. Directional
selection on oviposition date has previously been demonstrated for this population
29], hence, if this variation has a genetic component, an evolutionary advancement in
average egg-laying date may take place as spring gets warmer. In contrast, we found
no statistical evidence for among-individual variation in trait plasticity (individual
reaction norm slopes). This result could reflect a true absence of variation in this
trait, implying that all females respond to annual fluctuations in spring temperature
in a similar way. However, another potential explanation for this apparent ‘null’
result is a lack of statistical power to detect individual variation in plasticity.
Our data set contained 131 females with ?2 observations, hence contributing to variation
in reaction norm slopes, but 77 of these females were observed only twice, which limits
the robustness of our slope estimation. A lack of variation in laying-date plasticity
could be due to strong stabilizing selection on the shape of reaction norms as mis-timing
reproduction at this high latitude location can have drastic consequences 40]. This would constrain evolution towards new reaction norm optima associated with
an altered climate, however, exposure to environmental conditions outside those previously
experienced by a population, under which selection has not yet had an opportunity
to act, could release hidden phenotypic and genetic variation, and thereby allow for
an evolutionary change 41].

Interestingly, significant variation among individuals in their plastic response was
obtained when modelling within-individual residuals as homogeneous across years. This
agrees with findings of several other studies that also considered both approaches
36], 37], 42]–46], and confirms concerns raised that ignoring the possibility for heterogeneity in
residuals may lead to erroneous conclusions (e.g., 35], 36]). Patterns in residual variances can have important implications for evolutionary
trajectories of phenotypic traits as they may cause estimates of heritability to change
across environments (e.g., 36], 44], 47]). Exploring how environmental conditions affect residual variances may therefore
improve our understanding of how organisms respond to long-term directional climate
change. Our results show that in some years, egg-laying date deviated more from the
individual’s estimated reaction norm than in others, indicating heterogeneous among-year,
within-individual residuals. Heterogeneity in residual variances across space or time
is commonly associated with differences in environmental quality (e.g., 36], 42], 43], 47]), but we found no such associations. We speculate that timing of oviposition in this
species is a function of multiple environmental factors, and hence that heterogeneous
residuals reflect variation in their interactions and correlations among years 48].

Thus far, our assessment of how a warming climate is likely to affect this northern
population of lizards has solely been based on potential effects on oviposition timing.
Climate change is, however, likely to affect a whole suite of traits, having a combined
effect on viability and, ultimately, fitness. Most studies to date investigating the
vulnerability of lizards to a warming climate focus on sensitivity and adaptability
of thermal physiology and behavioural thermoregulation (e.g., 23]–26], 49]). However, as reproductive success is a prerequisite for the long-term persistence
of a population, potential effects on reproductive traits should also be considered.
Furthermore, several recent studies of conservatism of lizard thermal physiology indicate
that thermal tolerance is conserved across lineages, suggesting a limited potential
for local evolutionary adaptation 50]–52]. In contrast, intra-specific divergence in response to variation in local thermal
environment has been reported for a number of reproductive traits, including age and
size at maturity 53], 54], timing of ovulation and parturition 53], 40], and sex-determination system 55]. This argues for the importance of taking reproductive traits into account when assessing
population sensitivity to climate change.

In summary, our results suggest that a rise in spring temperature (such as associated
with climate change) may have positive fitness effects in this northern population
of sand lizards by allowing for an advancement of oviposition. Our analyses revealed
both population-level plasticity and individual-level variation in average laying
date, hence, this prediction is consistent over shorter and potentially also longer
time scales. In contrast, evolution of laying-date plasticity towards new trait optima
may be constrained by a lack of among-individual variation. To verify that an evolutionary
response to selection is possible, the genetic basis for the observed individual differences
should be investigated using an ‘animal model’ approach. That is, a linear mixed model
using pedigree-information to estimate genetic variances and covariances of adaptive
phenology 56], 57].