Technical Comments

Comment on “Erosion of Lizard Diversity by Climate Change and Altered Thermal Niches”

See allHide authors and affiliations

Science  29 Apr 2011:
Vol. 332, Issue 6029, pp. 537
DOI: 10.1126/science.1195193

Abstract

Using a regionally calibrated model, Sinervo et al. (Reports, 14 May 2010, p. 894) predicted potential climate change impacts on lizard populations and estimated that many extinctions are under way. We argue that this model is not sufficient for predicting global losses in lizard species in response to anthropogenic climate change.

Sinervo et al. (1) recently used a thermal model that incorporates mean body temperature (Tb) and mean maximum environmental temperature (Tmax) to predict the cumulative hours during which lizards cannot maintain activity, or the hours of restriction (hr = time spent above preferred body temperature, Tp), a variable that strongly affects population persistence and extinction probability. The model was validated with a data set including operative temperatures (Te) for one species, and Tmax, Tb, and known local extinctions of several Sceloporus populations in México. Global extinctions were then predicted assuming that the relationship between Tb, Tmax, and hr for Sceloporus spp. was suitable for other species and geographic regions. Although we agree that this approach highlights the potential effects of heat excess on populations, and thus the negative consequences of climate warming, we argue that further refinements for species groups and regions should take place before the model is extrapolated to a global scale.

The determination of hr is dependent on operative temperature availability in time and space. Therefore, accounting for the breadth (high or low variance) and shape (skewness) of the Te frequency distribution is essential for forecasting hr (2). For example, in two environments with the same mean Te but different Te breadths (Fig. 1A), the high-variance scenario (blue) would result in a lower hr simply owing to the increased range of thermal opportunities. Moreover, the skewness of the Te distribution indicates where most thermal opportunities occur and the range of infrequent Tes on either end of the distribution. Therefore, in the example shown in Fig. 1B, hr is greater for a right-skewed (red) than a left-skewed (blue) Te distribution, given the limited availability of optimal sites during much of the day in the right-skewed environment. In skewed distributions, the median, rather than the mean Te, is the more appropriate parameter for estimating hr.

Fig. 1

Hypothetical scenarios of daily frequency distributions of operative temperatures (Te) available to lizards in two environments (red and blue). Each bar represents 50% of the data, and whiskers are the remaining 50%. In (A), the blue environment has the same mean but higher variance of Te compared with the red environment, resulting in an increased period in which the preferred body temperature (Tp) overlaps with Te. In this case, lizards in the red environment are highly constrained by warm temperatures in the middle of the day compared with lizards in the blue environment. In (B), Te range is the same, but the shapes of Te distributions differ between the red (right-skewed) and blue (left-skewed) environments. In this scenario, although Tp overlaps with Te in both the red and the blue environments, lizards in the blue environment have increased availability of favorable Tes to maintain Tp during the warmest periods of the day compared with the red environment.

The distribution of Tes available for thermoregulation is strongly dependent on habitat type (sand, rock-outcrops, or vegetation) and its heterogeneity. For example, open-sky habitats with limited shade or refuges will generally have a wide diurnal temperature range with a frequency distribution strongly skewed toward high temperatures during daylight hours. In savannas and forests, the distribution of Te will be strongly influenced by the diversity and quantity of microsites (e.g., the number and size of shrubs that offer shade in savannas or the size of canopy gaps in dense forests). Moreover, the spatial arrangement of different microsites and the degree of complexity in the landscape will also determine the distance that lizards need to travel to exploit favorable Tes (3, 4). These factors will therefore influence the costs of maintaining Tp, such as those associated with increased predation risks and energy expenditure, thus affecting hr. Macroclimate data and energy-balance equations that include properties of the organism and of the environment (57) can provide the kinds of Te frequency distributions that would improve estimates of hr for different organisms in different environments (8).

Given that the projections made by Sinervo et al. (1) are for a wide range of habitat types incorporating much variation in the form of the frequency distributions of Te, it is highly likely that the estimates of hr will be in error in many instances. Similarly, the range of Tb (stenothermic versus eurythermic species) and the shape of the Tb frequency distribution (as opposed to mean Tb) will affect potential activity periods (8, 9). Therefore, the use of equation S2 in (1) as a general formula for predicting extinctions is inappropriate in its current form and needs to be reevaluated by taking into account Te and Tb variation for a variety of habitats and taxonomic groups.

Accounting for breadth and skewness of temperature distribution data (both spatially and temporally) and environmental constraints to thermoregulation should be a priority when forecasting mortality risks of ectotherms facing global warming. The inclusion of such measures would do much to provide better estimates of uncertainty for predictions of the kind made by Sinervo et al.

References

View Abstract

Navigate This Article