Effects of directional environmental change on extinction dynamics in experimental microbial communities are predicted by a simple model

Abstract

Global temperatures are expected to rise between 1.1 and 6.4°C over the next 100 years, although the exact rate will depend on future greenhouse emissions, and will vary spatially. Temperature can alter an individual's metabolic rate, and consequently birth and death rates. In declining populations, these alterations may manifest as changes in the rate of that population's decline, and subsequently the timing of extinction events. Predicting such events could therefore be of considerable use. We use a small-scale experimental system to investigate how the rate of temperature change can alter a population's time to extinction, and whether it is possible to predict this event using a simple phenomenological model that incorporates information about population dynamics at a constant temperature, published scaling of metabolic rates, and temperature. In addition, we examine 1) the relative importance of the direct effects of temperature on metabolic rate, and the indirect effects (via temperature driven changes in body size), on predictive accuracy (defined as the proximity of the predicted date of extinction to the mean observed date of extinction), 2) the combinations of model parameters that maximise accuracy of predictions, and 3) whether substituting temperature change through time with mean temperature produces accurate predictions. We find that extinction occurs earlier in environments that warm faster, and this can be accurately predicted (R 2  0.84). Increasing the number of parameters that were temperature-dependent increased the model's accuracy, as did scaling these temperature-dependent parameters with either the direct effects of temperature alone, or with the direct and indirect effects. Using mean temperature through time instead of actual temperature produces less accurate predictions of extinction. These results suggest that simple phenomenological models, incorporating metabolic theory, may be useful in understanding how environmental change can alter a population's rate of extinction. Synthesis

Multiple environmental changes interact to modify species dynamics and invasion rates

“…Unconsidered interactions between many environmental changes may prove problematic for the development of predictive frameworks mechanistically linking environmental changes to species dynamics. Metabolic theory, for example, has shown promise, in tightly controlled laboratory microcosms, as a method for scaling species feeding rates, interaction strengths and extinction risks with temperature (Petchey et al , O'Connor et al , Rall et al , , Clements et al ). However in natural systems, these scaling rates may vary continuously as other interacting facets of the environment change.…”

Big answers from small worlds: a user's guide for protist microcosms as a model system in ecology and evolution

“…; Clements et al . ), sampling the entire microcosm is highly desirable. This can be achieved by using a vessel with a transparent bottom that can be placed directly under a dissecting microscope.…”

Crossing regimes of temperature dependence in animal movement

“…At the organismal level, temperature can affect the behavior of ectotherms (Davis, 1989;Turner et al, 1993;Long et al, 2012) and alter the body size and body condition of both ectotherms and endotherms (Atkinson, 1994;DeLong, 2012). Changes in body size and condition due to temperature can dramatically affect population dynamics (Vasseur & McCann, 2005;Ims et al, 2008;Johnson et al, 2010;Amarasekare, 2015) by altering per-capita growth rates (Frazier et al, 2006;Jandricic et al, 2010;Amarasekare & Savage, 2012) and genetic diversity (Pauls et al, 2013), leading to an increase in extinction risk (Clements et al, 2014). Additionally, changes in temperature can result in changes in the structure and stability of entire communities (Brose et al, 2012) through their effects on interaction strengths (Rall et al, 2010) and food web body size structure (Albouy et al, 2014;Gibert & DeLong, 2014).…”