Applying Metabolic Scaling Laws to Predicting Extinction Risk

By Claire Asher, on 25 September 2014

The Earth is warming. That much were are now certain of. A major challenge for scientists hoping to ameliorate the effect of this on biodiversity is to predict how temperature increases will affect populations. Predicting the responses of species living in complex ecosystems and heterogenous environments is a difficult task, but one starting point is to begin understanding how temperature increases affect small, laboratory populations. These populations can be easily controlled, and it is hoped that the lessons learned from laboratory populations can then begin to be generalised and applied to real populations. Recent research from GEE academics attempted to evaluate the predictive power of a simple metabolic model on the extinction risk of single-celled organisms in the lab. Their results indicate that simple scaling rules for temperature, metabolic rate and body size can be extremely useful in predicting the extinction of populations, at least in laboratory conditions.

Current estimates suggest that over the next 100 years we can expect a global temperature rise of between 1.1°C and 6.4°C. This change will not be uniformly distributed across different regions however, with some areas expected to experience warming at twice the global average rate. Temperature is known to be a crucial component in some of the most basic characteristics of life – metabolism, body size, birth, growth and mortality rates. These characteristics have been shown to scale with temperature in an easily predictable way, formalised in the Arrhenius equation. This equation yields a roughly 3/4 scaling rule, so that as temperature increases, metabolism increases around 75% as fast. This relationship appears to hold true for a variety of taxa with different life histories and positions in the food chain. Models based upon this rule can be designed that are very simple, which makes it easy for scientists to collect the data needed to plug into the model. But are they accurate in predicting extinction?

Recent research conducted by GEE and ZSL academics Dr Ben Collen and Prof. Tim Blackburn, in collaboration with the University of Sheffield and The University of Zurich, investigated the predictive power of simple metabolic models on extinction risk in a single-celled protist Loxocephallus. They first collected data on the population and extinction dynamics of a population held at constant temperature. This data was fed into a model based on scaling laws for metabolic rates and temperature, which in turn attempted to predict extinction risk under different temperature changes. The researchers tested how real protists responded to temperature changes – for 70 days they monitored populations of the protist Loxocephallus under either decreases or increases in temperature. Populations began at 20°C and increased to 26°C or decreased to 14°C at different rates (0.5°C, 0.75°C, 1.5°C or 3°C each week). Most populations eventually went extinct, but these extinctions happened sooner in hotter environments, and mean temperature showed a strong correlation with the date at which the population went extinct. Extinction tended to happen sooner in populations subjected to more rapid warming.

None of this is particularly surprising, but what the researchers found when they ran their models was that, even with relatively minimal data to start out with (population dynamics under constant ‘normal’ conditions), and using only simple scaling laws to predict extinction, their model was able to accurately predict when populations would go extinct under different warming or cooling conditions, with an accuracy of 84%. One important factor was the specifics of the temperature changes that were input into the model – using average temperature across the experiment rather than actual temperature changes produced much less accurate results.

This research is a first step in creating models that may help us predict the future extinction dynamics of wild populations subjected to unevenly distributed climatic warming over coming decades. It is a long way from a simple model of a laboratory population to a model that can accurately predict the future of complex assemblages of wild animals that are also subject to predation, disease and a healthy dose of luck. But the fact that these models can work for simple systems in laboratory conditions is a great first start – if they didn’t work for these populations, we could be fairly sure they wouldn’t generalise to natural populations. This shows that simple phenomenological models based on basic metabolic theory can be useful to understand how climate change will effect populations.

Original Article:


This research was made possible by funding from the Natural Environment Research Council (NERC).

Predicting Extinction Risk:
The Importance of Life History and Demography

By Claire Asher, on 28 July 2014

The changing climate is no longer simply a concern for the future, it is a reality. Understanding how the biodiversity that we share our planet with will respond to climate change is a key step in developing long-term strategies to conserve it. Recent research by UCL CBER’s Dr Richard Pearson identifies the key characteristics that are likely to influence extinction risk due to climate change, and shows that existing conservation indicators such as the IUCN red list may contain the data necessary to make these predictions.

Human activities have been negatively impacting biodiversity for centuries, and conservationists have developed a number of different indicator lists which attempt to classify species’ extinction risk. However, these lists were created to measure human impacts such as as habitat loss, hunting and introduction of invasive species. These impacts will continue to be a major issue for biodiversity, but may be dwarfed in the future as climate change takes hold. Can the indices and data we already have be used to predict extinction risk from climate change? Or does climate change represent a new type of threat, needing new indices?

Studies have previously identified the ecological and biological traits that are characteristic of threatened or declining species. However, it is not clear how well these traits predict the future risk of climate-induced extinction. In February this year, GEE’s Dr Richard Pearson, in collaboration with colleagues at the American Museum of Natural History, Stony Brook University and the University of Adelaide, published a paper in Nature which attempted to address these questions. Most studies that have considered the impact of climate change on species’ extinctions have attempted to predict changes in the distribution of suitable habitats and measure extinction risk in terms of whether the species is likely to be able to find habitat to live in. However, such studies rarely consider how a species’ traits such as life history and spatial characteristics will influence their ability to persist through changes in climate. In this study, Pearson and colleagues coupled ecological niche models with demographic models, and developed a generic life history method to estimate extinction risk over the coming century.

Modelling Extinction
The authors then tested their models on ecological and spatial traits for 36 reptile and amphibian species in the USA. Using commonly available life history variables, they found that their models could accurately predict extinction risk between 2000 and 2010. They then utilised the same traits and models to predict future extinction risk under two climate models – a high emissions scenario and a policy scenario aimed at curbing emissions. Average extinction risk for the 36 species studied was 28% under the high emissions scenario, dropping to 23% under strict policy intervention. This seems like a very small difference for a significant intervention – it’s important to note that the same estimates indicated an average extinction risk of just 1% in the absence of any climate change at all.

One of the most important determinants of extinction risk in reptiles and amphibians was occupied area, which represents the range of climatic and habitat conditions the species can survive in. Species with a larger occupied area tended to be more robust to climate change, presumably because they are already adapted to a wider range of habitats and climates. Other key variables influencing extinction risk include population size and generation length. In many cases, traits interacted to determine species risk, for example extinction risk was strongly influenced by interactions between occupied area and generation length. Including many different traits can therefore greatly improve the accuracy of predictions. Recent trends tended to be less informative than spatial, demographic and life history traits, particularly under the high emissions scenario, suggesting that the impacts of climate change we have observed so far are likely to become less and less relevant as climate change accelerates.

The majority of variables that showed a significant impact on extinction risk are already included in major conservation assessments and indices, meaning that data and monitoring programs already in place may be better at predicting extinction risk under climate change than we might have expected. Climate change may not be fundamentally different from other human threats such as habitat loss and hunting, at least in terms of our ability to assess extinction risk. Conservation initiatives should focus on species who currently occupy a small and declining area and have a small population size. Regardless of the policy future, conservation actions will need to consider and account for climate change if they are to prove effective.

Original Article:

() Life History and Spatial Traits Predict Extinction Risk Due to Climate Change Nature


This research was made possible by funding from the National Aeronautics and Space Administration (NASA) and the Australian Research Council