By Isabel Smallegange & Matty Berg. Originally published in Amsterdam Science 11(3).

Climate variability is increasing. How will this affect different plant and animal species? The answer to this question is important to inform conservation strategies. Our research shows that we should not rely solely on big-data research to find the answer. Instead we should consider the mechanistic underpinnings of biological variation as a starting point when we want to extrapolate species responses to future environmental changes.

In our analysis, we focus on the so-called life-history speed or ‘pace of life’. The life-history speed of animals and plants ranges from slow to fast. Slow life-history species are typically characterised by late maturation, long life span and not many offspring. Fast life-history species have the opposite characteristics. The pace of life of species plays a crucial role in how they respond to variations in their environment. Environments can change in a myriad of ways, but one that is receiving increasing interest is the way in which the autocorrelation of environmental conditions changes over time. In highly autocorrelated environments, the state of the environment in the near future is strongly related to the current environmental state (Fig. 1, right graph). Hence, it is predictable for the organism. In uncorrelated environments, there is no correlation between environmental states at different moments in time, leading to unpredictability (Fig. 1, left graph).

Analyses of life-history data from hundreds of plant and animal populations using big-data approaches have revealed that species on the fast end of the life-history speed continuum are more sensitive to shifts in temporal autocorrelation than slow life-history species [1]. These analyses rely on big biodiversity databases comprising data on species diversity, abundance and traits in order to identify general patterns of how the life-history speed of species determines their response to environmental changes. Typical of big-data approaches, however, the species’ life histories in these analyses are phenomenological descriptions that lack a mechanistic representation of the biological processes that give rise to observed life-history variation. This means that, without a mechanistic underpinning to data patterns, we do not understand why and how species react to novel environmental conditions. Therefore, extrapolating beyond the range of existing data is problematic.

We set out to explore in simulations how a fast life-history species, the beach hopper Orchestia gammarellus (Fig. 2, left), and a slow life-history species, the reef manta ray Mobula alfredi (previously Manta alfredi) (Fig. 2, right), respond to changes in food conditions [2]. We found that the beach hopper responds strongly to how often food conditions are favourable, whereas the manta ray responds strongly to how predictable food conditions are. These results are opposite to those found in big-data analyses based on phenomenological descriptions of life histories [1].

Our simulations have a mechanistic underpinning based on individual energy budgets; hence, we can explain why the two species react differently. The population growth rate of the beach hopper is very sensitive to perturbations to individual reproduction. As rich food conditions directly affect individual reproduction, any shift in how often food conditions are favourable will have direct consequences for population growth. In contrast, the population growth rate of the reef manta ray is very sensitive to perturbation of individual growth, which impacts the population growth rate only indirectly. A shift in how often conditions are favourable has much less of an effect on the population growth of the slow life-history species than a shift in the predictability of food conditions.

Our findings highlight the importance of focusing on the mechanistic underpinnings of biological variation as a starting point for extrapolation of species responses to novel (climate) change. While big-data research methods are increasingly used to tackle complex eco-societal problems, we should not discard conventional scientific methods of inquiry. The empirical cycle starts with collecting data, but the purpose of such data is to inform theories, not to be a method in itself.

---

References

[1] M. Paniw, A. Ozgul and R. Salguero‐Gómez. Ecology Letters 21, 275–286 (2018). doi: 10.1111/ele.12892

[2] I.M. Smallegange and M. Berg. Ecology and Evolution 9, 9350-9361 (2019). doi: 10.1002/ece3.5485

---