Temporal variation in demography of temperate bats: consequences for population dynamics and disease

Using a spatial, individual-based model of serotine bats, this study demonstrates that increased inter-annual demographic variation, particularly in survival rates, can drive significant population declines and reduce disease persistence, highlighting the limitations of relying solely on mean demographic rates for predicting population dynamics and disease outcomes.

The Gist

The Big Picture: Bats, Weather, and a Invisible Enemy

Imagine a colony of Serotine bats (a specific type of bat found in the UK) as a bustling, ancient city. This city has a population of about 5,000 residents living in 45 different neighborhoods. These bats are the "long-lived" type of animal—they don't have huge families, but they live a long time and take their time growing up.

The scientists in this paper wanted to answer two big questions:

1. How does unpredictable weather affect the size of this bat city?
2. How does this weather affect a virus (like a synthetic rabies) trying to spread through the city?

Usually, scientists predict the future by looking at the average weather. They say, "On average, 90% of bats survive the winter." But in the real world, weather isn't an average. It's a rollercoaster. Some years are perfect (great food, mild winters), and some years are disasters (freezing cold, no bugs to eat).

This paper asks: What happens if we stop looking at the "average" and start looking at the "rollercoaster"?

---

Analogy 1: The "Bank Account" of Survival

Think of the bat population like a bank account.

• The Average View: If you earn $100 a year on average, you think you are safe.
• The Real View: In reality, you might earn $200 one year, but then lose $150 the next year.

The researchers built a computer simulation (a "digital twin" of the bat city) to test this. They found that variability is dangerous.

Even if the average survival rate stays the same, having years that are too good and years that are too bad causes the population to crash.

• The "Bad Year" Trap: Imagine a harsh winter kills off a huge chunk of the adult bats. Because these bats are slow breeders (they only have one baby a year), the city can't just "bounce back" quickly. The loss of those experienced adults is a hole in the foundation that takes years to fill.
• The "Good Year" Myth: A perfect summer with lots of food and babies is great, but it doesn't fully fix the damage done by a terrible winter. You can't "save up" extra babies to replace dead adults later.

The Takeaway: Just because the average looks stable doesn't mean the population is safe. Increasing weather chaos (due to climate change) could cause these bat cities to shrink or even collapse, simply because the "bad years" hurt them more than the "good years" help them.

---

Analogy 2: The "Ghost in the Machine" (The Virus)

Now, imagine a ghost (the virus) trying to haunt the city. The scientists introduced a fake virus into their simulation to see how it behaved.

• In a Stable World: If the weather is always the same, the ghost finds it easy to stay around. It moves from house to house, infecting a few, then hiding, then infecting more. It becomes a permanent resident.
• In a Chaotic World: When the weather gets wild, the ghost gets confused.
  • In a "Bad Year," the bat population shrinks. The houses are empty. The ghost can't find anyone to infect, so it fades away (dies out) faster.
  • However, the scientists found something tricky: The total number of infections didn't necessarily go down. Sometimes, in a chaotic world, the virus would explode in a "Good Year" when everyone is crowded together in maternity roosts (like a high school dance), infecting hundreds at once, before dying out the next year.

The Takeaway: Unpredictable weather makes the virus's life shorter on average (it dies out sooner), but it doesn't make the virus disappear completely. It just makes the outbreaks more erratic—sometimes silent, sometimes explosive.

---

The "Adults vs. Babies" Lesson

One of the most important findings is about who matters most.

• Babies (Reproduction): The scientists thought that if the weather was bad for having babies (e.g., cold springs), the population would crash.
• Adults (Survival): They were wrong about that. The population is surprisingly good at buffering bad baby years. If you have a year with no babies, the city can wait it out because the adults are still there.
• The Real Killer: The population crashes only when the adults die.

The Metaphor: Think of the bat city as a tree.

• If the tree drops fewer leaves (babies) one year, it's fine. It will grow new ones next year.
• But if the wind snaps off the main branches (adults), the whole tree is in trouble. It can't grow new branches fast enough to replace the ones lost.

Since climate change is predicted to bring more extreme weather (heatwaves, harsh winters), the biggest threat to these bats isn't that they won't have babies; it's that the adults might die in droves.

---

Why Does This Matter?

1. For Conservation: We can't just look at the "average" bat population numbers to say, "Everything is fine." We need to look at how much the numbers bounce around. If the weather gets more chaotic, these bats are at risk of disappearing, even if they seem stable today.
2. For Public Health: Bats carry viruses that can jump to humans (zoonotic diseases). If the bat population crashes, the virus might die out. But if the population survives but becomes chaotic, the virus might still be there, waiting for the right moment to jump. Understanding these "rollercoaster" dynamics helps us predict when a disease might flare up.

Summary

This paper is a warning against relying on "averages." Nature is messy and unpredictable. For long-lived animals like bats, stability is key. If climate change turns their world into a chaotic rollercoaster, the "bad drops" will hurt them far more than the "good hills" will help them, potentially leading to population collapse and unpredictable disease outbreaks.

Technical Summary

1. Problem Statement

Estimating wildlife population trends and disease dynamics typically relies on mean annual demographic rates (survival and reproduction), assuming these rates are fixed over time. However, for temperate bat species, demographic rates are highly stochastic and influenced by environmental conditions (e.g., weather extremes, resource availability).

• The Gap: Current models often ignore inter-annual variation, potentially leading to inaccurate predictions of population persistence and disease spread, especially under climate change scenarios where environmental volatility is expected to increase.
• The Specific Context: The study focuses on the Serotine bat (Eptesicus serotinus), a species of conservation concern in the UK (listed as vulnerable) and the primary reservoir for European Bat Lyssavirus-1 (EBLV-1). The authors aim to quantify how increasing inter-annual variation in demographic parameters affects population size trajectories and the persistence of a synthetic lyssavirus-like pathogen.

2. Methodology

The authors developed a spatially explicit, individual-based model (IBM) to simulate the life history of a Serotine bat metapopulation.

• Model Structure:

  • Landscape: A 30km x 30km arena containing ~5,000 bats distributed across 45 social communities. Each community consists of a maternity roost and satellite roosts.
  • Life Cycle: Simulated in monthly time-steps. Bats progress through classes: pup, juvenile, immature, and reproductively mature.
  • Behavior: Includes seasonal hibernation (Oct–Mar) and active periods (Apr–Sep). Dispersal is male-biased (100% of juvenile males disperse), while females show natal philopatry.
  • Demography: Survival and reproduction are influenced by age, breeding status, season, and an annual "quality" factor ($Q$).
    • Good years: High survival/reproduction.
    • Bad years: Low survival/reproduction.
  • Calibration: Parameters were fitted using Latin Hypercube Sampling against long-term capture-mark-recapture (CMR) data from a single UK site and national monitoring data (Bat Conservation Trust). The model was calibrated to produce a stable population with realistic sex ratios (60:40 male:female) and breeding proportions.

• Disease Simulation:

  • A synthetic SEIR (Susceptible-Exposed-Infectious-Recovered) model representing a lyssavirus-like pathogen.
  • Pathology: Latent periods pause during hibernation. Infection can be fatal (clinical) or sub-clinical (leading to temporary immunity).
  • Transmission: Contact rates vary by roost type (high in maternity roosts, low in satellite roosts). Transmission occurs via local contact and dispersal of infected juveniles.

• Experimental Scenarios:

  • The model tested increasing levels of inter-annual variation ($\sigma$) in demographic parameters, ranging from 0 (mean rates only) to 1 (high variability).
  • Scenarios included: (1) All parameters co-varying, (2) Survival parameters varying only, and (3) Reproductive parameters varying only.
  • Sensitivity Analysis: A global sensitivity analysis (LHS-PRCC) was conducted to identify which parameters most influenced population change and disease persistence.

3. Key Contributions

• Methodological Innovation: The integration of a spatially explicit IBM with a synthetic disease model specifically parameterized for Serotine bats, allowing for the disentanglement of demographic stochasticity from environmental stochasticity.
• Decoupling Survival and Reproduction: The study explicitly separates the effects of variation in survival versus reproduction, challenging the assumption that both drive population dynamics equally in K-selected species.
• Disease-Demography Interaction: It demonstrates how demographic fluctuations alter the "force of infection" and the likelihood of disease fade-out in a metapopulation context, rather than just a single closed population.

4. Key Results

A. Population Dynamics

• Impact of Variation: Increasing inter-annual variation led to a substantial decline in mean population size over time. While some simulations showed growth, the probability of catastrophic population crashes increased significantly.
• Survival vs. Reproduction:
  • Survival is the primary driver: Variation in adult survival rates was the dominant factor causing population decline.
  • Reproduction is buffered: Variation in reproductive rates (fecundity, primiparity) had minimal impact on mean population size. The species' long lifespan allows it to buffer fluctuations in reproductive success, provided adult survival remains high.
• Climate Change Implication: As climate change increases environmental volatility (e.g., heatwaves, harsh winters), the risk of population collapse rises, even if mean conditions remain favorable.

B. Disease Dynamics

• Reduced Persistence: Increasing demographic variation reduced the duration of disease persistence. In high-variation scenarios, the disease went extinct (fade-out) more frequently due to population bottlenecks in "bad" years, which lowered contact rates below the threshold for transmission.
• Stochasticity: Disease outcomes were highly variable between simulations. Persistence relied on stochastic events, such as the dispersal of a latently infected juvenile to a new community.
• Total Infections: Despite shorter persistence durations, the total number of infected bats over the simulation period remained relatively stable, as "bad" years with low persistence were sometimes offset by "good" years with high transmission spikes.

C. Sensitivity Analysis

• Population Size: Highly sensitive to uncertainty in adult survival. Reproductive parameters (fecundity, primiparity) were less influential.
• Disease Persistence: Also most sensitive to adult survival. Interestingly, the rate of transition from non-breeder to breeder (primiparity) was also influential; higher transition rates increased disease persistence by moving more individuals into high-contact maternity roosts.

5. Significance and Implications

• Limitations of Mean Rates: The study concludes that relying solely on mean demographic rates is insufficient for predicting future trends. Models must account for the variance in these rates, particularly for long-lived species where adult survival is critical.
• Conservation Strategy: Conservation efforts for temperate bats must prioritize mitigating extreme mortality events (e.g., protecting roosts from extreme weather) rather than just focusing on reproductive success.
• Public Health: While high environmental variation may reduce the long-term persistence of lyssaviruses by causing population crashes, the stochastic nature of transmission means outbreaks can still occur. Understanding the metapopulation structure and dispersal is crucial for predicting zoonotic spillover risks.
• Future Directions: The authors suggest linking demographic parameters directly to weather forecasts and incorporating spatial heterogeneity (source-sink dynamics) to better predict range shifts under climate change.

In summary, this paper provides a robust framework demonstrating that environmental volatility acts as a double-edged sword: it threatens the persistence of bat populations through increased mortality but simultaneously disrupts the persistence of their associated pathogens by breaking transmission chains.