Discussion
Our results show that, in our experimental microcosms, increased connectivity can buffer a population from extinction caused by drought. Specifically, habitats connected by good quality corridors, not only enabled populations to persist longer when they are threatened by drought severity, but slowed down the rate at which the number of patches across landscapes affected by drought decreased the populations. Surprisingly, connectivity differently affected the survival of populations in each individual patch, as the variability of patch abundance increased or slightly decrease then followed by a sudden decrease in the habitats connected by poor quality corridors, whilst it constantly decreased in the good quality corridors. In what follows, we first discuss the limitations of our approach, and then discuss the results in the context of the literature.
There are two main limitations to our approach. The first is that we did not count individuals in the corridors, as corridors can, potentially at least, provide extra habitats and serve as temporal refugia for species when stressed (Eversham and Telfer 1994). The aim of the work was to investigate how patchy populations response to environmental stressors, thus the individuals dispersing in corridors are largely ignored. The second is the potential confounding effect introduced the set up of the experiment which may explain some of the observed patterns in the abundance dynamics between treatments (Fig 2). This may be due to systematic differences in the body size distributions of the populations in different treatments, however the extinction dynamics largely conform to the expected patterns.
Species are expected to seek new habitats when natal habitats deteriorate. Field studies showed that Collembola have a high dispersal ability for new habitats when experiencing dry conditions (e.g., Ferrin et al. 2022). Consequently, increasing movement between habitats may save populations from local extinction when patches experience inhospitable environmental conditions. Our previous work showed that Collembola were more likely to colonise new habitats connected by good quality corridors, thereby increasing population growth rate in the colonised patches (Li et al. 2021). In this regard, corridor quality corridors may increase their probability of survival by facilitating movements and avoiding dry habitats. Collectively, we demonstrated that good quality corridors reduced the rate of population decline and increased time before extinction caused by drought extremes, highlighting the importance of connectivity in persisting populations in the face of climate change. This not only agrees with previous studies which show that increasing connectivity using corridors can increase the persistence of metapopulations (Gonzalez et al. 1998, Swart and Lawes 1996), but proves their ability to slow down extinction when species are exposed to disastrous environmental conditions. Our results suggest that increasing spatial connectivity across a landscape can to some extent mitigate the negative impact of climatic extremes on populations.
The most striking results were that the variability of patch abundance was negatively related with increased corridor quality, suggesting that connectivity can play an important role in shaping the viability of patchy populations. Thus, when corridor quality was poor, Collembola were constrained in patches as their dispersal ability was limited. Intriguingly, this implies that when connectivity is good the increased migration of individuals around the landscape could put increased pressure on habitats to support large population sizes, increasing the impacts of density dependence if such patterns persist in the long-term. Conversely, individuals being stuck in a patch could cause uneven survival rates among patches when drought extremes occur unpredictably. In such settings, local extinctions were more likely to occur when some patches were turned into ecological traps due to environmental change (Hale and Swearer 2016), thereby increasing the variability of patch abundance in habitats with poor habitat connectance. Similar findings have been previously reported in freshwater systems where drought disrupted the connectivity of habitat pools for some fish species (e.g., Vander Vorste et al. 2020). Furthermore, the changes in variability of patch abundance over time indicates a contrasting pattern of stressed populations towards extinction between habitats with good and poor connectivity. Specifically, a constant decrease of the variability found in well-connected populations indicates that the density of population in each patch became uniform when experiencing stress, whilst a large variability found in poor connected populations demonstrated an increasing risk of local extinction because of loss of connectivity. Our results suggest that the variability of patch abundance may be a good predictor of population status and monitoring how it changes over time might provide a useful guide to evaluate landscape connectivity for species.
Drought is a climatic extreme which has serious consequences for the persistence of metapopulations, seen here as an increasing drought severity significantly reducing the time that population persisted and increasing the rate of population decline, a phenomenon which is consistent with the prediction that severer drought would cause higher speed of extinction (Cady et al. 2019, Cayuela et al. 2016). Increasing the severity of drought is more likely to cause a strong effect of desiccation, causing a high rate of mortality when reaching species’ limits to drought. Collembola are well adapted in humid soil environments and need to absorb water vapor to over their entire life cycles (Bayley and Holmstrup 1999). Some essential life stages such as reproduction cycles and egg incubation in Collembola are highly depended on moisture levels (Holmstrup 2019, Waagner et al. 2011). Meanwhile, young individuals of Collembola are more vulnerable to desiccation than old (Hilligsoe and Holmstrup 2003), increasing drought severity are more likely to reduce the fitness of young adults and cause a further loss of fecundity. This may explain the extinction event occurred at the end of our experiment, and the extinction happens faster with increased severity. Indeed, extreme environmental stressors such as drought has been shown shaping the population dynamics and persistence of many species (e.g., Johansson et al. 2020).
The amount of the favourable habitats within a landscape has been shown to be important for maintaining the persistence of populations (Dytham 1995, Meli et al. 2014). Surprisingly, we found that increasing the number of patches affected by drought at different speeds only negatively affect the maximum rate of population decline, rather than the timing of extinctions, which may suggest a tipping point for environmental conditions as a result of increased drought habitats destabilising populations. The reasons underlying this pattern are largely unexplored though. One possible explanation is that Collembola were able to accumulate sugar and polyols when experiencing dry conditions, temporarily prolonging their survival from dehydration (Fountain and Hopkin 2005, Waagner et al. 2012). Hence, it is possible that individuals living in a habitat patch under drought manipulation (i.e., no water added for a week in our case) might not experience an immediate local extinction due to the accumulation of polyols and sugars, though drought was shown suppressing the viability of populations based on our results (Fig. 2). Adjusting their body conditions may allow them to temporarily survive over a short drought period, however, it may also contribute to a fast decline when the environmental conditions reach to tipping points (Dai et al. 2012). More empirical evidence is needed to examine how increased environmental stress, associated with tipping points, impacts the resilience of spatially structured populations.
In summary, our analyses add to the growing literature describing how increasing connectivity among habitats can buffer metapopulations against population decline and extinction. However, we also show that increased connectivity may increase the variability of populations between patches in a landscape, as individuals can move more easily into patches with more favourable conditions, potentially increasing the effects of density dependence in these highly utilised patches. As habitat fragmentation creates a mosaic of landscapes hosting a variety of spatially structured populations, maintaining good spatial connectivity allows species to move out the dangerous area, and/or allows conservation actions to be taken to avoid mass extinctions when facing extreme climatic events. Ultimately, our work highlights the importance of habitat connectivity in maintaining population viability in the face of climate change. Field evidence is needed to better understand the role of habitat connectivity in persisting populations.