### Post provided by Delphine Chabanne

Wildlife isn’t usually uniformly or randomly distributed across land- or sea-scapes. It’s typically distributed across a series of subpopulations (or communities). The subpopulations combined constitute a metapopulation. Identifying the size, demography and connectivity between the subpopulations gives us information that is vital to local-species conservation efforts.

**What is a Metapopulation?**

Richard Levins developed the concept of a metapopulation to describe “a population of populations”. More specifically, the term metapopulation has been used to describe a spatially structured population that persists over time as a set of local populations (or subpopulations; or communities). Emigration and immigration between subpopulations can happen permanently (through additions or subtractions) or temporarily (through the short-term presence or absence of individuals).

**Mark-Recapture Approach to Characterise Dynamics between Subpopulations**

For animals that we can identify individually (by things like permanent marks, fin shape, color patterns or tags for example) a mark-recapture approach can be used to investigate things like abundance and survival of multiple subpopulations. But the sampling of individuals from multiple subpopulations can make it more likely that some individuals are captured or seen more than once. Because of this, the population structure has to be considered when developing a sampling design for any study.

There are a few different things that can cause heterogeneity in individual capture probabilities (e.g. individuals seen more or less than others) associated with spatial distribution. A couple of the most common ones are particular environmental features (e.g. different habitats) and social sub-groups with little or no overlap between home ranges. When heterogeneity in capture probability occurs, the estimates of survival can be biased. Even small, the bias may have a great impact on what and how decisions for management will be made.

In our recently published research in *Methods in Ecology and Evolution *(‘Applying the Multistate Capture-recapture Robust Design to investigate metapopulation structure’), we explored the implications of heterogeneity in capture probabilities for mark-recapture studies investigating the characteristics of a metapopulation structure of Indo-Pacific bottlenose dolphins (*Tursiops aduncus*). To do this, we used an extension of the *Pollock’s closed robust design *(Pollock 1982*, *1990): the multistate closed robust design.

**How is Heterogeneity in Capture Probability Minimized?**

In mark-recapture studies, we often have to break the underlying assumption of homogeneity in individual capture probabilities because of practical constraints on sampling. By using the approach of Pollock’s closed robust design, we aimed to minimize this kind of heterogeneity as much as possible.

*First*, Pollock’s closed robust design is based on two different temporal scales:

*Primary periods*are longer periods of time (e.g. seasons, years). The time between primary periods is long enough that the population can be described as*open*. This means that additions to the population through births and immigration, and losses from deaths and emigration can occur (for examples see: Nicholson*et al.*2012 and Smith*et al.*2013).*Secondary Periods*are set within the primary periods and may refer to a survey being carried out on a daily or weekly basis. We describe populations as*closed*between secondary periods, meaning that the intervals between the first and last surveys conducted within a primary period should be so short that no gains or losses occur (again Nicholson*et al.*2012 and Smith*et al.*2013 have some great examples of this).

*Second*, by using this design we can estimate temporary emigration and immigration rates between primary periods. It also lets us estimate abundance and apparent survival parameters without having to assume equal capture probability over the entire study period. This minimizes biases due to heterogeneity in capture probability and, because they’re estimated at multiple occasions, we can be more precise with our estimates of abundance and apparent survival.

With the multistate closed robust design (an extension of the Pollock’s closed robust design), we can reduce the heterogeneity in capture probability by using sites associated with individual capture occasion. This lets us estimate apparent survival and abundance for each site within each primary period. Using such a complex design requires equal effort or number of surveys within each site per primary period.

The multistate closed robust design approach also helps us to understand the dynamic processes between subpopulations within a metapopulation. This is because, with this method, we can find out the probability that an individual will move from one site to another, between sites and primary periods.

**Tips for Using the Multistate Closed Robust Design in the Field**

To maximize the benefits of a multistate closed robust design approach, a number of considerations should be included when designing one sampling regime. Particular attention should be given to:

- Maximizing the coverage of the study area
- Limiting violations of assumptions associated with the mark-recapture approach

In ‘Applying the Multistate Capture-recapture Robust Design to investigate metapopulation structure’, we showed an example of a field sampling design to study bottlenose dolphins in a coastal and estuarine ecosystem. Some of our strategies were to:

- Design multiple zig-zag photo-identification routes within each site to optimize the coverage
- Survey the entire study area in the shortest possible time to minimize the possibility for animals transitioning between the geographic sites.

**A Powerful Tool**

When the geographic sites are associated with anthropogenic impacts or climate change, the use of the multistate closed robust design is a powerful tool for management and conservation of species that can be individually recognisable. The way individuals move from one site to another is particularly relevant for the conservation of highly mobile species (e.g., birds, larger mammals) in environments where anthropogenic pressures vary greatly from one site to another.

**To find out more about the characteristics of a metapopulation of dolphins and the limits of the multistate closed robust design approach, read our ****Methods in Ecology and Evolution**** article ‘****Applying the Multistate Capture-recapture Robust Design to investigate metapopulation structure****’.**