Loading...
 
Print

Marine spatial planning

Efficient conservation management of intra-specific genetic diversity require identification of population subdivisions. Subpopulations may harbour essential local adaptations, but may also be exposed to the risk of genetic loss due to drift and inbreeding. Traditionally, population structure is estimated from neutral genetic markers, but these methods may suffer from high costs and lack of samples. We developed a novel strategy based on dispersal models to identify potential barriers to gene flow in the marine landscape that is complementary to genetic analyses

Analysis of population structure from dispersal information

In a population the probability of interactions among individuals will differ depending on the distance between individuals leading to isolation by distance. Isolation by distance becomes more important for populations with large distributions in relation to individual dispersal distance. In addition, there may be features in the environment forming dispersal barriers, e.g. breaks in habitat distribution or ocean circulation, that constrain interactions within the population. Although in reality there is a continuum, populations may be grouped in three classes depending on how well the different parts are connected through dispersal:

  1. a single large population with high connectivity among it parts
  2. several isolated subpopulations with no or very low connectivity
  3. a population with reduced but significant connectivity among subpopulations, also known as a metapopulation


Conservation and management strategies may differ greatly for these three classes of population structure. Harvested populations, e.g. fish, that are structured into isolated subpopulations or form a metapopulation may require different regulation measures for different subpopulations (stocks). A single large population may be less prone to extinction than smaller isolated subpopulations, which may require stronger protection. When long-term conservation is considered a single large population often harbor greater genetic diversity that may be essential for adaptation to a changing environment. On the other hand subpopulations with restricted connectivity may evolve local adaptations that require local conservation measures. Regardless of the specific goals, the spatial planning of management and conservation actions thus strongly depends on the population structure of target species.

The analysis of population structure can be very complex depending on the population features of interest. Features may include population growth, age structure and genetic diversity. A common approach is to estimate the degree of genetic differences in neutral (non-adaptive) genetic markers from samples in different parts of a species distribution range. This approach has been successfully applied within BaltGene for selected species in the Baltic Sea. Although a very valuable tool, genetic methods have the disadvantage of being expensive and sufficient samples may be difficult to collect. In addition, current genetic methods may not resolve weak population structures that could be important for stock management. Within BaltGene a novel method of population structure analysis have been developed that is complementary to genetic methods.

The new method is based on information about dispersal within the distribution range, or a target area, of a population. The dispersal information may be empirically determined, e.g. from tagging and natural markers, or from dispersal models. In the marine environment many organisms are sessile (fixed) or sedentary (restricted movement) and disperse through microscopic spores, eggs or free-swimming larvae that mainly drift with ocean currents. In BaltGene models for larval dispersal have been developed to estimate dispersal probabilities within the Baltic Sea (HELCOM definition). Dispersal paths (Fig. 1) are calculated based on a 3-dimensional ocean circulation model (Rossby Centre Ocean Model) with a horizontal resolution of 3.7 km, a vertical resolution of 3-12 m, and a temporal resolution of 6 hours. The model is forced with empirical data on wind, sea level, temperature and precipitation and covers 25 years (1981-2005). As a demonstration project the larval dispersal of a typical marine invertebrate was simulated. Larval traits were similar to the blue mussel (Mytilus edulis/trossylus) with spring spawning and dispersal in the surface waters for 3 weeks. A large number of dispersal paths were simulated from all parts of the Baltic Sea with depths above 12 m (Fig. 2). In total 452 million dispersal paths were simulated and these were used to calculate the probability of dispersal between sites and summarized in a connectivity matrix (Fig. 3).A novel theoretical analysis was developed with a computational procedure that identifies subpopulations with high internal connectivity but below some threshold connectivity between subpopulations. This threshold connectivity is determined by the analyst and reflects the relevant biological process that is considered for a particular management or conservation action. Connectivities between subpopulations below 10% per generation has been suggested to result in demographic independence and separate stock characteristics. Connectivities below 0.1-1% may lead to genetic divergence between subpopulations resulting in independent evolutionary trajectories. It is then possible to map all subpopulations suggested by the analysis of the connectivity matrix. 

Figure 4 shows maps of subpopulations for a modeled invertebrate for four different connectivity thresholds. 

Setting the connectivity threshold low, e.g. reflecting conditions that may lead to genetic divergence, leads to fewer subpopulations (Fig. 4A), while higher connectivity thresholds, e.g. representing demographically independent stocks, lead to a finer subdivision (Fig. 4D). Note that species may differ in their connectivities depending on the combination of spawning season, duration as a drifting larvae, and possible larval behaviors. A particular connectivity matrix will produce a unique set of subpopulation maps. However, many species can probably be grouped together being sufficiently similar in their dispersal ability.

 

How can subpopulation maps based on connectivity be used in spatial planning?

In the absence of other information, e.g. genetic data, models of connectivity may be a first step to analyze the population structure. For the required coverage in space and time connectivity modeling is less expensive than genetic methods. Often it may be difficult to acquire genetic samples for many areas and it is essential to allocate samples where information may be highest. Connectivity-based maps of subpopulations can here be used to optimize the collection of samples for genetic analyses. Connectivity-based maps of subpopulations may also be directly used in spatial planning of management of harvested stocks. The maps suggest the presence of dispersal barriers which may prevent or slow down recolonization of depleted stocks. Sustainable production may also be stock-specific and require differentiated management. Connectivity-based maps can further be used in the spatial planning of conservation actions, e.g. when selecting sites for Marine Protected Areas (MPA). A subdivided population may contain many small subpopulations with increased risk for local extinction requiring stronger protection through MPAs and maybe that sufficient dispersal corridors are maintained.The presence of subpopulations also increases the likelihood for essential local adaptations, which may motivate a distribution of MPAs that includes many different subpopulations, as an insurance strategy. A preliminary analysis, using the connectivity-based maps in Fig. 4, shows that the present Natura 2000 network of MPAs is highly unevenly distributed among the suggested subpopulations (Fig. 5). Many subpopulations did not include any or only few MPAs. Future work could produce a complete set of connectivity matrices and subpopulation maps covering the most common larval dispersal abilities in the Baltic Sea.

 

Read more

A detailed description of the identification of subpopulations from connectivity matrices is found in:
Nilsson Jacob M, André C, Döös K, Jonsson PR. 2012. Identification of subpopulations from connectivity matrices. Ecography in press

 

 

CONTRIBUTOR 
Per Jonsson, University of Gothenburg, Sweden


Responsible editor: Per Jonsson, University of Gothenburg, Sweden 
You are welcome to contribute to the development of this section. Please add a comment or visit the Forum. 


 


Who is there...

9 online users