Assessing fine-scale pondscape connectivity with amphibian eyes: An integrative approach using genomic and capture-mark-recapture data

In the face of habitat loss, preserving functional connectivity is essential to maintain genetic diversity and the demographic dynamics required for the viability of biotic communities. This requires knowledge of the dispersal behaviour of target species, which can be modelled as kernels, or probability density functions of dispersal distances at increasing geographic distances. We present an integrative approach to investigate the relationships between genetic connectivity and demographic parameters in organisms with low vagility focusing on five syntopic pond-breeding amphibians. We genotyped 1056 individuals of two anuran and three urodele species (1732-3913 SNPs per species) from populations located in a landscape comprising 64 ponds to characterize fine-scale genetic structure in a comparative framework, and combined these genetic data with information obtained in a previous 2-year capture-mark-recapture (CMR) study. Specifically, we contrasted graphs reconstructed from genomic data with connectivity graphs based on dispersal kernels and demographic information obtained from CMR data from previous studies, and assessed the effects of population size, population density, geographical distances, inverse movement probabilities and the presence of habitat patches potentially functioning as stepping stones on genetic differentiation. Our results show a significant effect of local population sizes on patterns of genetic differentiation at small spatial scales. In addition, movement records and cluster-derived kernels provide robust inferences on most likely dispersal paths that are consistent with genomic inferences on genetic connectivity. The integration of genetic and CMR data holds great potential for understanding genetic connectivity at spatial scales relevant to individual organisms, with applications for the implementation of management actions at the landscape level.

Integrating dispersal, breeding and abundance data with graph theory for the characterization and management of functional connectivity in amphibian pondscapes

CONTEXT

Robust assessment of functional connectivity in amphibian population networks is essential to address their global decline. The potential of graph theory to characterize connectivity among amphibian populations has already been confirmed, but the movement data on which modelled graphs rely are often scarce and inaccurate. While probabilistic methods that account for intraspecific variability in dispersal better reflect the biological reality of functional connectivity, they must be informed by systematically recorded individual movement data, which are difficult to obtain for secretive taxa like amphibians.

OBJECTIVES

Our aim is to assess the applied potential of probabilistic graph theory to characterize overall connectivity across amphibian pondscapes using fine-scale capture-recapture data, and to inform conservation management based on the role of ponds on functional connectivity.

METHODS

We monitored an amphibian community in a pondscape located in a Spanish "dehesa" for 2 years. Photoidentification was used to build capture histories for individuals of six species, from which dispersal kernels and population sizes were estimated to model probabilistic graphs.

RESULTS

We obtained kernels of variable robustness for six species. Node importance for connectivity varied between species, but with common patterns such as shared road crossing areas and the presence of coincident interconnected pond clusters.

CONCLUSIONS

The combination of photoidentification, capture-recapture data and graph theory allowed us to characterize functional connectivity across the pondscape of study accounting for dispersal variability and identify areas where conservation actions could be most efficient. Our results highlight the need to account for interspecific differences in the study and management of amphibian pondscapes.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s10980-022-01520-x.