The antique genetic plight of the Mediterranean monk seal (Monachus monachus)

Abstract

Disentangling the impact of Late Quaternary climate change from human activities can have crucial implications on the conservation of endangered species. We investigated the population genetics and demography of the Mediterranean monk seal (Monachus monachus), one of the world's most endangered marine mammals, through an unprecedented dataset encompassing historical (extinct) and extant populations from the eastern North Atlantic to the entire Mediterranean Basin. We show that Cabo Blanco (Western Sahara/Mauritania), Madeira, Western Mediterranean (historical range) and Eastern Mediterranean regions segregate into four populations. This structure is probably the consequence of recent drift, combined with long-term isolation by distance (R² = 0.7), resulting from prevailing short-distance (less than 500 km) and infrequent long-distance dispersal (less than 1500 km). All populations (Madeira especially), show high levels of inbreeding and low levels of genetic diversity, seemingly declining since historical time, but surprisingly not being impacted by the 1997 massive die-off in Cabo Blanco. Approximate Bayesian Computation analyses support scenarios combining local extinctions and a major effective population size decline in all populations during Antiquity. Our results suggest that the early densification of human populations around the Mediterranean Basin coupled with the development of seafaring techniques were the main drivers of the decline of Mediterranean monk seals.

Keywords: Mediterranean monk seal; demographic history; genetic diversity; isolation by distance; marine mammals; population decline.

Figure 1.

Monachus monachus nuclear genetic structure. Individual posterior cluster membership coefficients for K = 4 represented using vertical barplot (a) and geolocated pie-charts (b). Representation of the samples from the six main sampling areas on the first two axes of PCA (c) showing a strong west–east cline of differentiation on the first axis. In (a), the samples are ordered by regions, sorted from west to east, separated by vertical black lines, and stars denote historical samples. In (b), black circles denote locations that include historical samples and the cyan area represents the putative ancient MMS distribution. In (c), full and empty symbols represent historical and modern samples, respectively. W-Sahara, Cabo Blanco (Western Sahara/Mauritania); W-Med, western Mediterranean Sea; C-Med, central Mediterranean Sea; E-Med, eastern Mediterranean Sea. (Online version in colour.)

Figure 2.

Monachus monachus nuclear genetic diversity. Boxplot of the microsatellites (a) allelic richness (A_R), (b) observed (H_O) and (c) expected (H_E) heterozygosity, (d) mean private alleles (P_A) of 500 resampling of the smallest sample size across each sampling area, and barplot of the number of private alleles (P_A), and (e) boxplot of the average individual inbreeding coefficient (F) per sampling area, expressed for modern (from 1989) and historical (before 1976) samples. In (a–d), historical samples were removed from the analyses for all extant populations apart from the W-Med where no modern population is clearly identified and for which we almost exclusively have historic samples. Letters and box colours in (a–d) illustrate the Tukey posthoc group assignment. W-Sahara, Cabo Blanco; W-Med, Western Mediterranean Sea; CE-Med, merged Central and Eastern Mediterranean Sea; C-Med, Central Mediterranean Sea; E-Med, Eastern Mediterranean Sea. (Online version in colour.)

Figure 3.

Isolation by distance in Monachus monachus. (a) Isolation by distance (IBD) relationship between individual-based genetic and maritime distances. (a,b) Mantel correlogram of spatial correlation, for 59 classes of 100 km, in all, modern, and historical samples (b), and all, males and females samples (c). Significant values, over 1000 permutations, are represented by filled symbols. Complementary analyses using alternative genetic distances are presented in electronic supplementary material, figures S27 and S28. (Online version in colour.)

Figure 4.

The antique decline of MMS. ABC-GLM posterior estimates of effective population size (N_e; a) and of the time (T₁ and T₂; b) of the major demographic event(s)—Population decline + local extinction + change in connectivity—estimated from the most likely scenarios (M180 and M181), within the ABC framework. (a): mode (symbol) and 50% quantiles (dotted arrows) of the effective population size (N_e) under the models 181 and 180, represented after (N₀: current populations) and before (N₁ and N₂: ancient populations) the demographic events occurring at T₁ and at T₂. (b): scaled-density distributions of the time of the demographic event(s) T₁ and at T₂. WS, Cabo Blanco; MAD, Madeira; WM, Western Mediterranean; CM, Central Mediterranean; EM, Eastern Mediterranean; Other N₁ and Other N₂, cumulated ancient effective population sizes of other populations before T₁ and T₂, respectively. (Online version in colour.)