In most mammals and birds, environmental change, population density, mating systems, resource defense, and inbreeding avoidance are some of the factors affecting the genetic diversity, gene flow, and population genetic structure (Berg et al., 2009; Chesser, 1991; Steiner & Gaston, 2005). Specifically, vagility and mating systems are factors that can determine the magnitude of gene flow (Lowe et al., 2004), determine philopatry (Ribeiro et al., 2012), and lead to genetic structuring (Zan et al., 2008). For example, if philopatry is present, then genetic differentiation may occur at small or large scales (Cristofari et al., 2015; Ibarguchi et al., 2011). The genetic structure may be strong between large and geographically distant populations if there is low or no exchange of individuals inducing genetic flow (Milot et al., 2008). Consequently, dispersion patterns result in a genetic structure at different scales (fine or coarse), at the level of social groups, and across territories (Beck et al., 2008; Finnegan et al., 2013; Van Dijk et al., 2015). Therefore, when genetics and behavioral patterns are linked, they can have a strong evolutionary effect even in relatively small areas (Bertrand et al., 2014; Sausner et al., 2016), and philopatry and local adaptation can lead to speciation processes, influence divergence between populations, and determine the degree of genetic subdivision (Abbott & Double, 2003; Dearborn et al., 2003).

Mitochondrial and nuclear DNA markers have been widely used for estimating philopatry and the genetic structure of colonial seabirds (Levin & Parker, 2012; Rayner et al., 2010). Among them, certain types of markers may be useful for different genetic analyses that can reveal different evolutionary scales (Avise, 2004; Rubinoff & Holland, 2005). mtDNA markers are subject to fewer recombination events than nuclear ones (Rokas et al., 2003). Moreover, it has been shown that at equilibrium, mtDNA markers could be sensitive to population structure (Zink & Barrowclough, 2008), and are sensitive for detecting population subdivision, improving spatial resolution (Ibarguchi et al., 2011). Different seabird species inhabiting the same geographical areas may not show genetic differentiation despite records of strong philopatry (Friesen, 2015). For example, Levin & Parker (2012) used microsatellites, NADH dehydrogenase subunit 2 (ND2), cytochrome b (Cyt-b), and cytochrome oxidase I (COI) genes to study the Galapagos Nazca Booby (Sula granti) and the Great Frigatebird (Fregata minor) from the Galapagos archipelago. Despite the short distances between nesting sites, they found evidence of genetic differentiation in the Nazca Booby, a pattern attributed to natal and breeding philopatry, while Great Frigatebirds showed no evidence at all.

Overall, gene diversity and genetic structure of many groups of seabirds are still poorly studied, as is the case of the storm-petrel species. This is a group of almost exclusively pelagic seabirds with global distribution and populations sizes ranging from thousands to millions of individuals (Spear & Ainley, 2007; International Union for Conservation of Nature, IUCN;  http://www.iucnredlist.org). In only a few instances the genetic structure of the storm-petrel species has been evaluated using mitochondrial, nuclear, or both markers. For example, Cagnon et al. (2004) found low genetic differentiation using cytochrome b data within the Mediterranean basin compared to the North Atlantic, with strong differences among populations. Smith et al. (2007) used mitochondrial control region and found significant population structure and differentiation among several colonies (Atlantic and Pacific, Cape Verde and all other populations, sympatric seasonal populations in the Azores, and sympatric seasonal populations in Madeira). Bicknell et al. (2012) found significant global population structure using mitochondrial control region in the Atlantic and Pacific populations of the Leach’s storm-petrel (Oceanodroma leucorhoa leucorhoa). Silva et al. (2016) found evidence of genetic differentiation also using mitochondrial control region. The most recent study by Antaky et al. (2020) (which used high-throughput sequencing) reported high levels of genetic diversity, with little differentiation between breeding colonies of the Hawaiian Band-rumped Storm-Petrel (Oceanodroma castro). Most of these examples found either strong and moderate evidence of genetic structure at large scales and regional scales, respectively (Antaky et al., 2020; Bicknell et al., 2012; Cagnon et al., 2004; Silva et al., 2016; Smith et al., 2007). Although studies such as the one by Silva et al. (2016) considered populations in the eastern Pacific near tropical American continent, none has been performed close to the coasts in this area.

The Black Storm-Petrel (H. melania) and the Least Storm-Petrel (H. microsoma) are pelagic seabirds showing similar ecological characteristics and distribution ranges. For example, they nest only on very isolated rocky islands with no natural predators (Everett, 1991). Historically, the largest breeding populations of these seabirds are found in Isla Partida Norte, Gulf of California (Figure 1) (Everett & Anderson, 1991; Spear & Ainley, 2007) and Islas San Benito, in the Pacific Ocean off the west coast of the Baja California peninsula (Velarde-González, 2008a, 2008b). However, other breeding populations included small islands in both the Gulf of California and the Pacific Ocean. For H. melania, this includes Los Coronados and San Benito in the Pacific and San Luis, Partida Norte and Roca Consag in the Gulf of California. For H. microsoma, they are the same islands (American Ornithologists’ Union, 1998; Bent, 1922); however, a small breeding site, islote La Lobera (close to Espíritu Santo Island) has also been reported (Carmona et al., 2020). After the breeding season, both species migrate long distances to the north off the coast of northern California and to the south to off the coast of Ecuador (American Ornithologists’ Union, 1998; Crossin, 1974). However, in recent times, threats such as introduced species of predators and human disturbance have caused the decline of many insular vertebrate populations (Bedolla-Guzmán et al., 2017). Particularly, islands with important nesting colonies of this species have been invaded by introduced predators, making Isla Partida Norte the most important predator-free colony (Everett & Anderson, 1991; Velarde-González, 2000). Although, the IUCN red list considers the conservation status of both H. melania and H. microsoma as “least concern” (BirdLife International, 2018a, 2018b), under Mexican environmental law (SEMARNAT-NOM-059–2010) they are considered “Amenazadas” (Threatened, in English). Here, we evaluated the genetic diversity and population structure of the Black Storm-Petrel (H. melania) and the Least Storm-Petrel (H. microsoma), and discuss explanations leading to divergent population genetic structure and potential consequences for their conservation. Different signatures of genetic structure and genetic differentiation may occur because of differences in the range and frequency of migration among colonies (Spear & Ainley, 2007). Despite of some overlap, distribution ranges are different southward because H. melania tends to visit the Galapagos Islands and even hundreds of kilometers in open sea, while H. microsoma distribution is mostly close to the coasts.

Figure 1.

Study Area in Isla Partida Norte, Gulf of California. The map highlights the locations where H. melania and H. microsoma were sampled (A–J letters). Asterisks represent sites were only H. melania individuals were found. In all other sites both species were sampled. The AMOVA groups of geographical zones used in the genetic analyses are identified by colored lines: blue (southwest hill, SW), yellow (land bridge, LB), and red (northeast hill, NE), as depicted in the text (see Methods section). Grouping of sampling localities maximally differentiated found using SAMOVA: Tessellation of H. melania groupings ({A} {I, J} {B, C} {D} {F, G, H} {E}), dashed lines; tessellation of H. microsoma groupings ({B} {C} {F, G, H} {I}), dotted lines.

Materials and Methods

Results

The mtDNA COI gene sequence length was 719 bp for both species, after alignment and editing. We recovered 113 and 70 sequences of H. melania and H. microsoma, respectively, each of them with 46 haplotypes (Table 1). Sequences were deposited in GenBank (accession numbers MH510865- MH510977 and MH510978-MH511047 for H. melania and H. microsoma, respectively). High and similar variability was found across sequences of both species; the number of segregating sites in H. melania was 26, while in H. microsoma was 28. The analyses of haplotype (Hd) and nucleotide (π) diversities showed to be extensive for both species; for H. melania Hd was 0.70 (SD = 0.05) and π was 0.00291 (SD = 0.0004), while for H. microsoma Hd was 0.95 (SD = 0.02) and π was 0.00559 (SD = 0.0006). However, H. microsoma showed higher values than H. melania. The Minimum Spanning Network was not fully resolved for H. melania, as many alternative connections between haplotypes were observed (Figure 2). Haplotype H4 was the most common and it was shared by 61 individuals, while only very few others were observed in only two individuals (H2, H5, H11, H15, and H27). All other haplotypes were unique and observed in only one individual. Most of the haplotypes that were connected to the most common haplotype H4 do so with one to three mutations, a condition observed in almost all other haplotypes. No clear relationship between the network and the sites defined for the AMOVA analyses was found (Figure 2). For example, haplotype H4 was found in all ten sampled localities. This was also the case between the network and the groupings inferred by SAMOVA results (data not shown). Conversely, the Minimum Spanning Network was better resolved for H. microsoma, with fewer alternative connections between haplotypes (Figure 3). The most common haplotypes were H11 and H3 (shared by 14 and 6 individuals, respectively), and four other haplotypes (H9, H10, H16 and H39) were observed in two individuals. All other haplotypes were unique and observed in only one individual. The network showed that very few haplotypes were connected to the most common, and many showed up to five mutations among them. As mentioned for H. melania, there was no relationship between the network and the sample localities defined for the AMOVA analyses or the groupings inferred by SAMOVA results. The networks show high frequency of single haplotypes in both species, especially in H. microsoma. The results of the AMOVA analysis, including both species Φ-statistics and P-values are summarized in Table 1. We found genetic differentiation for H. melania and H microsoma, although at different sources of variation. In the first case, no genetic differentiation was observed among groups (Φ_CT = −0.00657, P > 0.05) or within localities (Φ_ST = 0.0179, P > 0.05) for H. melania. However, genetic differentiation was detected among localities within groups (Φ_SC = 0.02431, P < 0.01). On the other hand, genetic differentiation was detected among groups (Φ_CT = 0.03456, P < 0.05) for H. microsoma, but not when other sources of variation were considered (Φ_SC = −0.02343, P > 0.05; Φ_ST = 0.01195, P > 0.05). The results of the SAMOVA analysis using both the geographically homogeneous groups (maximally differentiated from each other) (Figure 1; Table 2) and the maximally differentiated groups (non-constrained (by geography) groups) were almost identical. Therefore, only the results of the geographically homogeneous groups approach were included. The statistically significant and maximally differentiated number of groups identified (K) for H. melania was six (F_CT = 0.045, P < 0.001; tessellation: {A} {I, J} {B, C} {D} {F, G, H} {E}). A similar pattern was found for H. microsoma, as the corresponding maximally differentiated number of groups was four (F_CT = 0.073, P = 0.05; tessellation: {B} {C} {F, G, H} {I}) (Figure 1; Table 2). Finally, when geographical distances vs pairwise p-distances were plotted, an IBD pattern was observed only for H. microsoma (rM = 0.464; P = 0.033; y = 1E-06x + 0.0043). This was not the case for H. melania (rM = -0.139 P = 0.033; y = 1E-06x + 0.0043) (Table 3). Similarly, the plot of geographic distances vs Φ_ST/1– Φ_ST distances showed an IBD pattern for H. microsoma (rM = 0.497 P = 0.032 y = 9E-05x−0.053), while for H. melania this pattern resulted in a negative correlation (rM = −0.007 P = 0.05 y = −2E-06x - 0.058) (Table 3). Therefore, only the plots of linear regressions regarding pairwise p-distances and Φ_ST/1– Φ_ST distances showed a consistent IBD pattern for H. microsoma.

Table 1.

Genetic Diversity Estimates and Analysis of Molecular Variance (AMOVA) for H. melania and H. microsoma From Mitochondrial Sequences of Cytochrome Oxidase Subunit I (COI).

Figure 2.

Minimum Spanning Network Showing the Relationship Among Cytochrome Oxidase Subunit I (COI) Haplotypes From H. melania. Circles’ sizes are proportional to haplotype frequencies. Numbers in parentheses (when provided) are individuals sharing the same haplotype; otherwise, only one individual was found with that haplotype. Hatch marks represent the number of inferred mutations between haplotypes. Colors within the haplotypes represent the sample localities where the samples were taken.

Figure 3.

Minimum Spanning Network Showing the Relationship Among Cytochrome Oxidase Subunit I (COI) Haplotypes From H. microsoma. Circles’ sizes are proportional to haplotype frequencies. Numbers in parentheses (when provided) are individuals sharing the same haplotype; otherwise, only one individual was found with that haplotype. Hatch marks represent the number of inferred mutations between haplotypes. Colors within the haplotypes represent the sample localities where the samples were taken.

Table 2.

Fixation Indices Corresponding to the Groups Geographically Homogeneous and Maximally Differentiated Inferred by Spatial Analysis of Molecular Variance (SAMOVA) for H. melania and H. microsoma, Tested for Cytochrome Oxidase Subunit I (COI) Mitochondrial Sequences.

Table 3.

Isolation by Distance (IBD) Estimates and Mantel’s Test for H. melania and H. microsoma From Mitochondrial Sequences of Cytochrome Oxidase Subunit I (COI).

Discussion

In some species of the Procellariiformes group, the patterns of genetic diversity have been explained as the consequence of either ecological (current) and/or evolutionary (historical) events (Antaky et al., 2020; Silva et al., 2016). In the first case, low diversity was associated with intense predation pressure observed in the Atlantic populations of Oceanodroma leucorhoa (Leach's Storm-Petrel), while high diversity was observed in the Pacific populations, which in turn were subject to less intense predation pressure (Bicknell et al., 2012). In the second case, low but fixed nucleotide differences between groups of Hydrobates pelagicus (British Storm-Petrel) within Atlantic and Mediterranean basins were consistent with long-term geographic isolation, possible because of the existence of some barrier to gene flow or recent changes in demography and ecology (Cagnon et al., 2004). Conversely, high divergence has been linked with low potential for dispersal and intrinsic reproductive barriers, such as high philopatry and local adaptation (Silva et al., 2016). High diversity patterns are consistent with high diversity values found in this study for H. melania and H. microsoma, as there are no reports of intense predation pressure (at least until recent times) and their potential of dispersal is high in the eastern Pacific. However, the two species clearly showed different haplotype and nucleotide diversity values (both estimates were smaller in H. melania), with no overlap. The higher diversity values for H. microsoma may account for its much larger population size, which is possibly one order of magnitude greater than that of H. melania (Brooke, 2004). A possible explanation for this pattern is that H. microsoma may be experiencing population expansion, while for H. melania this may not be the case. If so, appropriate hypotheses testing of potential bottlenecks (i.e. sudden population reduction followed by rapid expansion), along with additional genetic markers, should be considered in any further studies. Moreover, the molecular diversity observed in our focal species was higher or similar to that found in other Procellariiformes. For example, the molecular diversity found in H. microsoma was higher than that reported for the O. leucorhoa control region (mean Hd = 0.729, ±0.14; mean π = 0.0054 ± 0.00073) (Bicknell et al., 2012) and H. pelagicus cytochrome b (Hd = 0.7, ±0.037; π = 0.00396 ± 0.00028) (Cagnon et al., 2004), but similar to that for O. castro control region (Hd = 0.965, ±0.007; π = 0.0408 ± 0.0017) (Silva et al., 2016).

Although most studies on the population genetic structure of seabirds have focused on large geographical scales, there are some instances where it has been uncovered at microscales. For example, using mtDNA and microsatellites Ibarguchi et al. (2011) evaluated the genetic structure of Uria lomvia (Thick-billed Murre) in a colony from Coats Island, Nunavut, Canada. They found no genetic structuring either among ledges or among the subareas level, regardless of grouping. However, in male individuals located in different ledges of the same cliff (less than 205 m long), evidence of low but significant genetic structuring was found, as well as on the east vs the west side of the colony. Similar findings have been reported integrating behavior and genetic information (Ibarguchi et al., 2011; Lecomte et al., 2009; Ribeiro et al., 2012). If the population genetic structure at fine-scale level is assessed with molecular markers, two possible scenarios are expected. On one hand, genetic variation can be distributed heterogeneously between groups; that is, genetic structure occurs because of philopatry and reduced gene flow (Ovenden et al., 1991; Rayner et al., 2011; Welch et al., 2012). Alternatively, genetic variation can be homogeneously distributed (i.e. panmixia) with reduced molecular diversity; that is, no genetic structure and absence of philopatry due to high gene flow (Austin et al., 1994; Bicknell et al., 2012; Milot et al., 2008). Despite the impossibility of considering differences between sexes due to the selected marker (mtDNA markers are only female inherited), our results agree with the first described scenario for H. microsoma, although evidence for H. melania was inconclusive. First, for H. microsoma the AMOVA analysis identified that variation among the a priori grouping of sites (northeast hill, land bridge, and southwest hill, Figure 1) was statistically significant. However, that was not the case for H. melania, as variation among groups of sites was lower although non-significant (Table 1). In this case, only variation among sample localities within groups was statistically significant. Second, we found significant differentiation in both species for the number of groups maximally differentiated (K) in the a posteriori grouping SAMOVA analyses. For H. melania, K was six (groups of sampling localities: {A} {I, J} {B, C} {D} {F, G, H} {E}), while for H. microsoma K was four (groups of sampling localities: {B} {C} {F, G, H} {I}). These values suggest that genetic population structuring is higher in H. microsoma, and that its maximally differentiated groups are akin to the sites defined a priori for the AMOVA analyses (Table 2; Figure 1). Further, an isolation by distance (IBD) pattern was clearly identified for H. microsoma, using both the paired genetic distances (p–distances) and genetic differentiation (Φ_ST/(1 - Φ_ST)); however, this was not the case for H. melania, as none of these estimators were significant (Table 3). Given that both species have common evolutionary origins (i.e. they are closely related due to phylogenetic relationships; Wallace et al., 2017) and present analogous ecological constraints (e.g., nesting sites are quite similar), it was expected that their population genetic structure was similar. A possible explanation for the contrasting genetic structure in our focal species is that they behave differently enough (e.g. in range and frequency of migration among colonies) as to develop subtle but distinct responses to their environment (e.g. differentiated foraging due to geographic partition of food availability), leading to divergent population structure signatures. Alternatively, although some evidence for genetic structuring was found, it is important to note that some bias may be caused by sex differences. This may be the case if such genetic structuring is driven mainly by philopatry. For example, previous observations in other seabird studies claim that males are more prone to show philopatry, while a greater dispersion capacity and less fidelity site has been observed in females (Greenwood, 1980; Ibarguchi et al., 2011; Ribeiro et al., 2012). If so, it would be necessary to test for male-biased dispersal (although different genetic markers are required to detect this). However, it is important to note that seemingly inconsistent results do not necessary mean absence of philopatry, because processes of divergence and differentiation can occur despite gene flow (Taylor et al., 2013), as well as panmixia may be inferred despite barriers to gene flow (Oomen et al., 2011). Finally, we acknowledge that mtDNA markers have a number of limitations and shortcomings, such as the nature of its matrilineal inheritance, a lower discrimination power at population level, the occurrence of discrepancies between nuclear and mtDNA inheritance patterns, and artifacts derived from pseudo-genes, among others (Rokas et al., 2003; Rubinoff & Holland, 2005; Zink & Barrowclough, 2008). However, it allowed us to rapidly describe genetic diversity and to test whether genetic structure could be observed in two species from one single island in a small spatial scale in the context of the Gulf of California. Further work should consider the exploration and adequate sample collection on other islands such as San Benito Island, off the coast of Baja California (which is where other major breeding sites for both species are located) will allow to test different hypothesis on genetic structure in a broader geographical and ecological context. Other nuclear markers (microsatellites, SNPs, NGS, etc.), as well as sex identification to test for sex-bias, should also need to be considered.

Implications for Conservation

We restricted our study to explore the genetic diversity and fine-scale genetic structure, disregarding other components such as population sink-source dynamics and multiple ecological and evolutionary factors. These components should be included in future conservation studies to describe more extensively the influence of philopatry at the population level within each species, as well as the consideration of the role of alloparenting, kin group selection, and the evolution of helping behavior, among others. Moreover, for both species the maintenance of their genetic diversity is an important factor for conservation management in future generations. Given their restricted breeding ranges, it may be necessary to elucidate which populations require major attention and protection. Our study provides baseline information for comparison with other breeding islands of the same two species, either experiencing similar demographic or ecological scenarios within the Midriff region of the Gulf of California. Although the present study focused on two seabird species of limited and restricted breeding area, it is important because it may be useful as a reference for other, similar studies, both in the same area with other seabird species or in any other areas worldwide with similar or equivalent ecological contexts. In the light of predicted adverse climate change for this region, changes in the demography and other ecological and evolutionary dynamics are expected. Despite the high genetic variation found in the two storm-petrel species, increasing human-induced factors (such as climate change, overfishing, and water and air pollution) threatens their conservation. We raise attention to the urgency to establish conservation programs for these and other seabird species in the region.