Genetic data and niche differences suggest that disjunct populations of Diglossa brunneiventris are not sister lineages
Parra, Juan Luis et al. (2022), Genetic data and niche differences suggest that disjunct populations of Diglossa brunneiventris are not sister lineages, Dryad, Dataset, https://doi.org/10.5061/dryad.jm63xsj9f
Disjunct distributions within a species are of great interest in systematics and biogeography. This separation can function as a barrier to gene flow when the distance among populations exceeds the dispersal capacity of individuals, and depending on the duration of the barrier, it may eventually lead to speciation. Here we describe patterns of geographic differentiation of two disjunct populations of Diglossa brunneiventris separated by approximately 1000 km along the Andes. Diglossa brunneiventris vuilleumieri is isolated in northern Colombia, while Diglossa brunneiventris brunneiventris has a seemingly continuous distribution across Peru, Bolivia, and Chile. We sequenced mitochondrial and nuclear DNA of the two Diglossa brunneiventris subspecies to evaluate whether they form a monophyletic clade, while including the other three species within the carbonaria complex (D. gloriosa, D. humeralis and D. carbonaria). We also constructed ecological niche models for each Diglossa brunneiventris subspecies to compare their climatic niches. We found that when using all available molecular data, the two D. brunneiventris subspecies are not sister lineages. In fact, each subspecies is more closely related to other species in the carbonaria complex. Our niche modeling analyses showed that the subspecies are occupying almost entirely different climatic niches. An additional, and not expected result was that the carbonaria complex might encompass more cryptic species than previously considered. We suggest reevaluating the taxonomic status of these brunneiventris populations, especially the northern subspecies, given its highly restricted range and potential threatened status.
We compiled a database of occurrence records from both populations of Diglossa brunneiventris from the Global Biodiversity Information Facility (GBIF) as well as from our own records (Fig. 1). Each record was evaluated for inconsistencies in the locality description, elevation, taxonomy and georeference. Suspicious records (i.e. locality not corresponding to coordinates, inconsistent elevation) were discarded. After all occurrence records were checked for quality, we used them to estimate and compare the climatic niches of each subspecies following the methodology proposed by Broennimann et al. (2012) in which occurrence density surfaces are generated via kernel density estimation based on the environments available and the frequency of occurrences in each accessible environment. By environment, we specifically refer to the 19 bioclimatic variables available from the WorldClim database (Hijmans et al. 2005, v1.4; Supplemental Material Table S1). These variables represent climate variation (mean, seasonality and extremes) for temperature and precipitation at ~ 1 km2 spatial resolution and for an interval of 50 years (1950-2000). Many of these climatic variables are highly correlated and the niche comparison is based on a principal component analysis of them. The methodology proposed by Broennimann et al. (2012) estimates climatic preferences by comparing the available or accessible climate (determined by all climates within the accessible area) in relation to the occupied or used climate (determined by all climates occupied by the species). The accessible area, understood as all areas in geographic space that the species is able to access, is referred to as M in ecological niche modeling theory (Soberón and Nakamura 2009, Barve et al. 2011). Since the method is based on occurrence frequencies in environmental space and to prevent correlation among detections, we chose to use only a single record per 1 km2 pixel. We determined the accessible environments for each subspecies as those within the ecoregions (Dinerstein et al. 2017) where at least one occurrence of the population was identified (Fig. 1). Finally, to formally evaluate how similar the niches of the two populations of D. brunneiventris were, we quantified their overlap using a similarity metric developed by Schoener (1968), but applied in environmental space as proposed by Broennimann et al. (2012). This metric essentially quantifies the amount of climatic space used commonly by both populations, acknowledging the climates that are accessible for each one. We used two null model approaches developed originally by Warren et al. (2008) to evaluate if similarity is higher than expected under a proposed null model. These two are known as the niche equivalency and the niche similarity tests. The null model of the niche equivalency test randomizes the identity of the population that each locality belongs to and calculates the overlap between populations (northern and southern subspecies) for a given number of iterations, while the null model of the niche similarity test randomizes the position of the occurrences within the accessible area for one population while holding the other fixed and calculates the overlap between populations (northern and southern subspecies) for a given number of iterations. In either case, the observed overlap is compared to the distribution of overlap values obtained by the given null model. The expectation under the null hypothesis for the niche equivalency (identity) test is that both populations share identical climatic preferences. If rejected, then the subspecies exhibit differences in their climatic preferences. For the niche similarity test, the null hypothesis is that subspecies differences in climatic preferences are due to differences in climate availability. Thus, if the null hypothesis is rejected, climatic preferences can be more or less different than expected under the climates available in each area.
Universidad de Antioquia, Award: Cooperation Agreement No. 13-13-014-347CE
Alexander von Humboldt-Stiftung, Award: Cooperation Agreement No. 13-13-014-347CE