late pleistocene to holocene distributional stasis in ... · late pleistocene to holocene...

Late Pleistocene to Holocene distributional stasis inscorpions along the Baja California peninsula

MATTHEW R. GRAHAM1,2*, ROBERT W. BRYSON, JR3 and BRETT R. RIDDLE1

1School of Life Sciences, University of Nevada Las Vegas, Las Vegas, NV 89154, USA2Department of Biology, Eastern Connecticut State University, Willimantic, CT 06226, USA3Department of Biology and Burke Museum of Natural History and Culture, University ofWashington, Seattle, WA 98195, USA

Received 5 July 2013; revised 5 October 2013; accepted for publication 5 October 2013

The biota of the Baja California peninsula (BCP) assembled in response to a complex history of Neogene tectonicsand Quaternary climates. We constructed species distribution models (SDMs) for 13 scorpion species from the BCPto compare current suitable habitat with that at the latest glacial maximum about 21 000 years ago. Using theseSDMs, we modelled climatic suitability in relation to latitude along the BCP. Our SDMs suggested that most BCPscorpion distributions have remained remarkably conserved across the latest glacial to interglacial climatictransformation. Three areas of climatic suitability coincide remarkably well with genetic discontinuities in otherco-distributed taxa along the BCP, indicating that long-term persistence of zones of abrupt climatic transition offera viable alternative, or synergistic enhancement, to hypotheses of trans-peninsular seaways as drivers ofpeninsular divergences. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014,111, 450–461.

ADDITIONAL KEYWORDS: ecological niche model – Mexico – Quaternary – Scorpiones – species distribu-tion model.

INTRODUCTION

Quaternary climate change has profoundly influencedthe associations of species, communities, ecosystemsand biotas. Cooler climates during the last glacialmaximum (LGM, c. 21 000 years ago) followed by theperiod of warming (albeit not a simple linear warmingtrend) from the LGM until the present represents anepisode of climatic transition that can be decipheredfor understanding how species and biotas mightrespond to current and future climatic warming.Molecular genetics (e.g. Klicka & Zink, 1997; Avise,2000) and palaeontology (e.g. Betancourt, VanDevender & Martin, 1990; Macfadden, 2006) haveprovided keen insights into the evolutionaryresponses of species to climate change during thisperiod of time. Species distribution modelling (SDM),also referred to as ecological niche modelling (see

Saupe et al., 2012), is emerging as an additional toolfor assessing potential species responses to changingclimate. By using climate data to derive SDMs, andthen projecting these models onto climatic simula-tions of the LGM (‘hindcasting’), investigators seekto reconstruct the impact of Quaternary climatechange on the distributional responses of species. Oneoutcome might be the emergence of common responsepatterns within and among co-distributed species,providing a tool for predicting the future assemblyand disassembly of regional biotas.

The Baja California peninsula (BCP) of Mexico isthe second longest terrestrial peninsula in the world,and possesses a diverse array of desert and tropicaldry forest environments. Originally, the BCP (Fig. 1)was thought to have a dispersal-dominated biotichistory (Savage, 1960), with desert species leavingand re-entering the peninsula in concert with glacial–interglacial climatic cycles. Later advances in ourunderstanding of plate tectonics, however, facilitatedcomplex vicariant hypotheses that better explained*Corresponding author. E-mail: [email protected]

bs_bs_banner

Biological Journal of the Linnean Society, 2014, 111, 450–461. With 4 figures

© 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 450–461450

mailto:[email protected]

the distribution of biodiversity along the BCP (e.g.Hess, 1962; Murphy, 1983; Grismer, 1994). Startingwith seafloor spreading, rifting and volcanic origina-tion of islands that coalesced into the modern penin-sula during the mid-Miocene (e.g. Lonsdale, 1989;Spencer & Normark, 1989; Stock & Hodges, 1989),geological activity is now widely held to have heavilyinfluenced the diversification, distribution and com-position of the BCP biota (Carreño & Helenes,2002).

Recent studies have revealed a number of spatiallycongruent genealogical breaks across the BCP (Upton& Murphy, 1997; Riddle et al., 2000; Murphy &Aguirre-Léon, 2002; Crews & Hedin, 2006; Lindell,Méndez-de la Cruz & Murphy, 2008; Bryson et al.,2012). The spatial congruence of molecular breaksacross these disparate taxonomic groups promptedthe hypothesis that vicariant events driven by theformation of trans-peninsular seaways could haveisolated populations that subsequently reunited at

Figure 1. Map of the Baja California peninsula with major geographical regions delineated and hypothesized trans-peninsular breaks shown: A, mid-peninsular break; B, Loreto break; C, La Paz break.

DISTRIBUTIONAL STASIS IN BAJA CALIFORNIA SCORPIONS 451


contact zones following a period of evolutionary diver-gence (Lindell, Méndez-de la Cruz & Murphy, 2005,2008). Others have suggested that genetic disconti-nuities can be explained by abrupt habitat transitionscreated by latitudinal gradients in climatic patterns,as an alternative to trans-peninsular seawayvicariance (Grismer, 2002; further discussed byRiddle & Hafner, 2006). A third possibility is thatabrupt habitat transitions operated synergisticallywith trans-peninsular seaways to reinforce long-termisolation of populations (Riddle & Hafner, 2006).

To examine the potential impact that historicalclimate change has had on biodiversity along theBCP, we modelled the current and past distributionsof 13 scorpion species representing four families(Caraboctonidae, Chactidae, Scorpionidae, Vaejo-vidae) distributed along the peninsula. Scorpions areparticularly diverse and well documented on the BCPand occupy an array of habitat types (Williams, 1980;Due & Polis, 1986). Quaternary climate fluctuations,such as the period between now and the LGM, havedriven profound distributional shifts in many desertspecies (e.g. Jezkova et al., 2009; Oláh-Hemmings

et al., 2010; Wilson & Pitts, 2012; Graham et al.,2013a, 2013b). Based on these studies, we predictedsimilar dynamic shifts in the reconstructed distribu-tions of scorpions along the BCP as desert habitatscontracted and expanded. The absence of large shiftsin suitable habitat between the LGM and the presentday would support an alternative hypothesis of sta-bility in climatically mediated distributions.

MATERIAL AND METHODSSCORPION OCCURRENCE DATA

Occurrence records for 13 species of scorpions fromthe BCP (Table 1) were obtained from museumrecords (California Academy of Sciences and SanDiego Natural History Museum), private collectionsand from the literature (Stahnke, 1968, 1969;Williams, 1970a, 1970b, 1970c, 1971, 1974, 1980;Williams & Lee, 1975; Sissom, 1991). Although over60 species of scorpion are found on the BCP(Williams, 1980), these 13 species were selectedbecause of their relatively wide distributions across

Table 1. Families, distributions, sample sizes and measures of niche shifting for 13 scorpion species from the BajaCalifornia peninsula

Species Family DistributionTotal no.of samples

Testsamples

Centroiddistance (km)

Change inarea (km2)

Anuroctonus pocockiSoleglad & Fet, 2004

Chactidae Northern 138 34 41.3 –104 606.4

Hoffmannius puritanus(Gertsch, 1958)

Vaejovidae Northern 38 9 188.9 –62 421.44

Kochius hirsuticauda(Banks, 1910)

Vaejovidae Northern 51 12 199.0 27 423.26

Hadrurus concolorousStahnke, 1969

Caraboctonidae Central 60 15 120.6 –12 821.9

Hadrurus pinteriStahnke, 1969

Caraboctonidae Central 21 5 5.0 –41 827.46

Kochius bruneus(Williams, 1970)

Vaejovidae Central 30 7 260.6 85 479.16

Bioculus comondaeStahnke, 1968

Scorpionidae Southern 19 4 33.5 –2672.04

Hoffmannius diazi(Williams, 1970)

Vaejovidae Southern 29 7 64.6 12.92

Hoffmannius vittatus(Williams, 1970)

Vaejovidae Southern 36 9 44.2 –31 697.43

Nullibrotheas allenii(Wood, 1863)

Chactidae Southern 47 11 101.6 6425.01

Bioculus caboensis(Stahnke, 1968)

Scorpionidae Cape 18 4 16.4 –33 073.46

Hadrurus hirsutus(Wood, 1863)

Caraboctonidae Cape 27 6 21.71 –15 927.77

Kochius punctipalpi(Wood, 1863)

Vaejovidae Cape 30 7 19.1 –25 568.22

452 M. R. GRAHAM ET AL.


one or more of the northern, central, southern andCape regions of the BCP. Although common along theBCP, we did not include Centruroides exilicauda inthis study because genetic data suggest that it mightrepresent a complex of cryptic species (Gantenbein,Fet & Barker, 2001). We used Google Earth (http://earth.google.com) to estimate latitude and longitude(Appendix S1) for specimens lacking GPS coordinatesusing standard georeferencing techniques (Chapman& Wieczorek, 2006). We used only records withmaximum error distances under 5 km to match theresolution of the climatic layers. We adhere to thetaxonomic nomenclature proposed by Soleglad & Fet(2008), but refer readers to Williams (1980) for alter-native classifications that are still commonly used.

SPECIES DISTRIBUTION MODELLING

We used the program Maxent v.3.1.0 (Phillips, Dudik& Schapire, 2004) to develop SDMs for the 13 scor-pion species. The Maxent algorithm has been shownto perform well in a comprehensive assessment ofmodelling approaches (Elith et al., 2006), and it usespresence-only species occurrence records and environ-mental data in the form of GIS layers to approximateenvironmental requirements for a species by findingthe distribution that maximized entropy, subject toconstraints imposed by the occurrence records. Inother words, the values of the environmental (in thiscase climatic) variables act as constraints on theunknown distribution, forcing the mean and varianceof the output SDMs to approximate that of the occur-rence data values.

Current (0 kya) and LGM SDMs were developed foreach of the 13 scorpion species using the logisticregression output of Maxent. For 0 kya climatic con-ditions, models were reconstructed using 19 globalbioclimatic layers (temperature and precipitationinformation) downloaded from the WorldClim data-base (Hijmans et al., 2005) at a resolution of 2.5 min.Current models were projected onto two general cir-culation models simulated for the LGM at approxi-mately 21 kya, the Community Climate SystemModel (CCSM; Otto-Bliesner et al., 2006) and theModel for Interdisciplinary Research on Climate(MIROC v.3.2; Hasumi & Emori, 2004). Since therecan be slight variation in SDMs depending on whichoccurrence record is used to seed each run, we createdsix current models and projected them on threeCCSM and three MIROC models for each species.SDMs were then concatenated and averaged inArcMap v.9.2 (ESRI, Inc.) to form consensus modelsfor each species under 0 kya and LGM climates, thuscapturing some of the intrinsic variation betweenmodel runs.

We used random seeding and default settings (regu-larization = 1; convergence threshold = 0.00001, itera-

tions = 500) in Maxent, with the exception of anadditional 0 kya run for each species used for modelevaluation where the random test percentage was setto 25%. To visually depict models, we used theaverage lowest training presence (LTP) threshold(average among models) for each species (with 100%of the data used for training) so that nearly all areaspredicted by SDMs would have logistic probabilityvalues greater than that of the occurrence record withthe smallest value. For better visualization of ourmodels, we averaged the LTP and the highest pres-ence value for each species and imposed this value asan additional threshold. This latter threshold wasused only to better visually represent SDMs and wasnot used in area or centroid calculations (see below).Models were validated using the Receiver OperatingCharacteristic for its area under the curve (AUC)value.

Our methodology outlined above and SDMs ingeneral are built on a number of assumptions dis-cussed elsewhere (e.g. Nogués-Bravo, 2009) but thatwarrant mentioning again. When projecting present-day models to past climatic landscapes such as theLGM, a depiction of suitable areas at the LGM isobtained for a species that matches habitat prefer-ences in the present-day (Peterson & Nyári, 2008).Accordingly, this method assumes niche stabilitywithin the species being analysed (Nogués-Bravo,2009). The absence of fossil data for scorpions acrossthe BCP prevents us from directly testing thisassumption of niche stability. Whereas many studieshave found that species niches have been stable sincethe LGM (e.g. Martínez-Meyer, Peterson & Hargrove,2004; Martínez-Meyer & Peterson, 2006), a growingnumber of recent studies have found contrasting evi-dence that suggests niche stability may vary (Elith,Kearney & Phillips, 2010), even over relatively shorttime scales (Ahmadzadeh et al., 2013). Given theevolutionary antiquity of scorpions (derived fromamphibious ancestors over 400 Mya) and long historyof several species in south-western North America(Bryson et al., 2013; Bryson, Savary & Prendini,2013), we make the critical assumption that ecologicalniches of the 13 species in the BCP have been con-served over the past 21 kya, but acknowledge that ourresults and discussion are built up from the founda-tion of this untested and potentially erroneousassumption.

AREA AND CENTROID CALCULATIONS

To qualitatively compare SDMs between species, thetotal size (km2) as well as the latitude and longitudeof centroid position (the estimated centre of each ofSDM) were calculated for each scorpion SDM inArcMap. Area was calculated by converting Maxent



http://earth.google.com

http://earth.google.com

output ascii files to rasters consisting of integers andthen transforming the rasters to polygons using theLTP. All grid cells with values greater than the LTPwere included in the polygons, for which their indi-vidual areas were calculated and summed for eachspecies. Centroids were calculated by converting eachgrid cell above the LTP threshold to a point (using theraster to point function in ArcToolbox) and averagingthe latitude and longitude of each point for eachspecies. Distances between centroids were calculatedusing the Haversine Formula, which incorporates cur-vature of the earth by giving great-circle distancesbetween two points on a sphere. Small to no shifts insuitable habitat from the LGM to the present amongthe scorpions suggest stability in the distributions ofthose species. All spatial analyses were conductedusing the Albers Equal Area Conic map projection.

RESULTSMODEL PERFORMANCE

We compiled 736 total occurrence records (Table 1).Current SDMs projected to CCSM and MIROCclimate layers yielded different reconstructions forthe LGM. MIROC layers produced SDMs thatdepicted more dramatic geographical range shiftscompared with CCSM layers, but often to the degreethat there was no or little suitable habitat anywhereacross the BCP during the LGM. This has beenobserved in other studies (e.g. Habel et al., 2010;Oláh-Hemmings et al., 2010; McGuire & Davis, 2013)and may be related to sensitivity of the modelsto differences in the modelled climate of the LGMbased on sea ice levels (MIROC simulates warmerand wetter conditions compared with CCSM; Otto-Bliesner et al., 2006). We therefore chose to use onlythe CCSM SDMs and excluded MIROC models fromsubsequent analyses (however, MIROC models areprovided in Appendix S2). Current SDMs had highpredictive success rates and performed significantlybetter than random in Receiver Operating Character-istic analyses, with AUC scores for testing datagreater than 0.95 (Table 1).

GENERAL DISTRIBUTIONAL PATTERNS

ALONG THE BCP

Current SDMs for the 13 species on the BCP fellprimarily into four regional categories: (1) northernpeninsular (extending into coastal California and theMojave Desert), (2) mid-peninsular, (3) southern pen-insular and (4) the Cape Region (summarized inFig. 2; shown fully in Appendices S2–S5). Based onthe models, the northern peninsular group consistedof three species with relatively large ranges thatextended from the mid-peninsula well into portions

of central California and southern Nevada. Thisgroup displayed the least congruence among suitableregions; high climatic suitability in the northdecreased southward at approximately 30.5°N, 25°Nand 26°N in Anuroctonus pococki, Hoffmanniuspuritanus and Kochius hirsuticauda, respectively(Appendix S7). At the mid-peninsula, climatic suit-ability within all three species quickly peaked atabout 24°N, but then began to drop off at differentlocations between 29°N and 30°N (Appendix S8). TheSDMs for the third group, comprising four specieswith relatively small ranges in the southern penin-sula between Bahia Concepción and Cabo SanLucas, depicted a nearly linear relationship betweenclimatic suitability and latitude, with high climaticsuitability found within each species at the southerntip of the peninsula that then steadily declined tothe north until it reached a sharp decrease at 26°Nfor each species (Appendix S9). Models suggestedthat the four species in the Cape Region group hadthe smallest ranges (all < 23 000 km2) and that eachshowed a nearly linear decline in climatic suitabilitystarting at the southern tip of the Cape to between24°N and 25°N (Appendix S10). Comparing trendsamong all SDMs revealed three areas along the BCPwhere climatic suitability tended to quickly shiftwith latitude: approximately at 24°N, 26°N and30°N.

The predicted habitat for all but one of the 13species was remarkably stable across the BCP sincethe LGM. The general trend in the distributions ofthese species was high climatic suitability alonga north–south axis of the peninsula flanked bytransitional areas of less than 1° latitude wheresuitability quickly decreased from high to low (sum-marized in Fig. 3; shown fully in Appendices S6–S9).The predicted habitat of Kochius bruneus was theonly deviation of this pattern. For this species,habitat at the LGM exhibited a change in centroid

Figure 2. Species distribution models based on currentand last glacial maximum (LGM) conditions for repre-sentative scorpion species distributed throughout thenorthern (Hoffmannius puritanus), middle (Hadrurusconcolorous), southern (Hoffmannius vittatus) and Cape(Bioculus caboensis) regions of the Baja California penin-sula. The CCSM (our preferred model) and MIROC LGMmodels are shown for comparison. Shading representsprobability of occurrence, with darkest shading being mostsuitable and lightest shading representing unsuitablehabitat. Black dots indicate occurrence records used in themodelling and white stars indicate model area centroids.Species distribution models constructed for the remainingnine scorpion species from the Baja California peninsulaare presented in Appendices S3–S6.

▶



position of 260.6 km and a change in area of85 479.2 km2 (Appendix S4). The change in centroidposition of the other BCP species (Fig. 4) rangedfrom 5.04 to 260.6 km (mean = 85.89 km), andchange in area ranged from −104 606.4 to 85 479.16km2 (mean = −16 251.98 km2).

DISCUSSIONPLEISTOCENE DISTRIBUTIONAL STASIS

Our results suggest that the suitable habitat for scor-pion species in the BCP has remained remarkablystable since the Late Pleistocene. This finding is in

Figure 3. Correlation between latitude (x axis) and habitat suitability (y axis) for representative scorpion speciesdistributed throughout the northern (Hoffmannius puritanus), middle (Hadrurus concolorous), southern (Hoffmanniusvittatus) and Cape (Bioculus caboensis) regions of the Baja California peninsula. Grey diamonds represent individual gridcells from Maxent species distribution models. The black line is a moving average trendline based on the average of every200 grid cells. Correlation charts for the remaining nine scorpion species from the Baja California peninsula are presentedin Appendices S7–S10.



stark contrast to dramatic distributional shifts sincethe Late Pleistocene inferred for arid-adapted taxaelsewhere in western North America, includingheteromyid rodents (Jezkova et al., 2009), rattle-snakes and gnatcatchers (Waltari et al., 2007),phrynosomatid lizards (T. Jezkova, unpubl. data) andinvertebrates (Wilson & Pitts, 2012) including scorpi-ons (Graham et al., 2013a, 2013b). This discrepancyhas several potential explanations. First, the proxim-ity of the BCP to ocean currents may have influencedthis pattern of distributional stasis in scorpions in theBCP. Oceans are well known to buffer coastal climates(Rahmstorf, 2002), effectively acting as large sourcesand sinks for latent heat. Stewart & Lister (2001)hypothesize that proximity to the relatively temper-ate conditions of oceans could have allowed for per-sistence of coastal coniferous forest refugia innorthern Europe during the Late Pleistocene. In asimilar fashion, the BCP, flanked by the Pacific Oceanto the west and the Sea of Cortés to the east, couldalso be predisposed to such a coastal effect. The SDMsof Anuroctonus pococki, a widely distributed speciesacross the northern BCP and into coastal California,suggest that this species’ suitable habitat remainedremarkably stable from the LGM onward (AppendixS2), lending additional support to the ocean-bufferinghypothesis. Distributions of mid-latitude species (c.40–20°N), if flanked by considerable bodies of water,might therefore be somewhat resistant to the oftensevere range shifts and contractions associated withglacial cycles.

The discrepancy between relatively stable distribu-tions of scorpions in the BCP compared with arid-adapted taxa in western North America might also beexplained by the more southern latitude of the BCP.The impacts of Pleistocene climatic fluctuations onspecies distributions are generally thought to bestronger as the distance from the equator increases(Hewitt, 2004). SDMs for species in the south-westernUSA and mainland Mexico at approximately the samelatitudes as the BCP often fail to predict dramaticdistributional shifts since the Late Pleistocene (e.g.Peterson, Martínez-Meyer & González-Salazar, 2004;McGuire et al., 2007; Oláh-Hemmings et al., 2010), orat least do not show the significant loss of habitat inthe north and massive shifts of suitable habitat to thesouth as observed in arid-adapted taxa at more north-ern latitudes (e.g. T. Jezkova, unpubl. data; Waltariet al., 2007; Jezkova et al., 2009; Wilson & Pitts, 2012;Graham et al., 2013a). These former studies providesupport for a latitudinal distance-from-the-equatorexplanation for distributional stasis in scorpionsin the BCP as an alternative to an ocean-bufferinghypothesis. In fact, the combination of both factorsmay explain our observed pattern of stable habitat forscorpion species in the BCP since the Late Pleistocene.However, the ocean-buffering and distance-from-equator hypotheses are both in conflict with fossil datafrom early Holocene packrat middens, which suggestthat vegetation assembles on the BCP have in fact notbeen resilient to range shifts (Rhode, 2002).

SCORPION SDMs AND BIOGEOGRAPHICAL

HISTORY IN THE BCP

Emerging genetic information from species along theBCP indicate that the biogeographical history of theregion is complex (Riddle & Hafner, 2006). Numeroustrans-peninsular genetic discontinuities have beenexplained by phenomena such as ephemeral trans-peninsular seaways (Minch & Leslie, 1991; Upton &Murphy, 1997; Riddle et al., 2000; Riddle, Hafner &Alexander, 2000a, b; Murphy & Aguirre-Léon, 2002;Mulcahy & Macey, 2009), abrupt ecological and cli-matic changes (Grismer, 2002), and unusual dispersalevents (Wood, Fisher & Reeder, 2008). Our scorpionSDMs add another piece to the puzzling biotic historyof the BCP.

Perhaps the most interesting of the genetic diver-gences across the BCP, especially in light of ourscorpion SDMs, are three trans-peninsular breaksoutlined and described by Lindell et al. (2005): themid-peninsular break, the Loreto break and theIsthmus of La Paz break (Fig. 1). Divergences in theseareas have been substantially debated, with the mid-peninsular break being the most provocative, anddiffering degrees of genetic divergence in these areas

Figure 4. Box and whisker plots depicting the change incentroid position (shaded boxes, km) and change in area(white boxes, km2) between current (0K) and last glacialmaximum (LGM) species distribution models for 13scorpion species distributed along the Baja Californiapeninsula.



have complicated molecular dating estimates (Leaché,Crews & Hickerson, 2007). Most authors, however,posit multiple or reoccurring seaways to explain thevariety of genetic divergences in these regions and thecorresponding array of estimated divergence dates(Lindell et al., 2005; Crews & Hedin, 2006; Leachéet al., 2007; Hafner & Riddle, 2011). Although specu-lated early in discussions of biotic diversification inthe region (Grismer, 2002), relatively little considera-tion has been given to the hypothesis that vicarianceof terrestrial lineages could be due to climatic con-straints, either alone or synergistically with seaways.

Twelve of our 13 scorpion SDMs show sharp tran-sitions in habitat suitability almost precisely acrossthe three trans-peninsular breaks, providing some ofthe first evidence that long-term climatic stabilitymay have produced patterns of divergence betweenspecies. On the Isthmus of La Paz, for example,scorpion SDMs reveal two potential sister speciesrelationships, perhaps associated with the Isthmusof La Paz break: Kochius punctipalpi and Hadrurushirsutus in the Cape Region (Appendix S5), andKochius bruneus and Hadrurus concolorous (Appen-dix S3) found north of the Cape. The SDMs of eachof these species display a sharp decrease in climaticsuitability in the area of the La Paz break. Similartrends are apparent in SDMs of other scorpionspecies in the vicinity of the mid-peninsular andLoreto breaks (Appendices S6–S8). If patterns of cli-matic and habitat stability across this admittedlyrelatively short time slice from the LGM to presentwere expanded to a considerably deeper timeframe,then we find it conceivable that genetic discontinui-ties could have arisen in association with localizedadaptation to differential climate regimes. Such apattern, however, would be contradictory to the con-clusions of Lindell, Ngo & Murphy (2006), whostated that climate-induced habitat fragmentationcould not adequately explain mid-peninsular geneticbreaks.

Although our SDMs suggest that vicariance acrossthe BCP could be due to climatic constraints, it seemsdifficult to reconcile accumulation of congruentgenetic breaks across multiple taxa (including marinefishes, Riginos, 2005) with a variety of ecologicaltraits and presumably niche requirements withoutstill invoking additional concerted events such as thepreviously mentioned trans-peninsular seaways.Nonetheless, hypotheses like that of the VizcaínoSeaway, for which little geological evidence exists(Lindell et al., 2006; but see Helenes & Carreño, 1999;Carreño & Helenes, 2002; and Ledesma-Vázquez,2002), might be due for reappraisal, and the possibil-ity that climate was largely responsible for thecomplex phylogeographical patterns along the BCPshould be given more consideration.

ACKNOWLEDGEMENTS

We thank S. C. Williams, V. F. Lee and C. E. Griswoldfor granting loans from the California Academy ofSciences. We are grateful to J. R. Jaeger for providingassistance with collecting and transporting speci-mens; T. Jezkova for support with distribution mod-elling; L. Hancock for georeferencing assistance; M. A.Wall for kindly allowing access to the scorpions at theSan Diego Natural History Museum; M. E. Solegladfor scorpion occurrence records; and E. E. Saupe, M.E. Eckstut and T. Jezkova for helpful reviews ofearlier versions of the manuscript. We thank threeanonymous reviewers for their important comments.

REFERENCES

Ahmadzadeh F, Flecks M, Carretero MA, Böhme W,Ilgaz C, Engler JO, Harris DJ, Üzùm N, Rödder D.2013. Rapid lizard radiation lacking niche conservatism:ecological diversification within a complex landscape.Journal of Biogeography 40: 1807–1818.

Avise JC. 2000. Phylogeography: the history and formation ofspecies. Cambridge, MA: Harvard University Press.

Betancourt JL, Van Devender TR, Martin PS. 1990.Packrat middens: the last 40,000 years of biotic change.Tucson, AZ: University of Arizona Press.

Bryson RW, Jaeger JR, Lemos-Espinal JA, Lazcano D.2012. A multilocus perspective on the speciation history ofa North American aridland toad (Anaxyrus punctatus).Molecular Phylogenetics and Evolution 64: 393–400.

Bryson RW, Riddle BR, Graham MR, Smith BT,Prendini L. 2013. As old as the hills: montane scorpions insouthwestern North America reveal ancient associationsbetween biotic diversification and landscape history. PLoSONE 8: e52822.

Bryson RW, Savary WE, Prendini L. 2013. Biogeographyof scorpions in the Pseudouroctonus minimus complex(Vaejovidae) from south-western North America: implica-tions of ecological specialization for pre-Quaternary diver-sification. Journal of Biogeography 40: 1850–1860.

Carreño AL, Helenes J. 2002. Geology and ages of theislands. In: Case TJ, Cody ML, Ezcurra E, eds. A new islandbiogeography of the Sea of Cortés. New York: Oxford Uni-versity Press, 14–40.

Chapman AD, Wieczorek J. 2006. Guide to best practices forgeoreferencing. BioGeomancer Consortium. Copenhagen:Global Biodiversity Information Facility.

Crews SC, Hedin M. 2006. Studies of morphological andmolecular phylogenetic divergence in spiders (Araneae:Homalonychus) from the American southwest, includingdivergence along the Baja California peninsula. MolecularPhylogenetics and Evolution 38: 470–487.

Due AD, Polis GA. 1986. Trends in scorpion diversity alongthe Baja California peninsula. American Naturalist 128:460–468.

Elith J, Graham CH, Anderson RP, Dudik M, Ferrier S,Guisan A, Hijmans RJ, Huettmann F, Leathwick JR,



Lehmann A, Li J, Lohmann LG, Loiselle BA, ManionG, Moritz C, Nakamura M, Nakazawa Y, Overton JM,Peterson AT, Phillips SJ, Richardson K, Scachetti-Pereira R, Schapire RE, Soberon J, Williams S, WiszMS, Zimmermann NE. 2006. Novel methods improveprediction of species’ distributions from occurrence data.Ecography 29: 129–151.

Elith J, Kearney M, Phillips S. 2010. The art of modellingrange-shifting species. Methods in Ecology and Evolution 1:330–342.

Gantenbein B, Fet V, Barker MD. 2001. MitochondrialDNA reveals a deep, divergent phylogeny in Centruroidesexilicauda (Wood, 1963) (Scorpiones: Buthidae). In: Fet F,Selden PA, eds. Scorpions 2001 in memoriam Gary A. Polis.London: British Arachnological Society, 235–244.

Graham MR, Jaeger JR, Prendini L, Riddle BR. 2013a.Phylogeography of the Arizona hairy scorpion (Hadrurusarizonensis) supports a model of biotic assembly in theMojave Desert and adds a new Pleistocene refugium.Journal of Biogeography 40: 1298–1312.

Graham MR, Jaeger JR, Prendini L, Riddle BR. 2013b.Phylogeography of Beck’s Desert Scorpion, Paruroctonusbecki, reveals Pliocene diversification in the Eastern Cali-fornia Shear Zone and postglacial expansion in the GreatBasin Desert. Molecular Phylogenetics and Evolution 69:502–513.

Grismer LL. 1994. The origin and evolution of the peninsularherpetofauna of Baja California, México. HerpetologicalNatural History 2: 51–106.

Grismer LL. 2002. A re-evaluation of the evidence for amid-Pleistocene mid-peninsular seaway in Baja California:a reply to Riddle et al. Herpetological Review 33: 15–16.

Habel JC, Schmitt T, Meyer M, Finger A, Rödder D,Assmann T, Zachos FE. 2010. Biogeography meets con-servation biology: the life history of the highly endangeredpostglacial relict butterfly Lycaena helle. Biological Journalof the Linnean Society 101: 155–168.

Hafner DJ, Riddle BR. 2011. Boundaries and barriersof North American warm deserts: an evolutionary perspec-tive. In: Upchurch P, McGowan AJ, Slater CSC, eds.Palaeogeography and palaeobiogeography: biodiversity inspace and time. Boca Raton, FL: CRC Press, 75–114.

Hasumi H, Emori S. 2004. K-1 coupled model (MIROC)description, Center for Climate System Research. Tokyo:University of Tokyo.

Helenes J, Carreño AL. 1999. Neogene sedimentaryevolution of Baja California in relation to regionaltectonics. Journal of South American Earth Science 12:589–605.

Hess HH. 1962. History of ocean basins. In: Engel AEJ,James HL, Léonard BF, eds. Petrological studies: a volumein honor of A. F. Buddington. New York: Geological Societyof America, 599–620.

Hewitt GM. 2004. Genetic consequences of climatic oscilla-tions in the Quaternary. Philosophical Transactions of TheRoyal Society B 359: 183–195.

Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A.2005. Very high resolution interpolated climate surfaces for

global land areas. International Journal of Climatology 25:1965–1978.

Jezkova T, Jaeger JR, Marshall ZL, Riddle BR. 2009.Pleistocene impacts on the phylogeography of the desertpocket mouse (Chaetodipus penicillatus). Journal ofMammalogy 90: 306–320.

Klicka J, Zink RM. 1997. The importance of recent ice agesin speciation: a failed paradigm. Science 277: 1666–1669.

Leaché AD, Crews SC, Hickerson MJ. 2007. Two waves ofdiversification in mammals and reptiles of Baja Californiarevealed by hierarchical Bayesian analysis. Biology Letters3: 646–650.

Ledesma-Vázquez J. 2002. A gap in the Pliocene invasion ofseawater to the Gulf of California. Revista Mexicana deCiencias Geológicas 19: 145–151.

Lindell J, Méndez-de la Cruz FR, Murphy RW. 2005.Deep genealogical history without population differentia-tion: discordance between mtDNA and allozyme divergencein the zebratailed lizard (Callisaurus draconoides). Molecu-lar Phylogenetics and Evolution 36: 682–694.

Lindell J, Méndez-de la Cruz FR, Murphy RW. 2008.Deep biogeographical history and cytonuclear discordancein the black-tailed brush lizard (Urosaurus nigricaudus) ofBaja California. Biological Journal of the Linnean Society94: 89–104.

Lindell J, Ngo A, Murphy RW. 2006. Deep genealogies andthe mid-peninsular seaway of Baja California. Journal ofBiogeography 33: 1327–1331.

Lonsdale P. 1989. Geology and tectonic history of the Gulf ofCalifornia. In: Winterer EL, Hussong DM, Decker RW, eds.The Eastern Pacific Ocean and Hawaii. Boulder, CO: TheGeological Society of America, 499–521.

Macfadden BJ. 2006. Extinct mammalian biodiversity of theancient New World tropics. Trends in Ecology and Evolution21: 157–165.

Martínez-Meyer E, Peterson AT, Hargrove WW. 2004.Ecological niches as stable distributional constraints onmammal species, with implications for Pleistocene extinc-tions and climate change projections for biodiversity. GlobalEcology and Biogeography 13: 305–314.

Martínez-Meyer E, Peterson AT. 2006. Conservatism ofecological niche characteristics in North American plantspecies over the Pleistocene-to-Recent transition. Journal ofBiogeography 33: 1779–1789.

McGuire JA, Linkem CW, Koo MS, Hutchinson DW,Lappin AK, Orange DI, Lemos-Espinal J, Riddle BR,Jaeger JR. 2007. Mitochondrial introgression and incom-plete lineage sorting through space and time: phylogeneticsof crotaphytid lizards. Evolution 61: 2879–2897.

McGuire JL, Davis EB. 2013. Using the palaeontologicalrecord of Microtus to test species distribution models andreveal responses to climate change. Journal of Biogeography40: 1490–1500.

Minch J, Leslie T. 1991. The Baja Highway: a geology andbiology field guide for the Baja traveler. San JuanCapistrano, CA: John Minch and Associates.

Mulcahy DG, Macey JR. 2009. Vicariance and dispersalform a ring distribution in nightsnakes around the Gulf of



California. Molecular Phylogenetics and Evolution 53: 537–546.

Murphy RW. 1983. Paleobiogeography and genetic differen-tiation of the Baja California herpetofauna. OccassionalPapers of the California Academy of Sciences 137: 1–48.

Murphy RW, Aguirre-Léon G. 2002. Nonavian reptiles:origins and evolution. In: Case T, Cody M, Ezcurra E, eds.A new island biogeography in the Sea of Cortes. Oxford:Oxford University Press, 181–220.

Nogués-Bravo D. 2009. Predicting the past distribution ofspecies climatic niches. Global Ecology and Biogeography18: 521–531.

Oláh-Hemmings V, Jaeger JR, Sredl MJ, Schlaepfer MA,Jennings RD, Drost CA, Bradford DF, Riddle BR.2010. Phylogeography of declining relict and lowlandleopard frogs in the desert Southwest of North America.Journal of Zoology 280: 343–354.

Otto-Bliesner BL, Brady EC, Clauzet G, Tomas R, LevisS, Kothavala Z. 2006. Last Glacial Maximum and Holo-cene climate in CCSM3. Journal of Climatology 19: 2526–2544.

Peterson AT, Martínez-Meyer E, González-Salazar C.2004. Reconstructing the Pleistocene geography of theAphelocoma jays (Corvidae). Diversity and Distributions 10:237–246.

Peterson AT, Nyári AS. 2008. Ecological niche conservatismand Pleistocene refugia in the thrush-like mourner,Schiffornis sp., in the neotropics. Evolution 62: 173–183.

Phillips SJ, Dudik M, Schapire RE. 2004. A maximumentropy approach to species distribution modelling. Proceed-ings of the 21st International Conference on Machine Learn-ing 69: 83–90.

Rahmstorf S. 2002. Ocean circulation and climate during thepast 120,000 years. Nature 419: 207–214.

Rhode D. 2002. Early Holocene juniper woodland and chap-arral taxa in the central Baja California peninsula, Mexico.Quaternary Research 57: 102–108.

Riddle BR, Hafner DJ. 2006. A step-wise approach tointegrating phylogeographic and phylogenetic biogeographicperspectives on the history of a core North American warmdeserts biota. Journal of Arid Environments 66: 435–461.

Riddle BR, Hafner DJ, Alexander LF. 2000a.Phylogeography and systematics of the Peromyscuseremicus species group and the historical biogeography ofNorth American warm regional deserts. MolecularPhylogenetics and Evolution 17: 145–160.

Riddle BR, Hafner DJ, Alexander LF. 2000b. Comparativephylogeography of Bailey’s pocket mouse (Chaetodipusbaileyi) and the Peromyscus eremicus species group: histori-cal vicariance of the Baja California Peninsular Desert.Molecular Phylogenetics and Evolution 17: 161–172.

Riddle BR, Hafner DJ, Alexander LF, Jaeger JR. 2000.Cryptic vicariance in the historical assembly of a BajaCalifornia Peninsular Desert biota. Proceedings of theNational Academy of Sciences 97: 14438–14443.

Riginos C. 2005. Cryptic vicariance in Gulf of Californiafishes parallels vicariant patterns found in Baja Californiamammals and reptiles. Evolution 59: 2678–2690.

Saupe EE, Barve V, Myers CE, Soberón J, Barve N,Hensz CM, Peterson AT, Owens HL, Lira-Noriega A.2012. Variation in niche and distribution model perfor-mance: the need for a priori assessment of key causalfactors. Ecological Modelling 237–238: 11–22.

Savage JM. 1960. Evolution of a peninsular herpetofauna.Systematic Zoology 9: 184–212.

Sissom WD. 1991. Systematic studies on the nitidulus groupof the genus Vaejovis, with descriptions of seven new species(Scorpiones, Vaejovidae). Journal of Arachnology 19: 4–28.

Soleglad ME, Fet V. 2004. The systematics of scorpionsubfamily Uroctoninae (Scorpiones: Chactidae). RevistaIbérica de Aracnología 10: 81–128.

Soleglad ME, Fet V. 2008. Contributions to scorpion system-atics. III. Subfamilies Smeringurinae and Syntropinae(Scorpiones: Vaejovidae). Euscorpius 71: 1–115.

Spencer JE, Normark WR. 1989. Neogene plate-tectonicevolution of the Baja California Sur continental margin andthe southern Gulf of California, Mexico. In: Winterer EL,Hussong DM, Decker RW, eds. The Eastern Pacific Oceanand Hawaii. Boulder, CO: Geological Society of America,489–498.

Stahnke HL. 1968. Some diplocentrid scorpions from BajaCalifornia del Sur, Mexico. Proceedings of the CaliforniaAcademy of Sciences 4: 273–320.

Stahnke HL. 1969. A review of Hadrurus scorpions(Vejovidae). Entomological News 80: 54–65.

Stewart JR, Lister AM. 2001. Cryptic northern refugia andthe origins of the modern biota. Trends in Ecology andEvolution 16: 608–613.

Stock JM, Hodges KV. 1989. Pre-Pliocene extension aroundthe Gulf of California and the transfer of Baja California tothe Pacific Plate. Tectonics 8: 99–115.

Upton DE, Murphy RW. 1997. Phylogeny of the side-blotchedlizards (Phrynosomatidae: Uta) based on mtDNA sequences:support for a midpeninsular seaway in Baja California.Molecular Phylogenetics and Evolution 8: 104–113.

Waltari E, Hijmans RJ, Peterson AT, Nyári Á, PerkinsSL, Guralnick RP. 2007. Locating Pleistocene refugia:comparing phylogeographic and ecological niche model pre-dictions. PLoS ONE 2: e563.

Williams SC. 1970a. A systematic revision of the giant hairyscorpion genus Hadrurus. Occasional Papers of the Califor-nia Academy of Sciences 87: 1–62.

Williams SC. 1970b. Scorpion fauna of Baja California,Mexico: eleven new species of Vejovis (Scorpionida: Vejovi-dae). Proceedings of the California Academy of Sciences 37:275–332.

Williams SC. 1970c. New scorpions belonging to theEusthenura group of Vejovis from Baja California, Mexico(Scorpionida, Vejoviae). Proceedings of the CaliforniaAcademy of Sciences 37: 395–418.

Williams SC. 1971. New and little known scorpions belongingto the punctipalpi group of the genus Vaejovis from BajaCalifornia, Mexico, and adjacent area (Scorpionida:Vaejovidae). Wasmann Journal of Biology 29: 37–63.

Williams SC. 1974. A new genus of North American scorpionswith a key to the North American genera of Vaejovidae.



Proceedings of the California Academy of the Sciences 40:1–16.

Williams SC. 1980. Scorpions of Baja California, Mexico andadjacent islands. Occasional Papers of the CaliforniaAcademy of Sciences 135: 1–127.

Williams SC, Lee VF. 1975. Diplocentrid scorpions fromBaja California Sur, Mexico (Scorpionida: Diplocentridae).Occasional Papers of the California Academy of Sciences115: 1–27.

Wilson JS, Pitts JP. 2012. Identifying Pleistocene refugia inNorth American cold deserts using phylogeographic analy-ses and ecological niche modelling. Diversity and Distribu-tions 18: 1139–1152.

Wood DA, Fisher RN, Reeder TW. 2008. Novel patterns ofhistorical isolation, dispersal, and secondary contact acrossBaja California in the rosy boa (Lichanura trivirgata).Molecular Phylogenetics and Evolution 46: 484–502.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:

Appendix S1. Coordinates (in decimal degrees) used to generate species distribution models.Appendix S2. Maxent models created using data from the Model for Interdisciplinary Research on Climate(MIROC).Appendix S3. Species distribution models created with Maxent for three scorpion species distributed insouthern California, USA, and northern Baja California, Mexico: (A, B) Anuroctonus pococki, (C, D)Hoffmannius puritanus and (E, F) Kochius hirsuticauda. Maps A, C and E are present predictions, with blackdots indicating occurrences used in the modeling; maps B, D and F are Maxent predictions at the LGM. Whitestars indicate model area centroids and shading represents probability of occurrence, with darkest shadingbeing most suitable and lightest shading representing unsuitable habitat (see Methods for modelling details).Appendix S4. Species distribution models created with Maxent for three scorpion species distributed along thecentral portion of the BCP, Mexico: (A, B) Hadrurus concolorous, (C, D) Hadrurus pinteri, (E, F) Kochiusbruneus. Maps A, C and E are present predictions, with black dots indicating occurrences used in the modelling;maps B, D and F are Maxent predictions at the LGM. White stars indicate model area centroids and shadingrepresents probability of occurrence, with darkest shading being most suitable and lightest shading represent-ing unsuitable habitat (see Methods for modelling details).Appendix S5. Species distribution models created with Maxent for scorpion species distributed along thesouthern portion of Baja California, Mexico: (A, B) Bioculus comondae, (C, D) Hoffmannius diazi, (E, F)Hoffmannius vittatus, (G, H) Nullibrotheas allenii. Maps A, C, E and G are present predictions, with black dotsindicating occurrences used in the modeling; maps B, D, F and H are Maxent predictions at the LGM. Whitestars indicate model area centroids and shading represents probability of occurrence, with darkest shadingbeing most suitable and lightest shading representing unsuitable habitat (see Methods for modelling details).Appendix S6. Species distribution models created with Maxent for three scorpion species endemic to the BajaCalifornia Cape Region, Mexico: (A, B) Bioculus caboensis, (C, D) Hadrurus hirsutus, (E, F) Kochius punctipalpi.Maps A, C and E are present predictions, with black dots indicating occurrences used in the modelling; mapsB, D and F are Maxent predictions at the LGM. White stars indicate model area centroids and shadingrepresents probability of occurrence, with darkest shading being most suitable and lightest shading represent-ing unsuitable habitat (see Methods for modelling details).Appendix S7. Correlation between latitude (x axis) and habitat suitability (y axis) for three scorpion speciesoccurring in northern Baja California. Grey diamonds represent individual grid cells from Maxent speciesdistribution models, and the black line is a moving average trendline based on the average of every 200 gridcells.Appendix S8. Correlation between latitude (x axis) and habitat suitability (y axis) for three scorpion speciesoccurring in middle of the BCP. Grey diamonds represent individual grid cells from Maxent species distributionmodels, and the black line is a moving average trendline based on the average of every 200 grid cells.Appendix S9. Correlation between latitude (x axis) and habitat suitability (y axis) for three scorpion speciesoccurring in southern Baja California. Grey diamonds represent individual grid cells from Maxent speciesdistribution models, and the black line is a moving average trendline based on the average of every 200 grid cells.Appendix S10. Correlation between latitude (x axis) and habitat suitability (y axis) for three scorpion speciesendemic to the Baja California Cape Region. Grey diamonds represent individual grid cells from Maxent speciesdistribution models, and the black line is a moving average trendline based on the average of every 200 gridcells.



late pleistocene to holocene distributional stasis in ... · late pleistocene to holocene...

Documents