Tectonics, climate, and the rise and demise ofcontinental aquatic species richness hotspotsThomas A. Neubauer1, Mathias Harzhauser, Elisavet Georgopoulou, Andreas Kroh, and Oleg Mandic

Geological-Paleontological Department, Natural History Museum Vienna, 1010 Vienna, Austria

Edited by Ole Seehausen, University of Bern, Bern, Switzerland, and accepted by the Editorial Board July 24, 2015 (received for review February 26, 2015)

Continental aquatic species richness hotspots are unevenly dis-tributed across the planet. In present-day Europe, only two centersof biodiversity exist (Lake Ohrid on the Balkans and the CaspianSea). During the Neogene, a wide variety of hotspots developed ina series of long-lived lakes. The mechanisms underlying the presenceof richness hotspots in different geological periods have not beenproperly examined thus far. Based on Miocene to Recent gastropoddistributions, we show that the existence and evolution of suchhotspots in inland-water systems are tightly linked to the geo-dynamic history of the European continent. Both past and presenthotspots are related to the formation and persistence of long-livedlake systems in geological basins or to isolation of existing inlandbasins and embayments from themarine realm. The faunal evolutionwithin hotspots highly depends on warm climates and surface area.During the Quaternary icehouse climate and extensive glaciations,limnic biodiversity sustained a severe decline across the continentand most former hotspots disappeared. The Recent gastropoddistribution is mainly a geologically young pattern formed after theLast Glacial Maximum (19 ky) and subsequent formation of post-glacial lakes. The major hotspots today are related to long-lived lakesin preglacially formed, permanently subsiding geological basins.

biogeography | hotspot evolution | freshwater gastropods | Cenozoic |species-area relationship

The term “hotspot” is variably defined in the literature. Hot-spots may characterize regions with particularly high species

richness, high levels of endemism, high numbers of rare orthreatened species, or refer to the intensity of threat (1, 2). Inthis paper, we specifically deal with species richness hotspots andtheir evolution over geological time. Species richness patterns andchanges therein have been tightly linked to climatic processes (2–5).In some cases, however, climate may only be a secondary orindirect cause for varying species richness (6). Regional climateconditions can be strongly influenced by tectonic processes causingclosing or opening of seaways or orogenesis (7, 8). Such configu-rations determine oceanic and atmospheric circulation dynamicsand, hence, climatic dynamics (9, 10). A good example is thetriggering of the Indian monsoon at the beginning of the Neo-gene by the collision of the Indian subcontinent with Eurasia andthe uplift of the Himalaya (11). Today, half of the world’s humanpopulation is affected by this particular climate regime that wasinitiated more than 20 Mya (11).Paleogeographic settings may affect faunal distributions in-

directly (through climate), but also directly. The initiation andcessation of continental basins, along with the related develop-ment and persistence of freshwater and brackish environments,controls the possibilities for dispersal and evolutionary radiationsfor nonmarine biota. Lacustrine basins that are tectonically andecologically stable on geological time scales provide opportunitiesfor settlement and diversification (12, 13). In various cases bio-diversity maxima might be related to a basin’s geographic position,whereas climatic and physiographic parameters may influence thespecies richness and composition.Diversification patterns have been linked to tectonic evolution

for various settings (14–17), yet such a relation has not beenestablished for inland aquatic biota. Freshwater and brackish biota

are ideal research objects because they have limited possibilitiesfor dispersal (18) and must therefore overcome environmentalpressure in their habitats. In particular, species originating fromintralacustrine radiations in long-lived lakes rarely colonize otherlakes (12, 19).We analyzed the distributions of Miocene to Recent European

nonmarine gastropods to identify the dynamics and drivers of spe-cies richness hotspots through time. We tested for correlations ofspecies richness trends, large-scale faunal turnovers, and bio-diversity maxima to climatic and to physiographic parameters, aswell as geodynamic setting. The specific aim of this paper is tounravel the evolution and history of the modern distributions.

Shifting Hotspots Through TimeAs measures of diversity we chose a number of proxies: (i) thetotal number of species per million years as a proxy for theoverall trend, (ii) the maximum number of species per environ-ment as an indicator for individual systems, (iii) the mean numberof species of all environments per million years as a measure of theaverage trend, and (iv) origination, extinction, turnover rates andWhittaker’s β diversity (4) between million year bins to estimatethe degree of faunal changes. The rise and demise of richnesshotspots is shown in Figs. 1 and 2. In general, the total number ofspecies increases constantly over time (linear correlation be-tween million year bin and number of species: r = 0.866, P <0.001). This trend is occasionally interrupted by marked break-downs, e.g., in the Middle Miocene and Pleistocene (Fig. 2). Asimilar picture is presented by the maximum number of species

Significance

To our knowledge, this study is the first investigation of theevolution of species richness hotspots in continental aquaticsystems. We demonstrate the development of European rich-ness hotspots over the last 23 My based on a comprehensivedataset combining recent and fossil occurrences of gastropodspecies. We show that changes in species richness patterns canbe related to geodynamic and climatic processes. The additionof tectonics, geological time, and spatial scales to ecology andclimate is essential for understanding hotspot development ingeneral. These insights also provide a foundation to explainthe modern, uneven distribution of species richness as a whole.The pattern for Recent European faunas is a geologically youngphenomenon, triggered by the ice sheet retreat after the LastGlacial Maximum.

Author contributions: T.A.N., M.H., and E.G. designed research; T.A.N., M.H., E.G., and O.M.performed research; T.A.N., E.G., and A.K. analyzed data; and T.A.N. and E.G. wrotethe paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. O.S. is a guest editor invited by the EditorialBoard.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: thomas.neubauer@nhm-wien.ac.at.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1503992112/-/DCSupplemental.

11478–11483 | PNAS | September 15, 2015 | vol. 112 | no. 37 www.pnas.org/cgi/doi/10.1073/pnas.1503992112

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020

per environment, which peaked in the Late Miocene to Plioceneand declined toward the Pleistocene. Marked episodes of faunalturnover occurred during the Middle Miocene (15 Ma), the LateMiocene (9 Ma), the Early Pliocene (5 Ma), and the Late Plio-cene (3 Ma). Particularly, the Pliocene events are characterizedby extremely high extinction rates.Preservation and sampling biases constrain the reconstructions of

past biodiversity to some extent. As the fossil record of continentalaquatic systems is commonly biased toward large, long-lived lakes(13), which typically hold the most species, the potential bias fordetecting biodiversity hotspots is expected to be minor.

Early Miocene. This period is characterized by a relatively lowaverage species richness (Fig. 1). Three small hotspots of aboutequal magnitude and geographical extent exist, corresponding tomajor geological basins present during this interval (Fig. 3). Thewesternmost center is within the Aquitaine Basin (20), which isfamous for its rich marine Cenozoic mollusk faunas. Addition-ally, a rich freshwater to brackish gastropod fauna is recordedfrom marginal deposits (Table 1 and Table S1). A second hot-spot is located in the German Hanau-Wetterau Basin along theUpper Rhine Graben, whose origins date back to the EarlyOligocene (21). Due to partial geographical restriction and cut-off from the North Sea Basin, brackish and freshwater environ-ments developed between the Late Oligocene and the earlyMiddle Miocene (21, 22). Maximum diversity was reached duringthe Early Miocene. A third, smaller hotspot existed in the southGerman North Alpine Foreland Basin and is also a relic of aformer sea. With the retreat of the western branch of theParatethys Sea in the late Early Miocene, small brackish basinsformed within the North Alpine Foreland Basin (22, 23). TheUpper Brackish Water Molasse (UBWM) is characterized by arich, endemic gastropod fauna, whose evolution around 18 Maboosted β diversity. From this time onward, β diversity remained

at a high level of 0.6–0.7. Later diversity fluctuations are minorand largely parallel main turnover events. The overall diversityincrease in the earliest Miocene coincides with a trend towardhigher temperature and humidity after the comparatively dry andcool Oligocene (24, 25).

Middle Miocene. Overall species richness in Europe increasedtoward a maximum at the Middle Miocene Climate Optimum(MCO) (24, 26). This peak is followed by the first big Mioceneturnover phase with simultaneously increased origination andextinction rates. Turnover is related to the rise and demise ofindividual lake basins. Many of the characteristic Early Miocenefaunas decline, mostly as a result of deteriorating settings in theAquitaine and Hanau-Wetterau basins. At the same time, a hot-spot emerged on the Balkan Peninsula, comprising several fresh-water lake basins in Croatia and Bosnia and Herzegovina (Fig. 3)that form the Dinaride Lake System (27). These isolated intra-montane lakes formed as a result of the folding of the DinarideMountains (27) and accommodated rich and highly endemicmollusk faunas in the early Middle Miocene (13, 28). Decreasingglobal (24) and regional (25) temperatures, as well as decreasingregional precipitation (26) around 14 Ma, coincide with a strongdecrease of total species richness and individual richness perlake, reflecting the ongoing continentalization of the Dinaridelakes (27). Afterward, species richness throughout Europerebounded in a prolonged (5 Ma), gradual rise in the total numberof species and in the maximum richness per lake, paralleling in-creased origination rates. Simultaneous faunal diversifications oc-curred in the late Middle Miocene Bakony wetlands near Budapest(Hungary), the Paratethyan Soceni wetland fringes (Romania), andthe small Lake Steinheim (Germany).

Late Miocene. With the final retreat of the Central Paratethys bythe onset of the Late Miocene, all of the previous hotspots dis-appeared. However, a huge new hotspot developed in Lake

Fig. 1. Heat maps for the six selected temporal units, using an equal-area grid of 100-km cell size in Behrmann projection. Colors are based on the ordinaryKriging algorithm with 100-km cell size and 200-km interpolation radius performed on grid-specific species richness. Coloration is equal in all maps, scaled tothe maximum grid richness of 218 species (Late Miocene, Lake Pannon).

Neubauer et al. PNAS | September 15, 2015 | vol. 112 | no. 37 | 11479

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020

Pannon in the Pannonian Basin at that time, hosting large faunalradiations (28–30) (Fig. 3). Its individual species richness reachedits maximum at 9 Ma, coinciding with the lake’s maximum extentand an overall peak in European diversity. That peak is followedby the second major Miocene turnover phase, characterized byelevated extinction rates. From ∼9 Ma onward, the diversity ofLake Pannon declined together with its shrinking size. A smallpeak at 7 Ma reflects the evolution of faunas in the semienclosedDacian Basin and adjacent Galati Seaway in southeastern Romania,Moldova, and Ukraine, both of which yielded brackish to marginallyfreshwater conditions at that time (31). An additional, smallerhotspot developed in the Bresse–Valence graben system in south-eastern France, coinciding with its successive isolation from theMediterranean during the Late Miocene and the development ofbrackish to fluvio-lacustrine conditions (32). In addition, severalrich faunas are known from the Iberian Peninsula (e.g., LakeCalatayud-Teruel) (13). Most of these, however, are poorly stud-ied on the species level and few of these could be included in thepresent dataset. Their potential status as hotspot requires furthertaxonomic investigations of the local faunas.

Pliocene. The total number of species reached a temporary max-imum, mostly because of the presence of the rich fauna of LakeDacia (31, 33). With the isolation of the basin in the LateDacian/Middle Zanclean, ∼4 Mya, freshwater conditions startedto develop (31), offering suitable conditions for mollusk settle-ment and evolution. Two biodiversity peaks are present in thePliocene, followed by major faunal turnovers. The peak at 5 Marepresents the final phase of the Lake Pannon fauna, which isassociated with a high extinction rate, and the small but richLake Metohia in Kosovo. A second peak at 3 Ma corresponds tothe highly diverse fauna of Lake Pannon’s successor, Lake Slavonia(28, 29), and the fauna of the fluvial plains along Rioni Bay in theeastern Black Sea depression, which had brackish conditions at

that time (34) (Fig. 3). This biodiversity peak, corresponding tothe mid-Pliocene warm period (35, 36), is followed by the ex-tinction of more than 300 species, representing the biggest turn-over event in the late Cenozoic.

Pleistocene. The successive richness decline early in the Pleisto-cene coincides with global and regional cooling (24, 35). Thetotal number of species drops to less than one half of the Plio-cene maximum. With the demise of long-lived lakes by the end ofthe Pliocene, such as Dacia, Slavonia, and Transylvania lakes,the maximum number of species per lake also declines (Fig. 2).Former mega hotspots vanished and gave way to dispersed smallhotspots such as Lake Bresse in France, Lake Tiberino in Italy,and Lake Kos in Greece. Beyond that, several low-diversitycenters formed in central Europe, none of which can be attrib-uted to a bigger paleo-lake or an evolving geological basin. Thepattern is to some extent biased by the varied availability ofoutcrops and time-averaging of glacial and interglacial deposits.Preliminary data suggest that the Caspian Sea, today a very large,brackish lake, likely formed a hotspot already during the Pleis-tocene, but our knowledge on early stage faunas is limited. Thepicture for the Pleistocene is not well resolved in detail. Never-theless, the low overall diversity compared with the Pliocene isevident.

Recent. With the onset of the Ice Ages and the associated gla-ciations of large parts of northern Europe and the Alpine region(37–39), many of the former lakes and their faunas vanished.Most extant lakes and their faunas emerged after the Last GlacialMaximum (20–19 ka BP) (39). Two areas with high species num-bers develop in the Caspian Sea and in a series of lakes in thesouthern Balkans, i.e., Ohrid, Prespa, Mikri Prespa, Pamvotis,and Trichonis (40–42). Their corresponding lake basins werealready present before the Pleistocene. Another area of rela-tively high biodiversity encompasses the Curonian Lagoon, alow-brackish lake close to the Baltic Sea. Contrary to the twoother hotspots, this area only contains geologically young lakes(<19 ka). The high diversity is likely related to the lacustrine-brackish development of the Baltic Sea in the Late Pleistoceneto earliest Holocene (38) and to the following migrations betweensurrounding lakes.

What Drives Species Richness Trends and Patterns?Using different datasets and resolutions (locality based, envi-ronment based, and overall richness per million years; TablesS2–S4), we carried out regression analyses to test for relations ofspecies richness with geographic, physiographic, and climaticparameters. Although species richness is known to vary acrossgeographic settings in general, particularly latitude (3, 43, 44),our data indicate very weak geographic relations throughout alltime intervals (Table S5). In contrast, species richness trendsover time are correlated with climate changes. Linear regress-ions of richness and δ18O as proxy for global temperatureyielded highly significant correlations for total number of species(log10-transformed, r2 = 0.514, P < 0.001; with range-throughassumption: r2 = 0.465, P < 0.001), maximum richness per en-vironment (log10-transformed, r2 = 0.245, P = 0.016), number oforiginations (r2 = 0.470, P < 0.001), number of extinctions (r2 =0.189, P = 0.038), and turnover rate (r2 = 0.433, P = 0.001)(Tables S2 and S3 and Fig. S1). Global cooling and pronouncedglaciations, paired with the decline of the Paratethyan long-livedlakes by the end of the Pliocene, dramatically reduced the avail-ability of biotopes for lacustrine gastropods. Throughout the Neo-gene, continental aquatic species richness maxima in Europecorrespond to climate maxima (e.g., MCO and MPWP; Fig. 2).Additionally, we performed regression analyses for species

richnesses of 30 selected systems to assess potential correlationswith latitude, longitude, lake surface area and temporal duration

Fig. 2. Species richness trends across time. From Left to Right: mean num-ber of species per lake (dotted line); maximum number of species per lake(dashed line); total number of species (solid line); number of originations(dotted line); number of extinctions (dashed line); faunal turnover rate(dash-dotted line; summed number of originations and extinctions); totalnumber of species with range-through assumption (solid line); Whittaker’s βdiversity (solid line); deep-sea oxygen isotope record (5 pt moving average)as proxy for global temperature (24); schematic indication of important cli-matic events (MCO, Middle Miocene Climate Optimum; MPWP, Mid-Pliocenewarm period). Horizontal bars denote marked faunal turnover events. Notethe logarithmized number of species was used for the first three curves, asotherwise trends of mean species richness per lake would be indiscernible.

11480 | www.pnas.org/cgi/doi/10.1073/pnas.1503992112 Neubauer et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020

(Table S4). Linear ordinary least-square regressions indicated astatistically significant association only for area (log-log, r2 = 0.489,P < 0.001; Fig. S2 and Tables S6–S18). A multiple regression in-cluding all parameters did not improve the model (SI Materialsand Methods and Tables S6–S18). The species-area relationshipserves as a common explanatory model for varied numbers ofspecies in geographically restricted environments, such as lakesor islands, and has been confirmed by many theoretical and em-pirical studies (4, 45–47).

Geodynamics as a Primary Driver of Hotspot DevelopmentIt is crucial to distinguish between the faunal development of ahotspot and its mere existence. The regression analyses indicatethat faunal diversification correlates with climate and surface area,confirming earlier studies (3–6). However, the presence of hot-spots and their spatial distribution are rarely discussed.

Here, we demonstrate that the shifting presence of speciesrichness hotspots through time is tightly linked to the developmentof geological basins accommodating long-lived, stable fresh-water or brackish environments (Fig. 3). Particularly large, long-lived lakes such as the Late Miocene Lake Pannon or LateMiocene-Pliocene Lake Dacia presented environments that werestable across geological timescales and offered a great variety ofhabitats. In these and other such systems, intralacustrine speci-ation gave rise to many hundreds of species over time (40, 41, 48).This process created diversities far above the average of typicalshort-lived systems such as most modern lakes. Conversely, notevery basin with freshwater habitats offered opportune condi-tions for settlement and evolution. Thus, the availability of apersisting, stable geological basin providing continual fresh-water or brackish environments is a prerequisite for hotspotevolution for aquatic gastropods. Further faunal evolution ismainly controlled by climatic factors and surface area (Fig. S1).

Fig. 3. European species richness hotspots in relation to geodynamic development. The Miocene-Pliocene palinspastic maps follow latest reconstructions (13,35). Pleistocene-Recent maps are created in European equidistant-conic projection to ease comparison with the palinspastic reconstructions. The boundariesof the Scandinavian and Alpine ice sheets represent their maximum extent during the Late Glacial Maximum (50). Hotspots with larger stratigraphic rangesare indicated for the time interval of their maximum richness. Numbers refer to the total number of species across a system’s full duration. CP, CentralParatethys; UBWM, Upper Brackish Water Molasse.

Neubauer et al. PNAS | September 15, 2015 | vol. 112 | no. 37 | 11481

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020

Hence, large-scale biodiversity patterns through time rise andfall with the presence of large lakes. As the regression analysesindicate, species numbers are not directly related to a system’stemporal duration per se.

The Cradle of Modern FaunasThe two greatest centers of continental aquatic biodiversity today,the Caspian Sea and Lake Ohrid, both reflect long-lived geo-logical lakes dating back at least into the Pleistocene (42, 49).The great majority of other lakes comprise low diversities andshare similar species compositions. This situation differs greatlyfrom the Miocene and Pliocene distributions, which were char-acterized by a centralization of biodiversity into few, highlydiverse long-lived lakes. Despite the low number of aquatic en-vironments available for past time slices compared with the hugenumber for the most recent interval, the earlier time slices stillpreserve comparable levels of biodiversity (Fig. 2 and Tables S2and S3).The discrepancy between Miocene-Pliocene and Pleistocene-

Recent distribution patterns is partly explained by deterioratingclimates during the Quaternary. Together with the disappear-ance of long-lived Paratethyan lakes at the end of the Pliocene,global cooling and large-scale glaciations dramatically reducedthe availability of suitable habitats. Ice sheet retreat after theLast Glacial Maximum triggered the formation of thousands oflakes in the successively emerging glacial depressions and valleys.The pattern for the Recent faunas is consequently a very youngphenomenon. Although current hotspots are confined to long-lived, geologically induced lakes outside the reach of the Pleis-tocene glacial sheet and permafrost belt, most present distributionsreflect immigration events to postglacial lakes after deglaciationstarting about 19,000 y BP (39). The here proposed scheme iscertainly not restricted to European continental aquatic systems

and not only to the late Cenozoic. The rise and demise ofspecies richness hotspots through time is tightly related to regionaltectonic phases.

Materials and MethodsThe data are based on an extensive literature research and derive from morethan 400 publications on Miocene to Recent continental aquatic gastropodfaunas (Dataset S1). Due to preservation issues, the fossil record is slightlybiased toward lacustrine systems (13), which is why we chose to include onlylacustrine faunas for Recent times. Only freshwater to brackish environ-ments were considered in this study. This approach yielded a total of 2,785species-group taxa (species and subspecies) from 5,414 localities. Althoughthe general geographic frame is Europe, we integrated Turkey and marginalterritories of Azerbaijan, Georgia, and Kazakhstan into the analysis becausebiogeographic entities do not adhere to political borders. For the fossilfaunas, no data were available for latitudes north of 52° due to erosion offormer sediments by the advancing ice sheet during the Pleistocene.

The heat maps were produced in ESRI ArcGIS 10.0 (Fig. 1). An equal-areagrid of 100-km cell size was created using the Behrmann projection to cor-rect for oversampling of certain areas. The available faunas were separatedinto six discrete temporal units: Early Miocene (23.03–15.97 Ma), MiddleMiocene (15.97–11.62 Ma), Late Miocene (11.62–5.333 Ma), Pliocene (5.333–2.588 Ma), Pleistocene (2.588–0.00117 Ma), and Recent. For each cell con-taining localities, the number of species was calculated. For each time slice,an ordinary Kriging algorithm (100-km cell size, 200-km interpolation radius)was performed on the grid-specific species richness. The coloration gradientis equal in all maps, defined based on the maximum range of speciesnumbers (Late Miocene). The maps are intended as graphical visualizationsupporting intuitive recognition of diversity hotspots and should be inter-preted carefully, because the interpolation method tends to exaggerateactual richness toward the margins.

For details on taxonomic and stratigraphic treatment, constraints on areaand duration data for the lakes and regression analyses, see SI Materialsand Methods and Figs. S3 and S4.

Table 1. Late Cenozoic freshwater and brackish gastropod species richness hotspots, reflecting lakes with more than 40 species

Lake/basinType of

environment

Number ofspecies acrossentire duration

Maximumnumber of coevallypresent species*

Interval ofmaximum

richness (Ma)

Maximumtemporal range ofenvironment (Ma)

Ohrid Freshwater lake 68 65 0 1.5–0Caspian Sea Brackish lake 105 92 0 0.88–0†

Curonian Lagoon Brackish lagoon 42 42 0 0.0135–0Tiberino Freshwater lake 42 32 1 3.1–1.55Kos Freshwater lake 41 26 1 3.0–1.4Bresse Fluvio-lacustrine system 64 44 2 4.5–1.5Rioni Fluvial plain 56 52 3 5.5–1.6Slavonia Freshwater lake 163 145 3 4.0–2.6Transylvania Freshwater lake 78 78 4 3.8–0.8Dacia Brackish to freshwater lake 303‡ 159 4 8.6–2.6Metohia§ Freshwater lake 70 36 5 6.04–2.588Galati Brackish embayment 133 131 7 8.6–4.4Bresse-Valence Fluvial plain 60 56 8 10.0–8.0Pannon Brackish lake 605 248 10 11.6–4.0Soceni Brackish embayment 51 51 12 12.7–12.3Bakony Wetlands (?) 87 48 13 17.5–12.5Steinheim Freshwater lake 42 42 13 15–13.9Sinj Freshwater lake 58 50 15 18.0–15.0Drniš Freshwater lake 43 43 15 15.7–15.0Upper Brackish Water Molasse Brackish lake 40 37 16 16.6–17.6Aquitaine Brackish wetlands (?) 59 29 20 28.1–11.6{

Hanau-Wetterau Brackish embayment 49 15 21 28.1–16.1{

*Per one million years for fossil lakes.†Time from last marine transgression.‡Excludes peri-marine Sarmatian assemblages.§Excludes Middle Miocene assemblage, which represents a separate lake development.{Poorly constrained, freshwater/brackish conditions not continuous throughout. For a more detailed version of the table with references for the age models,see Table S1.

11482 | www.pnas.org/cgi/doi/10.1073/pnas.1503992112 Neubauer et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020

ACKNOWLEDGMENTS. We thank a great many colleagues for assistance withliterature research and/or stratigraphic classifications, which markedly improvedthe data: J. Albesa, D. Alba, P. Anadón, V. V. Anistratenko, I. Casanovas-Vilar,A. Engelbrecht, D. Esu, M. Gross, G. Haszprunar, S. Herzog-Gutsch, R. Macalet,I. Magyar, N. Krsti�c, G. Mas Gornals, V. A. Prysjazhnjuk, M. W. Rasser, and

D. Vasilyan. We thank M. Stachowitsch and A. Neubauer for linguistic amend-ments. The comments of three reviewers and the editor greatly improved thework. The study was financed by the Austrian Science Fund (FWF; ProjectP25365-B25: “Freshwater systems in the Neogene and Quaternary of Europe:Gastropod biodiversity, provinciality, and faunal gradients”).

1. Reid WV (1998) Biodiversity hotspots. Trends Ecol Evol 13(7):275–280.2. Orme CDL, et al. (2005) Global hotspots of species richness are not congruent with

endemism or threat. Nature 436(7053):1016–1019.3. Mittelbach GG, et al. (2007) Evolution and the latitudinal diversity gradient: Specia-

tion, extinction and biogeography. Ecol Lett 10(4):315–331.4. Hof C, Brändle M, Brandl R (2008) Latitudinal variation of diversity in European

freshwater animals is not concordant across habitat types. Glob Ecol Biogeogr 17(4):539–546.

5. Ezard THG, Aze T, Pearson PN, Purvis A (2011) Interplay between changing climateand species’ ecology drives macroevolutionary dynamics. Science 332(6027):349–351.

6. Graham CH, Moritz C, Williams SE (2006) Habitat history improves prediction ofbiodiversity in rainforest fauna. Proc Natl Acad Sci USA 103(3):632–636.

7. Champagnac J-D, Molnar P, Sue C, Herman F (2012) Tectonics, climate, and mountaintopography. J Geophys Res 117(B2):B02403.

8. Godard V, et al. (2014) Dominance of tectonics over climate in Himalayan denudation.Geology 42(3):243–246.

9. Micheels A, et al. (2010) Analysis of heat transport mechanisms from a Late Miocenemodel experiment with a fully-coupled atmosphere-ocean general circulation model.Palaeogeogr Palaeoclimatol Palaeoecol 304(3-4):337–350.

10. Quan C, Liu YS, Tang H, Utescher T (2014) Miocene shift of European atmosphericcirculation from trade wind to westerlies. Sci Rep 4:5660.

11. Cook ER, et al. (2010) Asian monsoon failure and megadrought during the last mil-lennium. Science 328(5977):486–489.

12. Wesselingh FP (2007) Long-lived lake molluscs as island faunas: A bivalve perspective.Biogeography, Time, and Place: Distributions, Barriers, and Islands, ed Renema W(Springer, Dordrecht, Germany), pp 275–314.

13. Neubauer TA, Harzhauser M, Kroh A, Georgopoulou E, Mandic O (2015) A gastropod-based biogeographic scheme for the European Neogene freshwater systems. Earth SciRev 143:98–116.

14. Kohn MJ, Fremd TJ (2008) Miocene tectonics and climate forcing of biodiversity,western United States. Geology 36(10):783–786.

15. Renema W, et al. (2008) Hopping hotspots: Global shifts in marine biodiversity.Science 321(5889):654–657.

16. Hoorn C, et al. (2010) Amazonia through time: Andean uplift, climate change,landscape evolution, and biodiversity. Science 330(6006):927–931.

17. Keith SA, Baird AH, Hughes TP, Madin JS, Connolly SR (2013) Faunal breaks andspecies composition of Indo-Pacific corals: The role of plate tectonics, environmentand habitat distribution. Proc Biol Sci 280(1763):20130818.

18. van Leeuwen CHA, van der Velde G, van Lith B, Klaassen M (2012) Experimentalquantification of long distance dispersal potential of aquatic snails in the gut ofmigratory birds. PLoS One 7(3):e32292.

19. Schön I, Martens K (2004) Adaptive, pre-adaptive and non-adaptive components ofradiations in ancient lakes: A review. Org Divers Evol 4(3):137–156.

20. Biteau J-J, Le Marrec A, Le Vot M, Masset J-M (2006) The Aquitaine Basin. PetrolGeosci 12(3):247–273.

21. Kümmerle E, Radtke G (2012) Die Fossilien des Tertiärmeeres im Hanauer Becken. JberWett Ges Ges. Naturkunde 162:59–77.

22. Berger J-P, et al. (2005) Paleogeography of the Upper Rhine Graben (URG) and theSwiss Molasse Basin (SMB) from Eocene to Pliocene. Int J Earth Sci 94(4):697–710.

23. Reichenbacher B, et al. (2013) A new magnetostratigraphic framework for the LowerMiocene (Burdigalian/Ottnangian, Karpatian) in the North Alpine Foreland Basin.Swiss J. Geosci. 106(2):309–334.

24. Zachos J, Pagani M, Sloan L, Thomas E, Billups K (2001) Trends, rhythms, and aber-rations in global climate 65 Ma to present. Science 292(5517):686–693.

25. Mosbrugger V, Utescher T, Dilcher DL (2005) Cenozoic continental climatic evolutionof Central Europe. Proc Natl Acad Sci USA 102(42):14964–14969.

26. Böhme M, Winklhofer M, Ilg A (2011) Miocene precipitation in Europe: Temporal trendsand spatial gradients. Palaeogeogr Palaeoclimatol Palaeoecol 304(3-4):212–218.

27. De Leeuw A, Mandic O, KrijgsmanW, Kuiper K, Hrvatovi�c H (2012) Paleomagnetic andgeochronologic constraints on the geodynamic evolution of the Central Dinarides.Tectonophysics 530-531:286–298.

28. Harzhauser M, Mandic O (2008) Neogene lake systems of Central and South-EasternEurope: Faunal diversity, gradients and interrelations. Palaeogeogr PalaeoclimatolPalaeoecol 260(3-4):417–434.

29. Magyar I, et al. (2013) Progradation of the paleo-Danube shelf margin across thePannonian Basin during the Late Miocene and Early Pliocene. Global Planet Change103:168–173.

30. ter Borgh M, et al. (2013) The isolation of the Pannonian basin (Central Paratethys):New constraints from magnetostratigraphy and biostratigraphy. Global PlanetChange 103:99–118.

31. Jipa DC, Olariu C (2009) Dacian Basin: Depositional Architecture and SedimentaryHistory of a Paratethys Sea (GeoEcoMar, Bucharest, Hungary).

32. Sissingh W (2001) Tectonostratigraphy of the West Alpine Foreland: Correlation ofTertiary sedimentary sequences, changes in eustatic sea-level and stress regimes.Tectonophysics 333(3-4):361–400.

33. Stoica M, et al. (2013) Paleoenvironmental evolution of the East Carpathian foredeepduring the late Miocene-early Pliocene (Dacian Basin; Romania). Global PlanetChange 103:135–148.

34. Popov SV, et al. (2004) Lithological-Paleogeographic maps of Paratethys. 10 Maps.Late Eocene to Pliocene. Cour Forsch-Inst Senckenb 250:1–46.

35. Brigham-Grette J, et al. (2013) Pliocene warmth, polar amplification, and steppedPleistocene cooling recorded in NE Arctic Russia. Science 340(6139):1421–1427.

36. Koenig SJ, DeConto RM, Pollard D (2014) Impact of reduced Arctic sea ice onGreenland ice sheet variability in a warmer than present climate. Geophys Res Lett41(11):3934–3943.

37. Mangerud J, et al. (2004) Ice-dammed lakes and rerouting of the drainage ofnorthern Eurasia during the Last Glaciation. Quat Sci Rev 23(11-13):1313–1332.

38. Andrén T, et al. (2011) The development of the Baltic Sea basin during the last 130 ka.The Baltic Sea Basin, eds Harff J, Björck S, Hoth P (Springer, Berlin), pp 75–97.

39. Clark PU, et al. (2009) The Last Glacial Maximum. Science 325(5941):710–714.40. Hauswald A-K, Albrecht C, Wilke T (2008) Testing two contrasting evolutionary pat-

terns in ancient lakes: Species flock versus species scatter in valvatid gastropods ofLake Ohrid. Hydrobiologia 615(1):169–179.

41. Albrecht C, Hauffe T, Schreiber K, Wilke T (2012) Mollusc biodiversity in a Europeanancient lake system: Lakes Prespa and Mikri Prespa in the Balkans. Hydrobiologia682(1):47–59.

42. Wagner B, et al. (2014) The SCOPSCO drilling project recovers more than 1.2 millionyears of history from Lake Ohrid. Sci Drill 17:19–29.

43. Allen AP, Gillooly JF (2006) Assessing latitudinal gradients in speciation rates andbiodiversity at the global scale. Ecol Lett 9(8):947–954.

44. Mannion PD, Upchurch P, Benson RBJ, Goswami A (2014) The latitudinal biodiversitygradient through deep time. Trends Ecol Evol 29(1):42–50.

45. Whittaker RJ, Fernández-Palacios JM (2007) Island Biogeography. Ecology, evolution,and conservation (Oxford Univ Press, Oxford, UK).

46. Franzén M, Schweiger O, Betzholtz P-E (2012) Species-area relationships are con-trolled by species traits. PLoS One 7(5):e37359.

47. Wagner CE, Harmon LJ, Seehausen O (2014) Cichlid species-area relationships areshaped by adaptive radiations that scale with area. Ecol Lett 17(5):583–592.

48. Geary DH, Staley AW, Müller P, Magyar I (2002) Iterative changes in Lake PannonMelanopsis reflect a recurrent theme in gastropod morphological evolution. Paleobiology28(2):208–221.

49. Van Baak CGC, et al. (2013) A magnetostratigraphic time frame for Plio-Pleistocenetransgressions in the South Caspian Basin, Azerbaijan. Global Planet Change 103:119–134.

50. Ehlers J, Gibbard PL, Hughes PD (2011) Quaternary Glaciations—Extent and Chronology.A Closer Look (Elsevier, Amsterdam).

Neubauer et al. PNAS | September 15, 2015 | vol. 112 | no. 37 | 11483

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

Aug

ust 2

2, 2

020